Inside Bending Spoons: Finding Talent, Leveraging AI & Driving Operational Excellence | Luca Ferrari

探秘 Bending Spoons：发掘人才、善用 AI 与实现卓越运营 | Luca Ferrari 在本次与 David Senra 的对话中，Bending Spoons 的 Luca Ferrari 分享了公司从第一性原理出发、致力于打造“史上最优秀公司”的经营理念。受早期创业项目 Evertale 的经验启发以及米兰地理位置的相对孤立，Ferrari 认为在科技行业中，天赋和进取心远比过往经验更重要，因为技术经验会迅速过时。为此，Bending Spoons 将旗下 50 多家被收购业务的招聘与解聘工作集中管理，由约 50 名工程师组成的人才团队通过专有测试和 100 多个预测信号来评估候选人，同时摒弃了传统的奖金与机械化 KPI。此外，公司弱化了固定的职位边界，强调全员以公司整体成功为己任，并通过严格的优先级排序，剔除负投资回报率（ROI）的项目，专注于能够产生实际现金流和长期价值的举措。 Luca Ferrari 表示，Bending Spoons 致力于打造“史上最优秀的公司”，由于地处意大利米兰、远离主流创业中心，公司得以从第一性原理出发独立探索，并尽量淡化“创始人”标签，以实际贡献衡量每个人。 基于早期创业项目 Evertale 的教训，Ferrari 认为技术行业的经验很快就会过时甚至让人染上不良习惯，因此公司优先看重原始天赋（聪明的头脑与强烈的进取心），并为高水平团队减少繁琐规则、提供极大自主权。 Bending Spoons 将旗下 50 多家业务的人才选拔与解聘集中交给由约 50 名工程师组成的团队负责，利用专有测试和超过 100 个预测信号（包括对待招聘后勤人员的态度）来科学评估潜力，且不设奖金或机械化 KPI。 公司打破僵化的岗位界限，倡导全员都承担“帮助公司成功”的同一份工作，并通过严格的优先级排序砍掉大量投资回报率（ROI）为负的可选项目，摆脱单一资产虚荣指标的束缚，专注于创造净现金流。

Bending Spoons co-founder and CEO Luca Ferrari argues that digital companies achieve outsized performance by replacing conventional corporate playbooks with first-principles operational excellence, pairing high-density raw talent with aggressive architectural simplification. By rebuilding acquired software assets on a proprietary technology platform and operating them permanently, Ferrari contends the company extracts far more cash flow and product velocity than traditional operators or private equity buyers. Prioritizing Raw Talent Over Corporate Tenure (5:47) Ferrari dismisses standard corporate reliance on industry experience, noting that software engineering methods depreciate rapidly and prolonged tenure in politicized, low-standard environments can actively degrade an individual's effectiveness. Instead, Bending Spoons optimizes for raw cognitive horsepower, resourcefulness, and what Ferrari terms "extreme ownership"—an intense, intrinsic drive to see the organization succeed. To identify elite contributors with negligible resumes, a centralized 50-person talent team of engineers and specialists evaluates over 100 subtle behavioral signals, allowing the firm to hire fewer than 300 candidates out of more than 800,000 applicants in a single year. To preserve alignment, Bending Spoons completely centralizes hiring and termination across its 50+ businesses, stripping line managers of the authority to hire merely for short-term operational relief. Ferrari also eliminated variable bonuses and mechanistic key performance indicators across the organization. In his view, tying compensation to rigid metrics leads employees to game incentives rather than maximize enterprise value, whereas high fixed compensation combined with genuine operational trust yields superior long-term dedication. Relentless Simplification and Radical Headcount Compression (59:28) Left unchecked, corporate systems accumulate nonlinear complexity, where adding steps, features, or staff introduces compounding coordination friction. To counteract this decay, Bending Spoons enforces an asymmetric burden of proof: anyone advocating to add organizational rules, software features, or personnel must rigorously justify the systemic complexity, whereas team members advocating for deletion or maintaining simplicity carry no burden of justification. Ferrari applies this principle directly to software products, willingly pruning legacy features used by minor segments of the user base to eliminate technical debt for the overwhelming majority. This anti-complexity posture enables extreme operational leverage. Following the acquisition of Evernote for approximately $200 million—which was generating around $90 million in revenue at roughly break-even with 350 employees—Bending Spoons systematically dismantled technical debt and overhauled the codebase. The Evernote team was initially restructured down to 50–60 people and is currently maintained by a core team of approximately 20 people, while generating between $100 million and $200 million in revenue with company-wide adjusted operating income margins hovering around 54% to 55%. Proprietary Operating Infrastructure and AI Agents (70:54) Rather than relying on disjointed third-party software subscriptions, Bending Spoons runs its portfolio on a proprietary suite of over 50 internally developed tools spanning global payments, statistical A/B testing, lifetime value prediction, and algorithmic recruiting. Acquired products are migrated onto this shared platform, transforming fragmented standalone businesses into a standardized operational engine. Ferrari refuses to commercialize or license this toolset to third parties, asserting that its value stems from deep native interdependence and serves as an enduring competitive moat for acquisitions. Automation is further accelerated through internal artificial intelligence systems like "Alt Spooner," an autonomous agent integrated into Slack that mirrors each employee's exact operational permissions. Team leads use the agent to synthesize user support feedback, diagnose root causes within codebases, draft technical patches, and assign reviews to lead engineers within single commands. This platform-level leverage allows tiny product teams to ship updates faster than legacy organizations burdened by massive headcounts. Operational Re-Engineering Versus Traditional Private Equity (77:07) Ferrari rejects characterizing Bending Spoons as a private equity firm, emphasizing structural differences in holding period, operational depth, and portfolio integration. Unlike closed-end investment funds required to sell assets within three to seven years to return capital to limited partners, Bending Spoons acquires software assets to hold and run indefinitely. Instead of executing superficial financial engineering or basic corporate overhead trimming, Bending Spoons fundamentally rearchitects underlying codebases, cloud infrastructure, and monetization mechanics. This operational capacity alters Bending Spoons' acquisition strategy. Because its unified platform extracts substantially more cash flow from mature assets, Ferrari notes the company can comfortably submit fully informed, highly competitive bids that outprice traditional financial buyers while preserving high projected returns. Bending Spoons avoids aggressive haggling, typically submitting firm, generous offers and actively encouraging hesitant sellers to test the wider market, anticipating that sellers will return once discovering that rival acquirers cannot match the valuation. Bending Spoons plans an eventual public market listing to access deep capital pools for larger transactions, though Ferrari practices "productive procrastination" to defer the transition until real-world necessity demands it. An unresolved question is whether public market scrutiny and external pressures for smooth quarterly earnings will challenge Ferrari's commitment to unmeasured compounding logic, ruthless feature elimination, and indifference toward short-term vanity metrics.

[0:02] So, we're going to start this episode in [0:03] a locked in stance because we've just [0:05] been talking off camera and I was like, [0:06] "God damn it, we need to start recording [0:07] immediately." I didn't even want to [0:09] think to start here, but uh you noticed [0:10] the apploving mug and then you're like, [0:12] "Oh, Farugi." And then you laughed. What [0:14] do you think of Adam Farugi? [0:16] >> Well, it's great. I mean, uniquely [0:19] focused, uh, ruthless, and I mean it in [0:23] a, you know, in a positive way. When [0:25] there's a goal, goes for it. Very [0:27] rational, effective. I mean 10 out of 10 [0:31] on that in in those areas I think. [0:33] >> So when I published the episode that I [0:35] did with him, I think I titled it like [0:36] the best founder no one's ever heard of [0:38] because at the time he was running like [0:39] I don't know like $150 billion market [0:41] cap company with like 400 employees and [0:43] they're printing like 6 billion in cash [0:45] and he kind of lays out exactly like his [0:47] we were talking about fanaticism before [0:48] we started recording. It's like he's [0:49] just fanatical. It's like success of his [0:51] company goes before almost anything or [0:53] no it does go before almost any anything [0:55] in his uh life. He's just completely [0:58] obsessed and committed, you know, to [1:00] essentially like excellence. I think you [1:02] share that trait with him. So, we had [1:03] lunch together probably 6 months ago. I [1:06] talked to you right after. I was like, [1:07] man, you got to do the show because I [1:08] know a lot of founders. I don't know any [1:10] other founders that think like you. One [1:12] of the things that you said that I think [1:13] everything else that all the other ideas [1:16] flow from this is that you want to be [1:17] the best in the world at what you do, [1:19] even if that's not possible. Can you [1:21] talk a little bit more about that? I've [1:23] always been uh kind of polarized in my [1:26] in my interests. I I either choose to do [1:30] something and then I'll try to uh match [1:34] that out, try to be the best uh or part [1:37] of the best team or I will try not to do [1:40] it at all or if it really has to be done [1:43] uh then uh I'll literally try to kind of [1:46] just check the boxes for you know [1:48] minimum commitment. All sorts of rewards [1:52] emotional and material are at the [1:55] extremes. I I think I have a close to 10 [1:58] out of 10 relationship with my wife. I [2:00] think to me that's worth that's worth [2:02] 100 times more than having an adequate [2:04] relationship with my wife. Uh same with [2:07] with my job and my colleagues trying to [2:09] build best company there ever was. Um [2:13] and we understand that's aspirational [2:15] and and likely nearly impossible. But I [2:17] think if we get close to that [2:20] accomplishment, the the rewards, the the [2:22] fulfillment, the satisfaction, the [2:23] learning along the way, uh financial [2:25] rewards will be just exponentially [2:27] greater than just doing well enough. And [2:29] so I think you have limited time and [2:31] energy. You want to find one or very few [2:35] pursuits or to try to go all out and [2:39] everything else keep it uh you know, [2:41] eliminated if you can or keep it at the [2:42] bare minimum. So you just said you're [2:44] trying to build the best company there [2:46] ever was. [2:47] >> Yeah. Just again aspirationally don't [2:49] take it it's not it's not not meant to [2:50] be arrogant. I I know we have a very [2:52] sling chance but just you know the [2:54] waking up in the morning and thinking [2:55] we're not building a nice church or [2:57] trying to build the greatest cathedral [2:59] that anybody has ever built. That's a [3:01] lot more exciting to me. Gets you [3:03] further. It's more fun energizing. [3:05] Better people will want to work with [3:06] you. And I think one of the big ways in [3:08] which life is interesting is surrounding [3:10] yourself with amazing people. better [3:11] people than than you are if possible. [3:14] So, [3:14] >> do your co-founders feel the same way? [3:16] >> I mean, you'd have to ask them. I think [3:17] we probably for the most part, I'd say. [3:20] Yeah. [3:21] >> But is this something that you guys [3:22] repeat to each other like throughout the [3:23] company? You're trying to build the best [3:24] company ever. [3:25] >> We're not big on founders. I know this [3:27] is maybe uh ironic to say, you know, [3:29] given your your your podcast, but we we [3:33] try to eliminate the idea of founder [3:35] from this company as much as possible. [3:36] We we think it distracts people from [3:38] from the company. the company is a [3:40] center and and whether you're a founder [3:42] or you joined a little bit later all [3:44] that matters is your contribution your [3:47] trajectory the people at at the company [3:49] those at least I know the best and with [3:51] whom I work the closest regardless of [3:53] whether they're founders I think broadly [3:55] they share this ambition so yes but it's [3:58] not necessarily a founder thing it's [4:00] more of a banish thing [4:01] >> the way I've been describing uh you to [4:04] other founders is like it's almost like [4:06] Luca is like the goalpost island of like [4:09] entre- reneurship, right? Because if I'm [4:11] not mistaken, [clears throat] when we [4:12] were talking, you're like, well, I don't [4:13] really listen to like podcasts. I didn't [4:15] read books. I didn't study other [4:16] entrepreneurs. You've kind of evolved [4:18] the way you build your company like [4:20] completely independent of anything going [4:21] on around you. [4:23] >> Yeah. I think part part of that is I [4:25] don't know how much the audience knows [4:26] about Danish, but we we started in [4:29] Copenhagen, Denmark, quickly thereafter [4:31] moved to Milan, Italy. These are not [4:33] exactly especially Milan not exactly and [4:35] certainly not at the time over a decade [4:36] ago a center of uh entrepreneurial [4:40] pursuit and an ecosystem where you you [4:42] know you turn left and right and you [4:44] have all these other startup startups or [4:46] or advisers and whatnot. So and of [4:48] course we were absolute nobodyies so [4:50] it's not that we could pick up the phone [4:51] and call Jeff Bezos right. So we just [4:53] had to figure things out on our own. We [4:56] were trying to build aspirationally [4:58] speaking the best company in the world. [5:00] If you emulate what most people are [5:02] doing, you're pretty much guaranteed to, [5:05] you know, at best be mediocre, mediocre [5:07] plus, maybe you execute a little bit [5:08] better, but if you want to try to be the [5:10] best of the best, you need to take some [5:12] risks and uh rethink things. And so we [5:16] figured, okay, let's try to experiment, [5:19] invent, think from first principles. And [5:22] we will make more mistakes. It will take [5:24] longer than if we copied some of the [5:26] tried and tested approaches, but we [5:27] should be able to find at least a few [5:29] insights, a few new ways that will set [5:32] us apart. And I think being more [5:34] isolated geographically has probably [5:36] played to our advantage in that regard [5:38] so that we weren't under the influence [5:40] of of of the mantras that everybody, you [5:43] know, the big startup hubs over time was [5:45] was preaching. [5:47] >> Let's give a little bit of background of [5:48] Benny Spoons. You had a startup before [5:51] Benny Spoons that failed, right? Yeah. [5:53] >> Okay. [5:54] >> What did you learn from that failure and [5:55] then what lessons did you learn from [5:57] that that helped you start Bennington [5:59] then? [5:59] >> Yeah. So that was called Evertale. We [6:02] were it went on from 2010 to 2013. We [6:05] were using AI to write uh diaries [6:09] automatically. So you would install an [6:11] app and then it would collect data and [6:13] uh figure out what what you'd done, [6:14] where you had gone and whatnot. It was [6:17] actually pretty cool but never managed [6:19] to make it scale. Commercial failure. [6:22] some of the most uh crucial lessons were [6:24] one the importance of of talent. So we [6:27] had a very small team at peak maybe 12 [6:29] people but we saw that the contribution [6:32] uh of the the best person we had on the [6:34] team relative to like the say the medium [6:37] person forget about the you know the [6:39] bottom but was easily 10 times as great [6:43] like literally worlds apart. So that [6:46] taught us okay the range of productivity [6:48] at least in our field in digital [6:50] technology is massive. So the value of [6:53] having on board that sort of individual [6:56] uh is gigantic and also [6:59] that person who was performing at the [7:01] peak in that group was actually one of [7:02] the least experienced people. And so [7:04] that that showed us uh okay maybe [7:07] experience [7:08] you know certainly valuable but not as [7:10] critical as people sometimes uh tell you [7:13] it is. If you have someone who's really [7:14] smart and and really cares, often [7:17] they'll be able to deliver as much [7:18] value, if not a lot more value, than [7:21] someone with a lot more experience. [7:22] >> Let's give a concrete example. So, I'm [7:24] just going to pull up the notes. When we [7:25] were having lunch, I was like, "Oh, this [7:27] is too good." And I started just texting [7:28] on WhatsApp. I'm like, and I think at [7:30] the bottom I say like, "These are notes [7:32] for when you do the show." Even though [7:33] this was like 8 months ago or whenever [7:34] it was. And you you mentioned that [7:36] you're like, "Hey, um, you said [7:37] something interesting. You're like, "I'd [7:38] rather hire young graduates." Let's talk [7:40] about the Evernote story in one second. [7:42] So you said most executives are [7:44] overvalued or overrated in your opinion. [7:46] I'd rather hire young graduates [7:48] graduates find someone good and then [7:50] saturate their capacity. Can you give [7:52] examples of how you've done this? [7:54] >> Why talent let's say over experience? I [7:57] think there are a few few reasons for [7:58] that. Number one, most of the things we [8:00] do and I mean broadly in most industries [8:03] probably certainly in the technology [8:04] industry are not rocket science. They do [8:07] not require immense amounts of uh [8:11] notional knowledge uh and and repeated [8:14] uh you know uh extensive track records. [8:18] They they require um uh actually a a [8:22] good brain and a desire to do well to [8:24] achieve um first and foremost and also [8:28] our field technologically but also in [8:30] terms of customer expectations uh [8:32] evolves very quickly. So experience uh [8:36] gets stale relatively fast. [8:38] >> Wait, before you go on, sorry, I'm going [8:39] to interrupt you. Explain more about [8:41] customer expectations evolve rapidly in [8:44] your field. Yeah, I think you know I I [8:46] I'm not sure how it works if you sell [8:49] sold, but when it comes to selling uh [8:52] technical tools, what people consider [8:55] excellent today or an intuitive uh [8:57] interface or you know effective [8:59] monetization uh are very different from [9:02] what uh uh things looked like uh 10 or [9:05] 15 years ago. Completely different. I [9:07] mean, if uh I'm sure at least, you know, [9:11] people in the audience who were at [9:12] least, I don't know, 35 years old will [9:14] remember what software looked like in [9:15] the early 2000s. And, you know, by [9:18] today's standards, that's primitive and [9:20] almost unacceptably bad and people would [9:22] never use it or buy it and the ways you [9:24] build that software. And by the way, [9:26] that's just the say customerf facing [9:28] layer. But then behind the scenes how [9:30] and this is only something that probably [9:32] people can understand if if they've [9:34] built software before uh or work with AI [9:36] the way you efficiently wrote software [9:40] in the you know in 2010 bears very [9:44] little resemblance to how you do that [9:45] today in 2026. Whatever people learned [9:48] back then yes some of it will will will [9:50] port. I'm sure you know you are more [9:52] mature emotionally, you know how to work [9:54] with others and whatnot, but a lot of [9:56] that experience basically you can throw [9:58] it away. The value of accumulating many [10:00] years of experience is not as great I [10:02] believe as as some people think it is. [10:04] And additionally, not all experience is [10:06] created equal. Uh you can actually get [10:08] worse through experience. If you're [10:10] exposed to low standards, for example, [10:12] of performance, you'll normalize those [10:15] over time and you'll actually be a less [10:18] capable team member than someone who has [10:20] never been exposed to any standards and [10:21] maybe you know naturally is inclined to [10:24] believe idealistically perhaps that the [10:25] bar should be held higher. or if you're [10:28] if you've been working for a long time [10:29] in an organization where the way to [10:32] progress and succeed is by pleasing [10:35] others and uh doing what they tell you [10:37] to do even though you don't necessarily [10:39] think it's optimal for the organization [10:41] call it politics uh I don't know that [10:43] that experience will make you a lot more [10:45] capable uh necessarily if you for [10:47] example join a company like Venice where [10:48] I'd like to think we're a radical [10:50] meritocracy and we try to be rational in [10:52] in deciding and and and do what's right [10:54] for the company. So for all these [10:56] reasons, experience can be extremely [10:58] valuable, but but it's not necessarily [11:00] extremely valuable. But talent, meaning [11:02] a good brain and a massive eagerness to [11:04] excel, grow, make an impact, those never [11:07] fail to be valuable. And so, you know, [11:09] in a competitive labor market where you [11:12] can't have everything at the same time, [11:13] you need to prioritize. We tend to favor [11:16] talent also because experience, we can [11:18] we can give it to you. You know, we just [11:19] have to be a little bit patient. Make [11:21] sure we expose you to good challenges [11:22] and surround you with amazing [11:24] colleagues. you'll accumulate experience [11:25] very quickly. first principles really [11:27] and based on those anecdotes and [11:29] observations during the first company I [11:32] co-ounded but also at Bennis Poons in [11:33] the early days we repeatedly saw that [11:35] that thesis uh was supported by facts [11:39] and and so we kept investing in first of [11:41] all attracting excellent talent and then [11:45] creating uh ideally the perfect [11:47] conditions for that talent to to [11:49] flourish very very quickly because of [11:50] course you need to establish your [11:53] structure and operations to get the the [11:55] the most out of uh the human capital you [11:57] have. I would build a company uh [12:00] differently if I had uh to work with [12:03] inferior talent than than than we do [12:06] because we believe we have amazing [12:07] colleagues. [12:08] >> Well, say more about that. So, how did [12:09] you build the company? [12:10] >> Yeah. So, for example, I think if you if [12:12] you have uh and maybe there's nothing [12:14] you can do about it. If you have [12:16] mediocre talent, then I think the appeal [12:19] of of process and procedures becomes [12:21] greater. Checks, rules, cuz you need to [12:24] guide more. You can count on people to [12:26] problem solve autonomy as well. You can [12:28] count on them coming to work with the [12:30] fire in their belly as much. You know [12:33] process and procedures sometimes we say [12:35] they are terrible but honestly they can [12:37] be you know the the lesser evil if you [12:39] are in that situation. If you are uh [12:42] lucky enough or good enough for whatever [12:44] reason to have a very strong team then I [12:46] think generally speaking you want to [12:47] have as few rules as possible. It's not [12:50] that process and procedures are always [12:51] bad. are cases where you want to have [12:53] some of those but to the extent possible [12:55] get rid of them. Uh and give people [12:57] massive leeway to express and develop [13:00] their talent. Make them feel trusted so [13:03] that they will bring the best of [13:05] themselves to work. Uh and that will be [13:07] good for everybody. They they get to do [13:09] better work. They get to learn a lot [13:10] faster. Their careers can be [13:12] turbocharged. But again, that only works [13:14] if you have a very good team. I suppose [13:16] it's probably similar with with sports. [13:18] I would imagine that how you coach, and [13:20] I'm going to the extreme. I'm not saying [13:21] Ben is, you know, that, but if you if [13:24] you were coaching the, you know, the uh [13:27] Team USA dream team with with Jordan and [13:30] Barlay and those guys, you you would do [13:32] it a certain way that that would be [13:34] different the way you would optimally [13:36] coach a team of kind of modest talents. [13:39] Uh, you can probably win with both. It's [13:40] a lot easier to win with Jordan, but [13:42] certainly you're not going to tell your [13:43] more modest talents, okay, go and figure [13:45] it out. You will try to give them a [13:47] system that's a lot more guiding. So we [13:49] try to approximate as much as possible [13:51] like the dream team uh aspirationally um [13:54] and then uh give a lot of space for for [13:57] those people to to live up to [13:59] expectations. [14:00] >> I want to tell you about the presenting [14:02] sponsor of this podcast ramp. I have [14:04] been reading a lot about SpaceX lately. [14:06] SpaceX is one of the most valuable [14:08] businesses in the world and one of the [14:09] main themes in the history of SpaceX is [14:11] constantly attacking and questioning [14:13] your cost. Ramp helps many of the most [14:15] innovative businesses in the world do [14:17] exactly that. The median company running [14:19] on ramp cuts their expenses by 5%. And [14:22] one thing SpaceX has demonstrated is [14:25] that a religious dedication to [14:26] controlling costs can help actually [14:28] increase revenue because you can pursue [14:30] opportunities you couldn't otherwise. [14:32] And we see that in the RAMP data, too. [14:35] The median company running on ramp also [14:37] grows their revenue by 16%. So when [14:40] you're running your business on ramp and [14:42] your competitors are not, you have a [14:44] massive competitive advantage that [14:46] compounds over time. Ramp is the only [14:48] platform designed to make your finance [14:49] team faster and happier. Many of the top [14:52] founders and CEOs I know run their [14:54] business on ramp. I run my business on [14:56] ramp and you should too. Go to ramp.com [14:58] to learn how they can help your business [15:00] save time, save money, and grow revenue. [15:03] That is ramp.com. [15:05] Okay, but go back. How do you identify [15:06] talent when that talent doesn't have [15:08] experience? Have you done that? I think [15:10] you can think about life in general as [15:13] when it comes to people and [15:14] accomplishments as you control certain [15:18] inputs [15:20] you know how much you work what you do [15:22] when you work for example to make it [15:24] super simple and then there's a bunch of [15:27] u elements you don't control them [15:29] boundary conditions sometimes knowable [15:31] sometimes unknowable sometimes fixed [15:33] sometimes shifting and the combination [15:35] of those ingredients leads to outputs or [15:38] results Call it what you will. Outputs [15:41] results are fairly easily observable. [15:43] Sometimes there is a gigantic amount of [15:45] inputs and that go into achieving a [15:47] certain output and and those inputs go [15:49] into it for a long period of time. [15:50] Whereas the output can be very simple. [15:52] You know, you the company achieved a [15:54] certain amount of revenue whatever you [15:55] won a certain tournament you know no [15:56] matter the field. And so it's a lot [15:58] easier to just look at the the results [16:00] the outputs. It's the convenient uh [16:02] sometimes lazy way. But life typically [16:05] in most pursuits most most endeavors is [16:07] so complicated. The amount of inputs, [16:10] the amount of people contributing [16:11] different inputs, the amount of [16:12] variables you don't control, those [16:13] boundary conditions are such that if you [16:16] just look at the outputs, sometimes you [16:17] get a massively distorted picture of [16:20] what the person contributed. A lot of it [16:23] could be luck. Uh a lot of a lot of it [16:25] could be actual human performance, but [16:28] not by that individual, someone else who [16:29] just you you know you failed to know was [16:31] involved, maybe wasn't as [16:33] >> front-facing. Now the more extensive [16:36] someone's track record is, the more [16:37] results likely correlate with actual [16:40] talent, take an investor. You can get [16:42] lucky one year, two years, four years, [16:44] but unlikely 30 years. I don't think [16:46] anybody would question that Warren [16:48] Buffett is almost certainly insanely [16:50] good at investing. You can never prove [16:52] it definitively. You could have been [16:53] lucky for decades, but that's [16:54] astronomically unlikely, right? But if [16:56] you find any hedge fund who delivered [16:58] 50% performance in any given year, it [17:00] could easily be they got lucky with two [17:02] stock picks. Maybe they were those were [17:04] terrible ideas. Maybe they picked them [17:06] for the wrong reasons, but you know, [17:07] whatever. The boundary conditions [17:09] changed and they made a lot of money. [17:10] So, when you pick someone who's very [17:12] experienced, decades of of work, [17:15] probably you can get away with just [17:17] looking at the results, uh maybe some [17:19] reference calls, uh very likely you'll [17:21] get a reasonably accurate uh uh [17:24] assessment. But if you have to pick [17:25] someone who hasn't even graduated yet or [17:27] maybe he's been in the you know in the [17:29] workforce for a year or two then you [17:31] don't have the luxury of using this like [17:32] the the sample is too small and so [17:35] >> you need to find different ways um and [17:37] and something we [17:38] >> we do is uh making extensive use of [17:41] testing. So we we develop tests that [17:44] people go through um that uh we have [17:46] found over time uh proxy pretty well [17:51] their same mental capacity. in [17:54] faculties. For example, uh we have over [17:57] time developed uh we've really built a [18:00] science out of uh studying uh people's [18:03] track records including academic records [18:05] and personal projects and and similar [18:07] things whereby in some application we [18:10] identify over 100 different signals and [18:13] through those we predict their long-term [18:15] potential. It's not entirely dissimilar [18:17] for what a stock trader who's uh [18:20] algorithmically oriented would do. you [18:23] know the more quantitative hedge funds [18:25] they would identify and test hundreds or [18:28] even thousands of signals. Many of these [18:30] are only very marginally predictive but [18:32] in aggregate they they you know they [18:34] make you predictive enough that you can [18:36] succeed because you're just better than [18:37] others. And so we we have all these [18:39] signals some of which are completely [18:41] obvious. You know someone's GPA [18:44] tells you something. You know higher GPA [18:46] is better than lower GPA. It doesn't [18:47] prove anything. It's not definitive but [18:49] it's a good sign. some are more subtle [18:51] and and we keep working and investing [18:52] scientifically in identifying and and [18:54] and measuring these signals. So over [18:57] time I think we have developed a [18:58] competitive advantage in finding people [19:02] who despite their minimal track record [19:06] are very likely to excel. [19:08] >> Okay. So, this is one of the notes that [19:10] we talked about at lunch and this is one [19:12] of the things I text you where you [19:13] essentially said you need a lot of other [19:15] companies kind of uh like they don't [19:17] really not many companies have brilliant [19:19] people in HR and you're like no you [19:20] actually need brilliant people in HR. [19:22] You said that you've you can make hiring [19:24] a science that you had at the time a [19:26] team of like 50 and you actually these [19:28] are engineers these are not typically [19:30] people that you find right and then you [19:32] say you compare one two several years of [19:35] performance the signals from their CV [19:36] which is what you just described or [19:38] these 100 signals signals rather and [19:40] that you centralized hiring and firing [19:44] and I guess these are you call them [19:45] talent managers and they're in charge of [19:47] both entries and exits. The hiring and [19:49] firing is centralized for all the [19:51] companies that you own. Yeah, it's a [19:52] very unusual but and by the way the [19:54] signals are are not just from the CV but [19:56] they could be from email exchanges with [19:58] our recruiting team from uh the tasks we [20:00] ask a candidate to go through anything [20:03] really anything qualifies as as a signal [20:05] and we just care that it's predictive. [20:07] It doesn't have to be something that [20:09] let's say intuitively immediately makes [20:12] sense as long as we can prove it's [20:13] likely not a statistical flu but [20:15] actually [20:15] >> give me an example that I'm a little [20:16] confused. Well, I mean, for example, one [20:18] of the qualities that we value in people [20:20] is uh because what we do is such a team [20:23] sport. You need to be somewhat [20:24] collaborative. You don't have to be the [20:26] nicest person in the world, you know, [20:27] like but if you are arrogant or uh [20:30] dismissive of others or just an [20:32] that that doesn't work typically unless [20:34] you're a freaking genius, we might, you [20:36] know, accept it occasionally, but it's [20:38] almost nobody is. So, for most people, [20:40] you need to be nice enough. And uh we [20:42] find though that when they're [20:44] interviewing even especially [20:47] because these are smart because [20:49] they've already passed the more [20:51] cognitive oriented test, they are [20:52] generally quite nice cuz they know that [20:54] if they come across as super arrogant, [20:56] they're not going to get an offer. We [20:58] found that whether the interviewer felt [21:00] that the interviewee was uh you know [21:04] open to criticism and reasonably [21:06] pleasant to to to talk to wasn't a good [21:09] predictor of whether they were actually [21:11] collaborated on the job. So we we have a [21:13] role call it like a it's almost like [21:15] customer support people who help you [21:16] with the more the logistics of your [21:18] application process of you know [21:20] scheduling interviews or uh so it's more [21:22] of a support role which clearly does not [21:25] come across as in any way connected to [21:28] the final assessment how people interact [21:30] with those is a lot more predictive of [21:33] how they actually are as human beings. [21:35] And so we found that people who were [21:37] hurt and sometimes of course even [21:39] disrespectful occasionally that's rare [21:41] ultimately that that predicted uh poor [21:43] behavior in a social context uh much [21:46] better than how they interacted in an [21:47] interview. You know, again, one of [21:48] hundreds of signals uh in and of itself [21:51] is not definitive, but uh it helps form [21:54] an accurate picture at the end of the [21:55] day. And that that would be for [21:56] collaboration, but then we would have [21:57] others for hardworking attitude, uh [22:00] whether you're creative, whether you're [22:02] logical in your thinking, uh [22:05] perseverance, you know, the important [22:06] things that identifying what's important [22:09] is not rocket science. You could imagine [22:11] what's important to performance. The [22:12] difficult part is spotting it through [22:15] these kind of subtle signals. And the [22:17] interesting part is that you just said [22:18] how many companies do you own right now? [22:20] >> We bought um a little over 50 businesses [22:23] over time, [22:24] >> but you could see at like if you own 50 [22:26] businesses, there's multiple different [22:27] ways. You're kind of like a [22:28] conglomerate, you know, like the hiring [22:31] could be pushed down to the actual [22:32] individual company level and you're [22:34] like, "No, no, this is so important." [22:36] And in the talk I had all that matters. [22:38] >> I I was going to say maybe the most [22:40] important. So you just said it's almost [22:41] all that matters. So it is the most [22:42] important. So I'm going to centralize [22:43] this. Are you also then the [22:45] centralization allows you to kind of [22:47] take the insights that you learn from [22:48] one of the 50 and disperse it to the [22:50] other ones. Is that what happens? [22:51] >> I firmly believe that in business [22:55] entrepreneurship [22:57] the number one thing is call it strategy [22:59] meaning what we're trying to do how and [23:02] why do we think it's going to work. That [23:04] is if you have a terrible strategy you [23:06] can have you know the best team you're [23:07] not going to go anywhere. Um [23:11] but once you have a strategy that makes [23:13] sense, the team is almost all that [23:15] matters. I'd say the team and the [23:16] culture which is like the the rules of [23:19] how we engage with one another uh it's [23:21] basically almost all that matters and so [23:23] I don't think there's overinvesting in [23:25] creating a great team within vision and [23:26] we try to be uh generous in our time and [23:29] resources when it comes to that. So why [23:31] why centralizing hiring and parting ways [23:33] with with people? I think there are [23:35] plenty of good reasons for it. uh one is [23:37] that team managers in most by the way [23:40] most companies it's say you are you run [23:42] a team of 10 people most companies you'd [23:44] be deciding maybe there's a budget like [23:47] you know you can hire two people but [23:49] once that's in place you'd be deciding [23:51] who gets hired probably HR will screen [23:54] CVS and send you and maybe pre-in a [23:56] bunch of people and send you maybe five [23:57] candidates and then you pick the one you [23:59] prefer end of story we think that system [24:02] is bad for a few reasons first of all [24:06] hire Hiring managers meaning that that [24:07] person who runs the team have almost all [24:10] the wrong incentives in hiring. [snorts] [24:12] For instance, uh they probably don't [24:15] want to work late or week or on the [24:17] weekends. They they feel they they need [24:20] help. Uh so they will try to fill the uh [24:23] position as quickly as possible. I'm [24:26] sure they will not hire someone who they [24:28] think is a net negative for the team, [24:29] but as long as they find someone who [24:30] they think can get the job done somehow, [24:33] they'll probably get that person [24:35] Obviously as a far-sighted ambitious [24:39] organization you don't want to have you [24:41] don't want to hire the first person [24:42] who's adequate. You want to hire someone [24:44] who can be amazing over time. So first [24:46] uh bad incentive. The second problem and [24:48] it's connected to that is if you are [24:51] running the team probably most people [24:54] although they would be willing to coach [24:57] if they it comes down to it they would [25:00] much rather hire someone who's already [25:01] fully mature and competent. So they [25:03] again they can either do other stuff or [25:06] work class. If you leave it to a hiring [25:08] manager to decide, they'll favor very [25:11] experienced candidates over green [25:13] immature but potentially much over time [25:15] much better contributors. [25:18] >> I like that you identified the incentive [25:19] misalignment right that you find in [25:21] typical companies. What's the incentive [25:22] structure for your hiring managers in [25:24] your company then? [25:25] >> Well, there's none. Just trust. They [25:27] don't have any any any bonuses, any [25:29] variable pay. We just tell them we trust [25:32] them to build the best organizations [25:34] they possibly can and then that's it. [25:37] And we find that if we hire people who [25:38] are intrinsically motivated and who like [25:41] the project and and you work with them [25:43] and you're deserving of their uh [25:46] friendship and and admiration then they [25:48] will do their very best to achieve the [25:50] common goals. In fact, we find that [25:52] setting highly specific concrete um [25:55] objectives to which career progression [25:57] or pay are tied uh almost invariably [26:01] leads to bad outcomes or inferior [26:03] outcomes. Meaning maybe people will [26:06] occasionally try a little bit harder in [26:07] the short term, but then uh there's all [26:09] sorts of uh deviations from what would [26:11] be optimal holistically for the company [26:13] and that's instead optimal for [26:15] specifically checking the boxes of that [26:17] particular uh incentive system you [26:19] created. And so we just tell them we [26:21] trust you to create the best teams you [26:23] can. So hire well part ways. We go for [26:25] we don't need to part ways with a lot of [26:26] people but when it's necessary please do [26:28] that. Uh let's talk if you need help. [26:30] Let's discuss. But uh but ultimately [26:32] it's as simple as that. And by the way [26:34] it's it shouldn't come as a shock. I [26:37] mean most of us I think when we worked [26:39] in projects where we thought we were [26:41] doing incred incredibly well and uh [26:43] everybody was pushing in the same [26:45] direction. How frequently were there [26:47] super mechanistic KPIs with our pay tied [26:50] to it? I I've never seen it. I mean [26:52] generally in startups for example, yes [26:54] there's a broader idea if we do really [26:56] well maybe our equity will be worth more [26:57] but it's highly indirect and ambiguous [27:00] and when and how much people work hard [27:03] and try their best because you know they [27:04] feel a sense of ownership. They like [27:06] working with with one another. They care [27:08] about the project. So we try to recreate [27:10] that that same setup. We give them full [27:13] trust in leading hiring and and by the [27:16] way because the the c this is [27:17] centralized they also have a much bigger [27:20] sample and much better information both [27:22] in terms of uh what's available out [27:23] there and what works and what doesn't. [27:25] Again if you're a hiring manager in a [27:26] team of 10 it's probably bigger than [27:28] most teams. Uh at best you're going to [27:30] hire three people a year. I know I'm [27:32] just making it up. Something like that. [27:33] It's not a huge sample to learn from and [27:35] and you you're not focused on it. So you [27:37] you're not going to wake up in the [27:39] morning thinking how can I be a better [27:41] interviewer. Obviously it's not it's not [27:43] your core problem. Uh for our for our [27:46] centralized talent team that's all they [27:47] do. So you know they there is no [27:49] professional pride other than we're [27:51] we're good at this. They do it at scale. [27:53] So they have massive uh massive sample [27:55] size and they get to see what kind of [27:57] talent we can attract across all [28:00] different uh let's say roles and [28:02] positions. Therefore, they're much [28:04] better positioned to understand whether [28:06] someone is the right hire for a [28:08] particular role because they they've [28:10] seen what's coming in time across the [28:12] board. And so, they know they're better [28:14] positioned to know, okay, if we wait a [28:15] little bit longer, the statistically [28:16] speaking, we're likely or or are we not [28:19] likely to find someone who can be even [28:21] better. So they they're they have all [28:22] sorts of advantages in terms of uh their [28:24] focus their you know the [28:26] informationational [28:28] sample uh that supports their decision [28:31] making uh and also this efficiency that [28:33] they they are basically yes they're [28:34] hiring for a particular role but nothing [28:37] prevents them from from picking from [28:38] other pipelines potentially and you know [28:40] swapping uh as as needed. Again maybe [28:43] someone applied as a product manager but [28:45] they see that they could actually be [28:47] amazing as a growth manager. they they [28:48] can easily make the swap because they [28:50] are looking at the entire thing, not [28:52] just that particular. Uh [28:53] >> I really love your insight. It was like, [28:55] well, if you're you're running the team [28:56] and you feel the pain, you might just [28:59] take the first candidate comes along. [29:01] But your whole thing is like, we know [29:02] our strategy works, so now we're just [29:03] going to spend all of our time on [29:04] talent. You're the conclusion you just [29:06] shared here reminds me of Brad Jacobs, [29:08] who was on this show last year. He says [29:10] he has a great maxim where he's like, [29:12] uh, an empty seat is less damaging than [29:14] a poor fit. [29:15] >> Oh, yeah. He's just like, I'll leave the [29:16] position empty. Like, it's going to be [29:18] painful, but it's going to be not it's [29:20] going to be way worse than if we hire [29:21] the wrong person. And he'll just leave [29:22] it in indefinitely till they find the [29:24] right person. It's very similar to what [29:25] you're saying. [29:25] >> Yeah, completely. And look, I I I think [29:27] in general having sharp job descriptions [29:30] is bad. You want to have a there is a a [29:33] blob of work that needs to be done and [29:35] different things are differently [29:37] important, urgent. And if you have a [29:39] team where people don't feel siloed, [29:42] they're just responsible for for the [29:44] company success. Again, just like a [29:45] startup, uh if you're failing to hire [29:48] someone who's supposed to take care of [29:49] like a little part of this blob of [29:51] potential work, it's [snorts] not that [29:53] that blob is ignored if it's really [29:55] important. Someone will basically [29:58] postpone something that's a little bit [29:59] less important to take care of it, [30:00] right? And so I always say I generally [30:02] talk to new hires. Uh we have sessions [30:06] where we discuss some of our cultural [30:08] principles and other things. And one of [30:09] the things I sometimes say is that we [30:11] all have the same job at Benish Blooms. [30:13] All of us starting with me and that's [30:16] helping the company succeed on a daily [30:18] basis. It's helpful to say you're a [30:19] software engineer, I'm a product manager [30:21] just to so we don't step on each other's [30:22] toes too much, but essentially [30:24] everybody's job is the same. Do whatever [30:26] is needed to help the company succeed. [30:28] And so I'm not worried about a seat [30:29] being empty because I don't think the [30:31] concept of seat even exists really. Uh [30:34] we'll just adapt and take over and and [30:36] complete the work that needs completion [30:38] and you know we'll just uh not do some [30:39] other work at the end of the day. Very [30:42] little work in a company, especially a [30:45] digital business, is strictly [30:47] necessary. Almost all of it is elective, [30:50] optional. It's just a matter of what's [30:52] higher priority and lower priority. [30:54] >> Say more about this. [30:55] >> Well, I mean it almost everything you [30:56] do, you could also not be doing almost [30:59] all of it. Uh and so winning starts with [31:02] doing what's ROI positive, which is only [31:05] a small portion actually of the complete [31:08] universe of possible projects and tasks. [31:10] and then [31:12] doing things in order of priority. So [31:14] from say highest ROI again the risk of [31:16] being a little bit simplistic and and [31:18] your resources will be limited. I think [31:20] most companies do things that are ROI [31:22] negative. You know there is 100 things [31:24] they could be doing but only 10 are ROI [31:26] positive. Many companies are doing 40 [31:28] things. Hopefully at least they do the [31:30] 10 that are ROI positive too. In some [31:32] cases tragically they're not doing some [31:34] of the ROI positive things despite doing [31:36] so many other things. [31:37] >> Wait, so why why do you think they're [31:38] doing this? Is this a a lack of talent, [31:40] an issue of focus, not understanding [31:42] prioritization? Like what's going on [31:43] there? [31:44] >> Oh, I mean all sorts of reasons for the [31:46] companies you you buy cuz obviously [31:47] you're buying things that are there's a [31:49] brand that's wellnown, there's a [31:51] customer, there's a product there, but [31:53] in almost I think every single example, [31:54] you've massively improved everything [31:56] you've purchased. So what are like the [31:58] most common mistakes that people [32:00] previously under previous management [32:02] were making? A lot of the the reasons [32:04] for those opportunities not being seized [32:07] uh uh frankly lay outside of their [32:10] control. Some of it is perverse [32:12] incentives. If you're running a business [32:14] on a standalone basis, [32:16] especially if you're a public company, [32:18] but also private companies ultimately [32:19] they're aiming to to go public. So it's [32:21] kind of the same. You'll be judged on [32:24] what I often you'll be judged on what I [32:26] would uh let's say consider [32:29] ultimately secondary [32:32] if not even vanity metrics uh rather [32:35] than uh let's say uh value creation [32:38] through through cash generation. [32:40] [snorts] For instance, if you are [32:42] running um a business where most of the [32:44] revenue comes from subscriptions [32:46] um and you you you know maybe that uh [32:50] the optimal price is a higher price like [32:53] in every pretty much with any product in [32:56] a free market if you raise prices you're [32:58] going to have fewer customers which can [33:00] be fine. You know, maybe you have 30% [33:02] fewer customers, but each ultimately [33:04] contributes twice as much. You're better [33:06] off, right? However, often the markets [33:08] will will punish you h dramatically if [33:11] you do that because when they see that [33:12] the number of subscribers has gone down, [33:14] even if revenue has gone up, they will [33:17] not like it. Uh and we could debate why [33:19] that's the case. It's an interesting [33:20] discussion. But if you're a management [33:22] team ultimately in that particular [33:24] context you will have to heed the [33:27] opinion or expectations of or of the [33:29] market and you will not do that uh [33:31] pricing change even if you know that [33:32] it's going to be positive. However, if a [33:35] company, a business is run within the [33:36] broader bending bones where uh none of [33:39] the businesses uh let's say ends with [33:41] itself, but it's a piece of a broader [33:43] puzzle, a a a source of cash for further [33:46] deployment and growth, then it's much [33:49] easier to make those otherwise unpopular [33:51] decisions and even investors would [33:53] potentially support them because they're [33:54] not focused any longer on I want say [33:57] Evernote to have as many subscribers as [33:58] possible. Yeah, all all has been equal. [34:00] I want to have more subscribers, but I [34:01] would rather have an Evernote that [34:02] generates more revenue, more cash flow, [34:04] so that it's it's more accreditive to [34:06] the to the bigger benefits and we can go [34:08] after bigger acquisitions and thrive. So [34:10] there there are incentives and this is [34:12] one example. Um, another one is talent. [34:15] Sometimes businesses when they have [34:18] matured and you know everybody [34:20] understands and sees their somewhat [34:22] they've somewhat saturated their [34:23] opportunity, you know, maybe they're [34:25] growing 15%, maybe they're flat but [34:27] they're not doubling every year or [34:28] something. often they have long stopped [34:30] attracting some of the most hungry uh [34:33] ambitious talent. And so these executive [34:36] teams um have access to perfectly valid [34:39] talent, but maybe not against standout [34:41] talent. And wait a minute. So I just I [34:43] think it just clicked on one of the the [34:45] like unexpected benefits of what you're [34:47] doing. It's like you buy AOL and I'm [34:49] working on AOL. I don't think I'm [34:52] working for AOL. I think I'm working for [34:53] Benny Spoons. [34:54] >> Exactly. Okay. So I you know I I don't [34:59] consulting like um you know the big [35:01] strategy consulting companies McKenzie [35:03] BCG bank. [35:03] >> Did you have a you got hired there [35:05] right? [35:05] >> I got hired I got hired because we had [35:07] we so in parallel with the startup we [35:09] were talking about. [35:10] >> So I have a background [35:11] >> tell tell the story. We're going to go [35:12] back on betting soon. This is a [35:13] hilarious story dude. [35:15] >> So I have a background in engineering [35:16] physics and um with two friends of mine [35:20] also engineers. We had this idea of [35:22] building that company that AI [35:23] self-writing diary I was describing [35:25] earlier uh Everale but we had no money [35:28] you know all of us you know coming from [35:30] countryside in in the northeast of Italy [35:32] >> you come from like a town of what 900 [35:35] people or something like that [35:35] >> at least at the time yeah fewer than a [35:37] thousand [35:37] >> I don't think anybody in your family [35:38] went to college I think your parents cut [35:40] hair right okay [35:42] >> they are retired now but yes they they [35:44] are they used to uh to to [35:46] >> maybe some of those billions you got in [35:47] your pockets are helping them retire [35:49] >> [laughter] [35:51] >> Uh well it's all virtual. I haven't sold [35:53] a whole lot of stuff. Um [35:55] >> so um so anyway we we wanted to to build [35:59] that that that startup but we had no [36:01] money and uh [36:03] it wasn't at least we thought it would [36:05] wouldn't be easy to raise seed capital [36:07] either and maybe it's easier or was was [36:09] easier and certainly is easier in the [36:11] states. it wasn't for us. And so we [36:13] figured how do we do this? And so the [36:15] three of us very good friends in the [36:17] time and even more so today because [36:18] we've gone through so much over the [36:20] following 15 plus years. We we figured [36:23] okay we all all of us look for a job and [36:25] whoever gets the most uh lucrative offer [36:28] accepts it and pays for rent and food [36:30] for the other two. The other two would [36:32] work on a prototype and basically the [36:34] startup until we can convince someone to [36:35] give us some money so the whoever is [36:37] working can quit and we can all focus on [36:39] the start. we all look for a job and [36:41] well frankly one of us uh was uh doing a [36:44] PhD uh already so that was uh you know [36:46] our uh backup plan but not a super [36:49] lucrative job so that you know we were [36:51] hoping to do better than that and I I [36:53] happened to get a an offer from McKenzie [36:56] for a consulting job as strategy [36:58] consultant and so that was the best uh [37:00] offer we got and I remember I was [37:02] terrified because I I'm in I'm close to [37:05] incapable of lying or being opaque I I [37:10] always want to be honest and [37:11] transparent. But that's why I decided I [37:13] would tell the partner from Mckins who [37:15] extended an offer to me that uh yes, I [37:18] was going to work there if they wanted [37:20] me uh give it my 100% but the plan was [37:24] as soon as possible for me to quit to go [37:27] do the startup. Um and I was so [37:30] convinced that they would uh withdraw [37:31] the offer, you know, cuz who wants to [37:34] hire someone who uh was kind of as a [37:37] who's who's not planning to [37:38] >> be here for a little bit. I'm trying to [37:39] leave as fast as possible. [37:40] >> Exactly. Incredibly, that partner uh was [37:44] enthusiastic about the project and said, [37:46] "Yeah, it's great. You know, well, we [37:49] want to have you here." [37:52] So, very grateful um very inspired [37:56] worked very hard, as hard as I could. uh [37:59] was working on the on the startup in the [38:02] you know basically during the night but [38:04] not like when people say during the [38:06] night they mean from 7:00 p.m. to 9:00 [38:08] p.m. I mean like from midnight to 2 in [38:09] the morning and then uh on the weekends [38:12] and uh and then I remember after the one [38:15] year at Mckenzie I had my I don't know 3 [38:17] weeks of vacation or something like that [38:18] I I spent to work like full-time on the [38:20] start. Anyway about a year later we [38:23] managed to to raise uh uh about half a [38:26] million dollars and so I quit. I mean I [38:28] finished a project another two three [38:30] months and then I I I left. So yes, [38:33] that's my stint in consulting. Um, and I [38:36] and I think I saw something there that I [38:38] I think has some similarities to Benish [38:40] most of the time. Most of the people who [38:42] applied to work at Mckenzie and I'm [38:44] pretty sure it was the same for BCG [38:45] being these other consultancies were [38:48] very excited to to be working there. Uh, [38:51] they got some of the best graduates at [38:53] least from business, maybe not as much [38:55] from engineering. But then you would end [38:57] up serving telos, banks, insurance [39:01] companies to which you would never have [39:03] sent your CV. And I think Benny Spoons [39:05] is kind of similar for software [39:07] engineers, product designers, product [39:09] managers. I believe we got some [39:11] exceptionally good talent, especially [39:13] students, new graduates [39:16] for reasons we can discuss, but talent [39:17] density is one of them. Career [39:18] opportunities and then [snorts] end up [39:20] working on uh AOL or you know again [39:24] Evernote. Some of these businesses I [39:27] they they these people would then have [39:28] applied if the the the whole prospect [39:32] have been to to work there for for five [39:34] years or 10 years but they're incredibly [39:36] excited and rightfully so to spend say [39:38] 12 months or 18 months on AOL rebuilding [39:40] the technological foundation rethinking [39:42] the you know customer experience [39:43] monetization that's a very interesting [39:45] challenge because you get to change a [39:46] lot of stuff on a very large user and [39:49] customer base. So you you you get the [39:51] the best of a startup and a big corp. [39:54] From the big corp, you get we're working [39:56] on big user and customer bases. This is [39:58] not we're not trying to find product [40:00] market fit. We have a lot of resources. [40:01] But from a startup, you have a tiny [40:03] team, lots of responsibility, and we're [40:05] actually making big changes. So it's not [40:06] we're not refining a button or trying to [40:09] add the next 0.5% in revenue. We're [40:11] trying to to rebuild almost from the [40:13] ground up in many cases. And so going [40:15] back to what are some of these executive [40:17] teams from the acquire companies getting [40:19] wrong? Well, they're actually often [40:21] doing well, but they're they can only [40:23] work with the teams that realistically [40:24] they can attract. And we are often able [40:27] to bring in a lot of fresh talent with [40:30] the new perspectives and some excellent [40:32] skills. And so, it's a lot easier to [40:34] rethink and rebuild these companies when [40:36] you have access to this to this talent [40:38] pool. [40:38] >> One of my biggest partners is RAMP and [40:40] I'm really close with the the founders [40:42] there. And I was happened to be with [40:43] them the night [40:44] >> Eric [40:45] >> Eric and Kareem. But I was with Kareem [40:47] the night that uh one of their main [40:49] competitors who they didn't even view as [40:51] a competitor anymore, but everybody else [40:52] said Brax got acquired. And Kareem said [40:55] something interesting. He's like, "Oh, [40:56] uh, you know, I was like, "Oh, how do [40:58] you think about this?" He's just like, [40:59] "Well, you know, there people thought [41:01] there was like a war between Ram and [41:03] Brex. He didn't." And he's like, "Well, [41:05] if there was a war, it's definitely over [41:06] now." And I go, "Why?" He goes, "Best [41:08] talent's not going to go to Virginia and [41:09] go work for Capital One." He's like, [41:10] "We're going to keep getting better [41:11] talent." And even if that like in that [41:14] gap between, you know, the talent that [41:15] we have and the talent that they're [41:16] going to attract there, it's like it's [41:18] like you play it out year one, year two, [41:20] year, five years from now. It's like [41:22] it's over. It's all about people. It's [41:24] very similar [snorts] to what you're [41:25] saying. [41:25] >> Yeah. I don't know their industry well, [41:27] but it sounds possible. I I tend to [41:28] agree with that. [41:29] >> Deal is how the best founders turn the [41:31] world into their talent pool. I've been [41:34] studying how history's greatest founders [41:35] operate for a decade. [music] And one [41:37] thing they all have in common is they [41:38] understand that recruiting and hiring [41:40] the very best talent is your most [41:42] important priority. A players recognize [41:44] other A players, which is why top [41:45] companies like Ramp, Shopify, 11 Labs, [41:48] Uber, and Door Dash all use Deal. Many [41:50] of the top founders I know have [41:52] personally invested in Deal after using [41:54] their product. [music] And what they [41:55] discovered is that Deal is the best [41:57] company in the world at building [41:58] infrastructure for global hiring. deal [42:00] will help your business hire, pay, and [42:02] manage any [music] worker anywhere in [42:03] the world so you can retain the best [42:05] talent anywhere and spend the rest of [42:07] your time focusing [music] on what you [42:08] do best, delivering value to your [42:10] customers. The founder of 11 Labs has a [42:12] great description of the value deal can [42:14] [music] give your company. He said, "We [42:16] built 11 Labs to break down language and [42:18] communication barriers. [music] With [42:20] deal enabling us to hire and support [42:22] exceptional talent anywhere, we can [42:24] accelerate our innovation and bring more [42:26] [music] voices, stories, and ideas to [42:28] every corner of the world. Deal is [42:31] trusted by over 40,000 businesses. Learn [42:33] how they can help your business today by [42:35] going to deal.com/enra. [42:38] That is deal.com/enra. [42:41] I want to go back to this idea of [42:43] saturating their capacity because it's [42:44] still one of the most interesting things [42:46] you told me when we had lunch. So like [42:48] can you give an example? Okay, you've [42:50] you talked about, you know, it's better [42:51] to have no habits than bad habits. So, [42:54] I'm going to find graduates or in some [42:56] cases people that haven't even graduated [42:58] yet. They might start as like an intern [42:59] or very entry level at Betting Spoons [43:01] and then you're like, "Oh, we identified [43:02] this talented person." And what do you [43:04] mean by saturating their capacity? Like [43:05] give give like concrete examples of how [43:07] you've done this. [43:07] >> Yeah. Everybody at the company certainly [43:09] the people who have shown promise, they [43:12] should have way more on their plate that [43:14] than feels even remotely comfortable. [43:16] And the reason why you should do that is [43:18] is manifold. The first reason is every [43:22] time you choose what to work on, whether [43:23] you do it consciously or unconsciously, [43:25] you're prioritizing a set of work items, [43:28] each with its own return attached. [43:30] Again, you may be unaware of potential [43:32] returns or very deliberate. But either [43:34] way, that's what's happening. uh the [43:37] bigger the universe of uh let's say work [43:41] items that you can prioritize from [43:44] mathematically the higher the returns on [43:46] your time you'll deliver let's say you [43:48] have 10 possible tasks if I add and each [43:51] with a certain let's say ROI attached if [43:54] I add the 11th it's it's impossible [43:57] assuming that you select well it's [44:00] impossible that adding an 11 task will [44:03] lower the ROI of what you choose to do [44:06] cuz you're you still have the other 10. [44:08] So if this is lower ROI than the others, [44:10] you're still going to do the others. But [44:12] it's it's possible that it happens to be [44:14] the highest ROI of all and so you end up [44:16] doing something more valuable. So the [44:17] more work you give people, the better [44:20] the opportunity for them to to create [44:22] value. Now that's especially true if [44:25] they choose well and therefore it's very [44:28] important to work with people who are [44:29] smart and it's very important for for [44:30] for managers for leads probably the [44:33] number one thing they can do or [44:35] certainly you know one of the most [44:36] important things you can do and we try [44:37] to coach them in this regard is to help [44:39] the report select well. So that's the [44:41] most important thing. The other uh very [44:44] important thing uh in that that you [44:46] accomplished when you give people a lot [44:48] more work than uh than than feels [44:51] comfortable is [44:53] you're really forcing them to [44:57] uh come to terms with the immensity of [45:01] of the of the possible. We find that [45:04] sometimes people as they as they grow [45:06] and uh and the and their the aperture [45:09] for them professionally expands they get [45:12] overwhelmed or there's too much to do. [45:15] Uh we need more people for example on [45:17] the team and uh and I think that's uh [45:20] generally a a terrible way of looking at [45:21] at life or the world. Generally speaking [45:25] there's always a lot more you could be [45:26] doing than uh uh than you can do in [45:29] terms of your capacity. It's just that [45:32] some people don't don't realize it. It's [45:34] not that if you're a student and you [45:36] you're done studying for an exam, [45:38] there's nothing else you can do. There's [45:39] plenty you could do. You could launch a [45:40] startup. You could uh take a second [45:42] degree on the side. Uh it's just that [45:45] you may not be sufficiently proactive [45:47] and imaginative to to figure it out. And [45:49] so as people jump in a job and if you [45:52] give them just you know relatively short [45:54] task list so that they they'll be done [45:56] within their eight hours and there's [45:58] nothing in their on their mind. Uh that [46:00] initially feels easy and and comfortable [46:03] but you're failing to train them at a a [46:05] massively important skill which is [46:07] handling that immensity of the of the [46:10] possible. And once you become really [46:12] good at at being comfortable with having [46:14] 100 times more things you could be doing [46:15] than you can actually do. That's an [46:17] insane superpower to have because it [46:19] enables you again to to handle a vast [46:22] array of possibilities and and select [46:25] surgically select those with insanely [46:27] high returns. And it's something you can [46:29] only do if you're not uh thoroughly [46:30] overwhelmed. So, is it better to have [46:32] someone become overwhelmed by that [46:34] immensity once you know they're 10 years [46:36] in and they're running a 100% [46:39] organization and a billion dollars in [46:40] revenue or is it better to test them at [46:42] that and coach them at that I wouldn't [46:44] say on day one but maybe day seven and [46:46] for the first year so that first of all [46:48] you only promote to that that higher [46:50] level of responsibility people who have [46:52] proven that they can do that and if they [46:54] actually can do that they begin [46:56] benefiting benefiting from it much [46:58] earlier even if their their their scope [47:00] is is more limited. But you need to [47:02] completely eradicate the concept of I'm [47:04] only good at my job if I uh if I exhaust [47:08] my checklist, my my task list. There's [47:10] no such thing. You're always going to [47:11] have at least advantage spoons. But I [47:13] think again life if you look through the [47:15] veil is like that anyway. You're just [47:18] unconscious about this. times. You [47:20] really want to be able to handle that [47:22] enormous amount of possibilities and and [47:24] surgically identify those that uh have [47:27] insanely attractive returns and then be [47:29] laser focused on those and disregard [47:31] everything else. Is this related to what [47:33] you're saying earlier where you're like, [47:35] listen, they're not going to be AOL's if [47:37] it was a standalone brand, they're not [47:38] going to get the talent that we are [47:39] going to get at Benning Spoons. Uh we [47:42] can have a massive impact because they [47:43] have a huge customer base, but then we [47:44] can treat it like a startup. But then [47:45] you said something about like if they [47:46] can work on this for 10 to or 12 to 18 [47:49] months. So then you rotate teams [47:51] throughout the different companies. Is [47:52] that part of saturating their capacity? [47:53] It was like okay this opportunity on [47:55] this business this person is really [47:56] talented but there's no other ways to [47:58] utilize that talent to a higher degree [47:59] here. So let's move them to another [48:00] team. Am I understanding that correctly [48:02] or [48:02] >> no no that you you do there there yes we [48:05] do rotate people all the time. There are [48:07] very reasons for that. Part is uh I [48:09] think at some point when you have looked [48:10] at the same thing for a long time you [48:12] stop having good ideas. So it's good to [48:14] get you know new people in to to maybe [48:16] take a fresh look. Uh part of it is we [48:19] just uh uh find that if people keep [48:22] working with the same people you risk uh [48:24] developing subcultures and we're highly [48:27] opinionated on what the optimal c [48:29] culture looks like. We want it to be [48:30] uniform across the company and and so [48:32] and if someone comes up with a better [48:34] idea that's awesome but that has to be [48:36] spread across the company. We don't want [48:37] to have subcultures and so you want to [48:39] move people so and you know mix and [48:41] match and so so that they they they [48:42] don't get used to a different way of [48:44] working uh at least on the important [48:46] aspects. Another one is that they get to [48:48] learn more. So that that goes back to [48:50] what you were discussing. Uh it's [48:52] slightly different from having you know [48:54] you know an immense set of possible [48:56] tasks. Uh there is an element of of of [48:59] diversity there. You need new challenges [49:01] and diverse challenges to keep honing [49:02] your craft and and finding new ways of [49:04] growing. So that helps too and it keeps [49:06] also enthusiasm levels higher because uh [49:08] you know humans tend to get bored uh and [49:11] so we want to try new things to stay [49:14] motivated. And last but not least, as we [49:16] keep acquiring new things, um, as an [49:20] organization, you know, the the universe [49:22] of of the things we could be doing [49:23] expands with new acquisitions and, uh, [49:26] often those in working on those new [49:29] things yields the highest expected [49:31] returns. And so we regret, you know, [49:33] regrettably, we have to remove resources [49:36] from from businesses that would still [49:37] have, you know, plenty of opportunity in [49:39] them. But, you know, relatively [49:40] speaking, it's better to to work on a [49:42] new on the new on a new business. So for [49:44] all these reasons we we do rotate people [49:46] all the time and I think it's been quite [49:48] uh quite successful for us to do it that [49:49] way. [49:50] >> You just said you're you have uh very [49:52] strong and you're highly opinionated on [49:54] the culture they should have. Do you [49:55] want to share some of those opinions? [49:56] >> The main quality we look for in people [49:58] we call it extreme ownership. We [50:00] [snorts] try to work with people who [50:02] care tremendously about being the best [50:05] in the world of what they do about [50:06] bringing value to the team. [50:08] >> Hold on. Did you get that from Jaco's [50:10] book? [50:10] >> Actually the the the name Yes. the [50:13] concept is not exactly the same. There [50:15] are similarities, but I thought the the [50:17] terminology extreme ownership was so [50:20] immediately evocative of what you look [50:23] for that I said, "Okay, we need to use [50:24] that for sure." [50:25] >> Obviously, you know, I only read [50:26] biographies and in history, but I don't [50:29] read business books, but I I always tell [50:30] people like that's one of the few [50:31] business books I'd actually recommend [50:33] reading. You can read on a weekend and [50:34] yeah, it's very direct. It's just like [50:36] Jaco is in I've met him in real life. [50:37] It's the exact same person. [50:39] >> Yes. [50:40] >> So, extreme ownership. [50:41] >> Yeah. Extreme ownership. And we I think [50:42] we define it a little bit differently. [50:44] But the [50:44] >> So how do you define it then? [50:45] >> Extreme ownership is is caring [50:48] in your belly tremendously about being [50:50] the best at what you do about helping [50:52] the team in the company succeed. Uh it's [50:54] a matter of priority and intense is a [50:56] priority. Um and we want to work with [50:58] people who feel that way about their [51:00] work at vantage points. We'd rather not [51:02] work with someone who's really really [51:04] smart, very competent, but for whom [51:06] doing well here would only be priority [51:08] number three or four. You know, like uh [51:12] we have seen it time and again we've had [51:13] people who were probably close to genius [51:16] level IQ fail here because ultimately [51:18] they saw their job as a way to make uh [51:21] you know to earn a living to to make [51:23] ends meet rather than actually [51:25] transcending apparent limitations and uh [51:27] and and winning and and being amazing at [51:29] what [51:30] >> Okay, hold on. You just got done saying, [51:31] "Hey, we're going to centralize hiring. [51:32] We have a bunch of engineers. You need [51:33] to have brilliant people in HR. We went [51:35] through this like very unique way that [51:36] you think about this, right?" But how do [51:39] you screen for that? How do you screen [51:41] for being successful here and helping [51:44] this company be successful? Is that [51:46] their top one of their top priorities? [51:47] You know, maybe the top priority in [51:49] their life. [51:49] >> Yeah. I mean it's never going to be I [51:50] mean we understand u obviously if you [51:53] have a family that will be number one [51:54] but if you start telling me after my [51:56] family then there's uh you know being a [51:59] great gamer at night plus [52:01] >> people are going to apply for a job and [52:02] say that so like what what are the [52:04] actual things that you're [52:05] >> Yeah. So I think there's first of all [52:06] you try to I mean I don't want to give [52:09] too much away but let's say [52:10] >> if it's a proprietary dog [52:13] >> no no but I would say the first of all [52:15] you want to see if there is a capacity [52:17] to to express extreme ownership. There's [52:18] a bunch of pe well I suppose every human [52:20] being in theory has it but I find there [52:23] a lot of people are they don't seem to [52:25] be at least not inclined to develop in [52:28] extreme ownership or for almost [52:29] anything. So they just struggle to care [52:31] tremendously about things in life and uh [52:33] there's no moral judgment but I'm just [52:35] saying I want to be a part of a team [52:36] that has a real chance of redefining [52:38] what's possible and and succeeding at at [52:40] really high level. Of course that type [52:42] of profile is not going to be highly [52:44] appealing. I don't think I'm saying [52:46] anything shocking here. So you look for [52:47] signs in someone's past of of that [52:51] extreme ownership at work. Uh maybe [52:54] they, you know, they they were fully [52:56] focused on their side is okay. Did they [52:58] do incredibly well at least, you know, [53:00] uh maybe they um they did a lot of work [53:05] next to studying because maybe they [53:07] didn't have the financial means or they [53:08] wanted to learn a craft. Maybe they were [53:10] into open source. Did they is is their [53:14] contribution extremely you know this [53:17] small or is there something just in [53:19] terms of it looks like they put in a lot [53:21] a lot of effort maybe they didn't have a [53:23] breakthrough but you can tell through [53:24] the sheer volume of contributions that [53:26] they really care. Did they launch a [53:28] startup? Was it uh because it's cool for [53:32] months it didn't work out too bad or [53:35] they ground at it for 3 years and it was [53:37] incredibly unsuccessful but you can tell [53:39] they wouldn't let up. you know, [53:41] something that shows they are capable of [53:42] of of putting their passion into [53:44] something. [53:44] >> So, in your S1, I think you referenced [53:46] Singleton, Henry Singleton and Tom [53:49] Murphy. I just read I did another [53:51] episode on my other podcast founders on [53:52] Singleton. And what was remarkable, [53:54] Singleton made a very early investment [53:56] in Apple, right? And then he ended up [53:58] joining the board and he was asked by [54:01] his partner, he's like, "Well, there's a [54:03] million companies, not a million, but [54:04] there's a bunch of companies trying to [54:05] make the personal computer. Like, why [54:07] did you choose Apple? like how you pick [54:09] the best one of the bunch and there's a [54:12] bunch of them and he said two things one [54:13] he thought that people were going to be [54:14] intimidated because they never dealt [54:16] with there was no such thing as a [54:17] personal computer these things could be [54:19] intimidated less likely to be [54:20] intimidated by a computer name uh [54:22] computer called Apple but more important [54:24] than that he's like the the founders of [54:27] Apple had he goes there's a lot of these [54:29] founders that I met of other computer [54:31] companies that they wanted to start a [54:33] computer company but if it didn't work [54:34] out they'd be okay he goes the founders [54:36] of Apple had to make it work they had [54:38] nothing else. There was no way that [54:40] Steve Jobs was going to give up. And the [54:41] idea that Singleton being the genius [54:43] that he was identified that in a 19 or [54:46] 20-year-old Steve Jobs is incredible. [54:49] >> I think sometimes of course one of the [54:51] greatest entrepreneurs, right, to ever [54:53] do it. And I think a lot of people focus [54:55] on the eye for detail. Certainly had [54:58] that um the perception of what consumers [55:02] would want. Certainly good at it. I [55:04] think maybe [55:06] that wasn't even like his main thing, [55:08] but I believe what we would call extreme [55:10] ownership in his case just [55:13] we'll probably call it differently, but [55:14] I think deep down it would be the same [55:16] thing. just he cared so badly about [55:20] seeing Apple succeed uh the way he [55:23] thought it should like by building those [55:25] amazing products and and when you want [55:27] something so badly uh you're not [55:29] guaranteed to win obviously but it just [55:31] sets you apart [55:32] >> cuz there's a million thousand different [55:34] little decisions you're going to have to [55:35] make and you're just going to pay [55:37] attention and care more about every [55:38] single little decision. I I've mentioned [55:40] this quote so many times on on the these [55:42] conversations we get to have with these [55:43] founders on the show, but um I think [55:44] Josh Kushner just it's one of my [55:46] favorite quotes I've ever heard. And [55:47] Josh's point was just like if you have [55:49] to pick the person that is the smartest [55:51] or the person that has the most [55:52] experience or the person that wants it [55:53] more, you always pick the person that [55:55] wants it more. [55:56] >> Oh yeah. [55:56] >> I think that's kind of what you're [55:57] getting with extreme ownership [55:58] >> completely. Look, we what we found is [56:00] that there is a level of uh let's call [56:02] it intelligence broadly speaking not [56:05] just purely let's say logical analytical [56:08] that's uh necessary in at least in our [56:10] endeavors. So but after you pass certain [56:13] threshold which is a you know admittedly [56:15] a fairly high threshold but we're not [56:16] talking about again genius level just a [56:19] then it's almost all badly you want it [56:21] like you you really want to be amazing. [56:23] I was talking about uh this with again [56:25] some new joiners uh yesterday or the day [56:28] before. I brought up brought up the [56:29] example of Rafa Nadal in tennis. I think [56:31] I've said this before. Most experts I've [56:34] talked to believe he wasn't even [56:36] probably a top 50 talent in his [56:39] generation, but went down as one of the [56:42] three best for sure. Some say the best [56:43] or second best to ever do it. And where [56:46] he really stood out was that extreme [56:47] ownership. He just woke up in the [56:49] morning and he was like, I'm going to be [56:50] the best tennis player I can possibly [56:52] be. I'll give it my all 100%. And [56:55] >> did you read his autobiography? It's [56:56] called Rafa. [56:57] >> I haven't. [56:58] >> You should. I think you'd be interested [56:59] cuz like a lot of people don't know the [57:01] amount of injuries that he had when he [57:03] was younger. He shouldn't have even been [57:04] able to play at all, much less be one of [57:06] the best to ever do it. [57:07] >> Unbelievable. Absolutely. Like I what I [57:10] said to those new joiners, I told them, [57:12] I don't believe there is almost any [57:15] chance you will fail to have an amazing [57:17] career at least at Bendy Spoons. [57:19] Probably almost anywhere. Certainly [57:20] business where we try to be extremely [57:22] meritocratic if you really bring it like [57:25] if you are an extreme owner I would bet [57:27] there's less than 1% chance you failed [57:28] to have an amazing career because we [57:30] know you're smart we tested that we're [57:32] unlikely unless you cheated somehow [57:34] unlikely to be wrong uh you have the you [57:36] know you studied what you needed to [57:38] study so you have some of the [57:39] foundations it's almost all about do you [57:42] come to work to be amazing to be better [57:45] today than you were yesterday to see [57:46] your team be better to see the company [57:48] take a step in the right direction or [57:50] you come to work basically waiting for [57:52] the day to be to be over. You know, like [57:54] of course you're going to try to be okay [57:55] at it, but you don't really care too [57:57] much as long as you're you have a job [57:59] and you're in your mind. And if you're [58:01] part of the former group, you'll you'll [58:03] do extremely well for sure. So that's, [58:05] you know, really a a key cultural tenant [58:07] for us. We we we select for it. We try [58:09] to to foster it. We will much rather [58:12] have a smaller team of people feel that [58:13] way than than vice versa. And by the [58:15] way, it's contagious. So if you have if [58:17] you have a high density of people who [58:19] feel like that highly accountable, [58:21] proactive because if you're an extreme [58:22] owner, so you care tremendously. You're [58:24] going to be entrepreneurial because [58:25] you'll be paranoid about things that [58:27] could go wrong and and enthusiastic [58:28] about new ideas and how can you can [58:29] improve things. If I ask you to do [58:31] something, you will not forget you get [58:33] it done more so you'll even come back [58:36] and do more things than I expected you [58:37] to and that will be incredibly exciting [58:40] for me and I don't want to show show you [58:42] that I can be just as good. So it's it's [58:44] just there [snorts] is an escalation of [58:46] positive reinforcement that you [58:48] accomplish if there is a high density of [58:49] that feeling within the team and as soon [58:51] as it you dilute that the people who [58:53] feel that uh either leave or lose it you [58:57] can't have an extreme like Steve Jobs [58:59] famously was was looking for a players [59:01] and I I think he was looking for people [59:03] who had that desire that drive to be to [59:06] to do something amazing even even before [59:08] looking for people who are brilliant. [59:10] Obviously you want to have both if you [59:11] can but uh you can never have a think I [59:14] think a high performance team where more [59:16] than a small fraction of people lack [59:18] extreme ownership. So that's really [59:20] number one and then there's you know [59:22] many other things that are important but [59:24] but secondary to to this what are some [59:26] of the things al that are also [59:28] important. Yeah, one thing that we we [59:29] call relentless simplification. [59:32] Um we believe that [59:36] most most things don't matter. Uh most [59:39] things do more harm than good. However, [59:42] humans have a tendency to to add the [59:46] complexity do things and do those things [59:48] that destroy value. So if you leave an [59:51] organization almost any environment uh [59:54] uh let's say unattended and you don't [59:57] provide guidance in this regard it will [59:59] tend to to become more complicated [60:01] people will be adding parts and when I [60:03] say parts I mean it could be expanding a [60:05] team it could be adding a step to a [60:07] process adding an entire new process if [60:09] it's a product adding a feature to the [60:11] product new rules uh it really you know [60:14] I'm making a general point but it [60:15] applies to almost any human endeavor [60:18] people will tend to add pieces very [60:19] rarely remove pieces. And with every [60:21] piece you're that you're adding to this [60:24] ensemble, this this system, you're not [60:27] adding complexity linearly because [60:30] you're not just adding with the piece, [60:31] you're also adding interdependencies, [60:33] interconnect interconnections with some [60:35] and sometimes all of the other pieces. [60:37] So if you go from three to four pieces, [60:40] you're not only getting, you know, the [60:42] the system is not getting say 33% more [60:45] complicated, it's maybe getting 40% more [60:47] complicated or 50 depending on on again [60:49] the connections and how these new [60:51] connections impact the other [60:52] connections. Most human reorganizations [60:54] if you don't make a conscious effort to [60:57] achieve simplicity. So [61:00] avoiding this this [61:02] this in increasing complexity and [61:04] embodying complexity will will go down [61:07] that that path and that's how we got to [61:08] you know our modern society with all the [61:10] bureaucracy and complicated regulation. [61:12] And a lot of it or almost all of it [61:14] probably when it was introduced it had [61:16] it was meant to be a good thing and [61:17] maybe in a vacuum it was but then people [61:20] failed to account for this the cost of [61:21] these net connections and and frictions. [61:24] And so we have this value or this [61:25] principle whereby we ask everyone who [61:27] works here to first of all every time [61:29] someone is suggesting that we should be [61:30] adding complexity the burden of proof is [61:34] on those making that suggestion. The [61:36] people who support the thesis that we [61:38] shouldn't be adding that complexity [61:41] don't need to prove it. They're done. [61:43] It's just to raise a flag and say I [61:44] don't think we should. So the burden of [61:45] proof is on those who want to add [61:47] complexity which helps reduce uh the [61:50] addition of complexity dramatically and [61:51] that the complexity you add tends to be [61:53] hopefully more often than not good [61:55] complexity because you have to prove it. [61:56] And so hopefully if you're [61:57] intellectually honest that should be a [61:59] good a good idea. Then the other part of [62:01] relentless simplification is that we [62:03] want people to be on the lookout for [62:04] existing complexity and and suggest that [62:06] we should be removing it. Understanding [62:08] how we operate and the biases that [62:11] accompany us through throughout our [62:13] lives is very important. Charlie Mer [62:14] famously studied biases and I think [62:17] knowing your weaknesses or likely [62:18] weaknesses is you know 50% of of of [62:21] avoiding them or or overcoming them. So [62:23] knowing that we as humans tend to be [62:25] disc called this thing called [62:28] consistency bias but also inertia bias I [62:30] mean I've heard it you know slightly [62:32] different things called with with [62:33] slightly different names but essentially [62:35] we tend to assume the status quo is fine [62:38] and we focus on deltas that happen new [62:40] things that are added or changes we we [62:42] stop seeing we become blind to our [62:45] surroundings as they stay the same day [62:47] after day and so we ask our colleagues [62:49] and all of us to make a conscious effort [62:51] to question what's already there and the [62:55] the longer it's been there the more we [62:57] should be questioning it whether it's [62:58] still enough positive so we can look for [63:00] things we can get rid of. [63:01] >> Have you paid attention to how Elon [63:02] talks about this at all? [63:03] >> Maybe maybe not. [63:04] >> Okay. I mean it's one of the things that [63:06] he probably repeats the most. Obviously [63:08] he has that famous like four-part [63:09] algorithm that he applies to every [63:11] company he does. But there's like emails [63:13] from him and I think he might have even [63:15] treated this. It's just like go ultra [63:16] hardcore on deletion. He is obsessed [63:19] with exactly what you're saying. You [63:20] call it relentless simplification. His [63:23] is like he wants to delete delete delete [63:26] delete as much as possible. Simplify [63:27] simplify simplify. We had Toby Luke on [63:29] the podcast a few months ago and he said [63:31] something that was very interesting. [63:32] He's like well in technology the world [63:34] belongs to the fast. It's to these teams [63:36] that actually actually can get ahead [63:38] [clears throat] by reduction. He's like [63:40] very few teams have understand the skill [63:42] and the genius of getting ahead by [63:44] reducing. And the the illustration of [63:46] his point which he did beautifully. He's [63:48] like, "Well, you know, the modern day [63:49] Picasso would be the the picture of the [63:51] Raptor engine that SpaceX designed where [63:54] it's like you see the first one. [63:55] >> It got super simple." [63:56] >> Yeah. It's got all kinds of weird [63:57] and wires coming out of there and then [63:59] the second version is a little less and [64:01] then the third one's just like [64:02] beautiful. And I accident I actually [64:04] posted the clip like two days ago of [64:06] Toby saying this on the podcast and then [64:07] I just quoted it with the picture of the [64:09] the Raptor and then somebody asked for [64:12] Elon's explanation. He goes into and [64:14] responds to like how he thinks about [64:16] this process, but he's completely [64:18] obsessed with the going ultra hard [64:19] hardcore on simplification on deletion. [64:21] >> It is a superpower, super super [64:23] powerful. Um because yes, it breeds [64:26] speed scalability also besides which is [64:29] a slightly different thing. [64:30] >> But even you just nailed it. He's like [64:32] even goes into like well if I have like [64:34] the the the comp the complexity is [64:36] nonlinear like you just said. If I have [64:38] a 100 parts in this engine compared to [64:39] if I have five like what does supply [64:41] chain look like? what does the [64:42] manufacturing look like? What is the [64:43] repairing it? Figuring out what actually [64:45] went wrong. Like there's just a million [64:46] other things that uh get more [64:48] complicated with more complexity. So [64:50] it's both people don't focus on [64:51] simplification for some reason. I think [64:53] it's really probably there are [64:55] anthropologic reasons. There are [64:57] certainly societal reasons, but people [64:58] do not focus on simplification unless [65:01] again they're, you know, uh unusual, you [65:03] know, radical lateral thinkers like Elon [65:05] or or you teach them. H but when they do [65:08] the second problem, they tend to be [65:09] incremental in it. But often by far the [65:12] biggest wins in terms of simplification [65:13] is complete removal. Uh for example uh I [65:16] mean you just said you know Elon is a [65:18] master at that in our context I remember [65:20] we were uh banging our heads against the [65:23] wall a decade ago approximately with [65:26] with job titles. So we like pretty much [65:29] every company we had we were very small [65:31] but still enough people that job titles [65:32] were a thing. So you wanted to you know [65:34] maybe have a senior this uh staff that [65:38] or director and H [clears throat] we [65:41] were trying to develop definitions who [65:43] who should be a director you know like [65:44] you need to if if that exists if it's a [65:46] thing you probably need to define it you [65:48] know uh so spending time trying to [65:50] define it and then you assign someone [65:51] that title whether it's senior engineer [65:53] and then the other guy who's not senior [65:55] engineer is disappointed it's like why [65:57] why is she senior engineer not senior [65:59] engineer well because of this or that so [66:00] you need to have that conversation and [66:02] then it's emotional drain it takes so at [66:04] some point we were looking for ways to [66:05] streamline it and simplifying it and [66:07] someone said why do we even have titles [66:09] what's the benefit benefit of titles and [66:10] and someone else is like well you need [66:12] titles or everybody has titles and why [66:14] do people have titles let's really try [66:16] to dig deep into the root cause because [66:18] I agree I mean everybody has said [66:20] probably there's some benefit I mean [66:21] let's not be arrogant there's probably [66:22] some benefit what what what's that [66:24] benefit and we ultimately determined [66:26] that the benefit was that people really [66:28] needed titles uh for let's say bragging [66:31] rights it feels good to be able to to [66:33] show progress in one's career and [66:36] they're useful if you need to find a new [66:37] job to to be able to um very [66:41] conveniently and efficiently convey a [66:43] level of experience or capability you [66:45] have you have achieved. And so we were [66:47] like, "Okay, but all we're saying here [66:49] is probably true, but also not something [66:51] that uh the company needs to be involved [66:53] with, right?" And so we we just got rid [66:56] of titles. And we told people, you can [66:59] pick your title for your CV, LinkedIn, [67:01] whatever. We we don't need to know. We [67:03] don't want to know. We don't need to [67:04] approve it. We don't want to see it. Uh [67:05] just don't embarrass us. Like, you know, [67:07] don't you're new hire. Don't say you're [67:08] the CTO because then people will [67:10] question our integrity as a company. But [67:12] as long as it's broadly reasonable, [67:14] we're good. We have never reintroduced [67:16] them again. We don't have any titles. I [67:18] mean the person runs product which [67:20] technically would be called a CPO. We [67:22] have an algorithm. It's just product [67:24] management lead. As simple as that. It's [67:26] completely automated. And and is [67:28] >> what do you mean it's completely [67:29] automated? [67:29] >> The organization is based on algorith [67:32] algorithmic rules. So that if you have [67:34] direct reports and if these reports are [67:38] um product managers automatically this [67:40] tool will call you product management [67:41] lead. But whether you have two or 200, [67:44] you're a product management lead. So [67:45] there's no discussion. We don't need to [67:47] agree whether you are or aren't. There's [67:49] no senior, junior, director, VP. Uh and [67:52] I just made the example of the let's say [67:54] topmost leader in product. For us, it's [67:56] just has the same quote unquote job [67:58] title as a person leading one person. [68:01] And if he needs to, you know, do [68:03] something with his LinkedIn, he could [68:04] put whatever he wants out there. And [68:06] it's just we never have to have this [68:08] discussion. So I we never looked back. [68:09] We probably saved easily hundreds if not [68:12] thousands of person hours in terms of [68:14] defining def you know terms and and [68:17] having emotionally draining discussions [68:19] with people never had a problem not a [68:21] single instance of someone complaining [68:23] that we didn't formally assign to them [68:25] my title ever ever out of at this point [68:28] many hundreds actually multiple [68:29] thousands of people. So that's an [68:31] example of something that everybody does [68:32] a certain way that if you are trying to [68:34] simplify incrementally maybe you achieve [68:37] a little bit of up uplift you know maybe [68:39] 5% but if you get rid of it completely [68:41] it's liberating it's a 10x improvement [68:44] potentially or whatever the baseline you [68:45] want to yeah however you want to measure [68:47] it um and often often not always but [68:50] often you find these opportunities on a [68:52] product get rid of an entire part of the [68:53] product 2% of people use it it's adding [68:56] complexity to codebase bugs issues and [68:58] sure someone will be disappointed But [69:00] you know the the 98% of people who don't [69:02] use it, you can serve them so much [69:03] better that one year on down the line [69:05] you'll be 2x as as well off. Uh just do [69:09] that. Don't slowly transition out a [69:12] million migrations, headaches, issues. I [69:15] found one of my all-time favorite quotes [69:17] when I was reading the book 0ero to1. [69:18] The quote says, "The single most [69:20] powerful pattern I have noticed is that [69:22] successful people find value [music] in [69:24] unexpected places and they do this by [69:27] thinking about business from first [69:28] principles instead of formulas." That is [69:31] exactly what Apploven has done with [69:33] their advertising platform. Apploven [69:35] connects you with over a billion [69:36] potential new [music] customers inside [69:38] mobile games. Apploven allows you to [69:40] capture undivided attention. Apploven [69:43] ads are full screen video ads that are [69:45] watched for an average of 35 seconds. [69:48] That is retention that blows other ad [69:50] platforms out of the water. And you can [69:53] launch on Apploven in minutes. You set [69:55] the goal and Apploven achieves it. [69:57] There's no complex setup, no expertise [70:00] needed, and Apploven scales quickly. [70:03] They can put your ads in front of over a [70:05] billion potential customers. Other [70:08] businesses have seen immediate results, [70:10] have scaled to [music] hundreds of [70:11] thousands of dollars of spend per day, [70:13] and increased their revenue by millions. [70:16] So, you want to get started quickly [70:17] before all of your competitors are on [70:19] Apploven. And you can do that by going [70:21] to apploven.com. [70:23] That's apploven.com. [70:27] Before we go back to these other [70:28] cultural tenants of you, tell me about [70:29] this like automated system you just [70:31] described. It's like running the company [70:32] in the background. What is this? [70:33] >> Yeah, I mean I wouldn't say it runs the [70:35] company in the background, but we we are [70:36] pretty uh fanatical about technology in [70:38] general. Again, I personally was [70:40] involved with AI in 2010, which uh at [70:43] the time nobody I mean it looked uh [70:46] weird because it wasn't a thing really. [70:48] Today, obviously, if you're now building [70:49] a startup with AI, people look uh look [70:51] you what the heck are you doing? Of [70:53] course, you should be building a [70:54] startup. You know, we carried uh with us [70:56] this passion for using technology and [70:58] cutting edge tools to to be more [71:00] productive, more more effective. uh as [71:02] we have invested pretty heavily at [71:04] Bennis Blooms over this point over a [71:06] decade to to develop uh basically you [71:10] could you could look at it as an [71:11] operating system at this point over 50 [71:13] proprietary tools that run almost [71:17] everything that we do or at least [71:18] supported through automation and then we [71:21] buy companies and they it's almost like [71:22] installing them on this operating system [71:24] and a lot of uh the operations are [71:26] subsequently run homogeneously [71:29] consistently and very efficiently [71:31] through it. For example, we have one [71:33] system to manage payments. We have one [71:35] system to run AB tests. We have one [71:37] system to predict user lifetime value. [71:39] We have one system for for recruiting [71:41] and talent uh uh predictions. We have [71:43] one system to orchestrate the many AI [71:46] models we use internally to to to run [71:48] our operations. So, we always use the [71:50] the ideal one in terms of cost uh [71:52] quality. Um we have one system to uh [71:56] authorize different colleagues to have [71:59] access to different systems. solistic [72:00] credentials management uh one system for [72:03] uh data aggregation and processing and [72:05] the list goes on and on and we keep [72:07] refining them and we have kind of a an [72:10] open source community internally whereby [72:12] we have platform teams who own who own [72:15] these different tools and make them [72:16] better by the day but then each of our [72:19] businesses as they use them they find [72:21] ways that they they they they come up [72:23] short they can add features fix bugs and [72:26] as they improve them they these [72:27] improvements are propagated and [72:29] automatically made available to to the [72:31] entire portfolio of businesses. So we [72:33] adding businesses actually makes us [72:35] better as a whole not just because we're [72:37] adding some revenue but because we are [72:39] adding another entry point for for [72:41] innovation improvement ideas on this [72:43] kind of operating system u and it's been [72:46] a a boon for us. It's hard to estimate [72:48] exactly how much in terms of efficiency [72:50] effectiveness it's added but it's [72:52] certainly transformative let's say. So [72:54] adding bit more businesses is better for [72:57] you. But then is that not in conflict [72:59] with I think you're now for your [73:01] acquisitions you want to do fewer and [73:03] bigger. [73:04] >> Yeah. I mean it's there's a trade-off [73:06] obviously the the like in almost [73:08] everything in life [73:10] fewer bigger acquisitions is better for [73:13] us to the extent that uh it means we can [73:15] focus our limited operational capacity [73:19] uh onto those transformations and those [73:22] and and and and getting those right. uh [73:25] we have seen that the in terms of of of [73:28] time and effort it doesn't take a lot [73:30] more time to transform a company that's [73:32] bigger in terms of revenue than a [73:33] company that's smaller. So same amount [73:35] of time invested [73:36] >> roughly speaking Evernote uh in 2023 [73:38] early 23 we uh we had a team a task [73:41] force of spooners these people from the [73:43] core team you know we've been talking [73:45] about uh probably about 50 people who [73:48] joined Evernote and and really drove [73:50] that transformation rewriting the [73:52] codebase rearchitected cloud [73:53] infrastructure rethinking monetization [73:56] and re reorganizing the company and all [73:58] that and and that was you know business [74:00] generating a little less than $100 [74:02] million in revenue [74:03] >> at the time you acquired it? [74:04] >> Yeah, at the time we acquired it. And [74:05] then uh in the first half of this year, [74:08] we uh we we did broadly speaking the [74:11] same thing with with Vmail uh with [74:12] roughly the same number of people, 50 to [74:14] 60, but Vmail uh is roughly $400 million [74:18] in revenue. So approximately four times [74:20] as large and and the team originally was [74:22] over a,000 people. Ver was a little over [74:24] 300 people. So three to fourx the scale [74:27] whether you want to look at revenue or [74:29] headcount. roughly the same number of [74:30] spooners introduced into the business to [74:33] to change it. So [74:34] >> that's incredible. [74:35] >> Part of that is I believe intrinsically [74:37] the complexity of transforming a [74:38] business doesn't scale linearly with the [74:40] revenue of that business. Partly is in [74:42] the meantime we've gotten a lot better. [74:44] For example, expanded and improved that [74:45] operating system. So we we're getting [74:47] more productive but so because of that [74:49] we prefer to acquire relatively few [74:52] businesses and make sure each counts. So [74:55] get it has to be larger and larger as we [74:57] scale as a company. Currently we're at [74:59] roughly $3 billion in run rate revenue. [75:01] So the business that moves the needle [75:02] for us today needs to be a lot bigger [75:04] than than when we acquired Evernote. In [75:06] terms purely of that operating system of [75:08] technologies, we do benefit from more [75:10] diversification because the more teams [75:12] we have who adopt these technologies, [75:15] the more likely we are we are to find [75:17] ways that could be made better, [75:19] innovated on. So u we we have [75:21] >> So how do you reconcile the two? you [75:22] want I [75:22] >> we we tend to to to prioritize the [75:24] former because I think there so bigger [75:26] businesses uh that [75:27] >> because you developed this operating [75:28] system over how many years decade and a [75:29] half something like that [75:30] >> but yeah we started 13 years ago [75:32] obviously you know when we when we [75:34] kicked off the project five five people [75:37] we didn't have the luxury of investing [75:38] in R&D in our technology I think we [75:40] started in earnest with significant [75:42] investments maybe 10 years ago something [75:44] like that [75:44] >> and has anybody I don't think you would [75:46] do this but anybody tried to come and [75:47] like buy these tools from you [75:49] >> first of all we we like to to keep them [75:50] for ourselves cuz competitive advantage [75:52] also you can't do everything you know in [75:54] life you need to talking talking about [75:56] prioritization and focus and we just [75:57] decided that we we use these tools [76:00] >> for own benefit to run the this business [76:02] as well as we can also I don't think [76:04] they would be all that appreciated by [76:06] the broader market for a couple of [76:08] reasons number one they tend to be very [76:10] very advanced most people out there who [76:13] run a digital business they actually [76:15] don't want maybe they think they do but [76:17] they don't want the most sophisticated [76:18] AB AB testing platform overwhelming. [76:21] They're not obsessive about AB testing. [76:24] They want something that's a little bit [76:25] more approachable. So, they wouldn't [76:27] actually necessarily take full advantage [76:28] of of the the real there are solutions [76:31] out there on the market that are more [76:33] mass market, a little bit more [76:34] intuitive, easier that I would recommend [76:37] to them rather than than our own, which [76:39] is again meant for high level of [76:40] sophistication. And lastly, a lot of [76:42] these technologies are doubly powerful [76:45] because they're fully natively [76:47] integrated with one another. they're all [76:49] built to function together. Uh, [snorts] [76:52] and so it's very difficult for a [76:54] business out there to choose to adopt 50 [76:56] different things. They're not going to [76:57] scrap everything they're doing. And so a [76:59] lot of the value fades away from only [77:01] giving you one thing. So I don't even [77:03] think that the the business opportunity [77:04] will be all that great to to to market [77:06] this stuff. I've heard people that don't [77:09] think pay attention to Benny Spoons. [77:11] They're like, "Oh, this is just like [77:12] another like PE play." And I was like, [77:13] "That's I don't think that's it at all. [77:14] Can you so let's I think walk through [77:16] one of the acquisitions right you [77:18] mentioned earlier I don't know if this [77:19] is a term you put on it but like you [77:21] know when you're starting a company [77:23] you're it's kind of you have to like [77:24] luck your way into product market fit [77:27] you don't want to do that you want to [77:28] buy you know a working product so let's [77:31] take Evernote for example I was an [77:32] Evernote customer for I don't know 8 [77:35] years so like what did you see in [77:36] Evernote what was the state of the [77:38] business and then what happened after [77:39] the fact I guess [77:39] >> and by the way I I think people who [77:42] compare benoons to to private equity [77:44] Yeah, they're they maybe have a [77:46] simplistic superficial view of the world [77:47] and they're like okay they acquire [77:48] companies [77:50] uh and you know they've raised prices [77:52] okay like okay but then uh you know [77:54] Google acquires companies that has [77:55] acquired hundreds of companies and has [77:57] raised prices hundreds of times so it's [77:59] a little bit of a pretty limited um set [78:01] of criteria to to compare. I'll give you [78:04] like the highlights on a high level and [78:05] then I'll translate to the very [78:07] specifics of Evernote or any business [78:08] you want me to to talk about. First [78:11] major difference we're not a fund. We [78:12] don't buy to sell. We have never sold a [78:14] material business. We buy to hold and [78:16] operate forever. The second uh very big [78:18] difference is that our interventions on [78:21] the business require are very very deep. [78:24] Again, I'll be very clear as I describe [78:26] Evernote, but we we transform them [78:28] sometimes uh sometimes beyond [78:30] recognition. I'd like to think for the [78:31] better. That's what we try to do here. [78:33] Uh and the third aspect is we integrate [78:35] these businesses very very deeply uh [78:38] into a shared platform including the [78:41] technological operating system we were [78:42] discussing but also this core team of of [78:45] uh we call them spooners who run the [78:48] businesses you know a lot of the R&D [78:50] marketing we move them around fluidly [78:52] across businesses and none of this is [78:55] bears any resemblance to what private [78:57] equity does because those are fun they [78:58] buy to sell after say five years they [79:00] generally intervene yeah maybe on some [79:02] cost or price, but they I've never seen [79:05] a private equity reinvent a product or [79:07] or um or rebuild the or um or rebuild [79:12] the technological infrastructure and [79:14] they generally don't integrate the [79:15] businesses together under a shared [79:16] platform because they don't have the [79:17] platform and even if they did uh they [79:19] need to sell them peace meal. So if you [79:21] integrate them you can't sell them or at [79:22] least it would be much more difficult to [79:23] sell them right. So we are almost as as [79:25] different as it gets uh other than we [79:28] acquire stuff for a living. Um so that's [79:30] for sure uh in common. Now Evernote [79:32] specifically, so um what we saw in it, [79:35] well Evernote in its history has been [79:39] used by a quarter of a billion people. [79:41] you know how many a quarter of a billion [79:43] people extensive reach and usage [79:45] ultimately build a brand naturally [79:47] there's that plus the experience needs [79:49] to be good which often was especially [79:51] for the first uh many years but it's [79:54] certainly it's a brand that almost [79:55] everybody has heard of is familiar with [79:58] often is perceived positively sometimes [80:00] not as much but certainly not negatively [80:01] more sometimes as well oh it something [80:03] from the past is probably not that [80:05] relevant but nobody has a negative [80:06] association with Evernote or very few [80:08] people so a very well-known powerful [80:11] brand, a pretty sizable user and [80:13] customer base, several million active [80:14] users and customers at the time of [80:16] acquisition and to this day. And uh and [80:19] we believe that a substantial [80:21] opportunity for for improvement across [80:23] the board. I'll describe the [80:24] improvements in a moment. Uh and lastly, [80:27] something we we always seek in [80:29] acquisitions is predictability. We like [80:31] to buy stuff where we have a good sense [80:33] of where it's going at least five years [80:35] out at least under management once it's [80:39] installed into our platform and and in [80:41] that case a few factors enabled us to to [80:44] predict its future. one uh uh the the [80:47] user and customer base was was highly [80:49] tenured on average I think a paying [80:52] customer had been on the uh on the [80:54] platform using Evernote for u 5 to 10 [80:57] years I don't remember exactly but I [80:58] assumely long periods of time um the um [81:02] most of revenue was from sub [81:04] subscriptions which we tend to be able [81:06] to uh predict uh in terms of their [81:09] future performance better than uh more [81:11] volatile revenue streams like [81:13] advertising uh most of the value uh lay [81:16] with [81:18] existing users and customers as opposed [81:20] to hypothetical new users and customers [81:22] to that to acquire out there. And we [81:24] find that it's much easier to bet on [81:25] existing customer bases than new [81:27] acquisition because new acquisition of [81:29] user and customers uh tends to be much [81:31] more volatile with the changes in [81:32] competition advertising uh uh dynamics [81:37] in terms of advertising for for [81:38] acquiring customers. Uh so we like the [81:41] whole package. We thought the price was [81:43] uh was reasonable. Do you disclose what [81:45] you bought it for? [81:46] >> Uh, well, it can be it can be seen [81:48] directionally from our financial [81:49] statements. It was about $200 million, [81:51] give or take. [81:52] >> Say that number again. [81:53] >> 200 million. [81:53] >> 200. Okay. More or less about 203 [81:56] something like that. [81:56] >> So, wait, they were doing 100 million in [81:58] revenue, [81:59] >> right? [81:59] >> A little less like 90. [82:00] >> 90. And they were they making any money [82:02] or no? [82:03] >> Uh, I would say roughly break even. [82:05] >> Okay. [82:06] >> Roughly break even. Slightly profitable. [82:09] >> And it's doing what now? So we don't [82:11] disclose profits by by individual [82:14] business but uh [82:17] I would say it's it's very very [82:19] profitable. You can see our overall [82:21] profitability as a group. Um adjusted [82:25] operating income margin we're at around [82:28] 54 55%. Individual businesses are tend [82:31] to be more profitable uh especially if [82:33] you've owned them for more than a couple [82:35] years. [82:35] >> So hold on before you go in there. So [82:37] like an Evernote case, right? You [82:39] drastically increase the profitability [82:41] based on these [82:42] >> revenue revenue. Revenue went up. [82:43] >> Well, that's what I was going to ask. [82:44] Does the revenue also have to go up or [82:46] you were just fine if you just make it? [82:48] >> We we try to improve revenue and improve [82:50] cost. Uh sometimes [82:52] >> we're successful on both fronts. [82:55] Generally, I would say some sometimes [82:57] more on one than the other, but uh on [82:59] Evernote we both increase revenue and [83:00] reduce cost. explain the difference of [83:02] what you were doing compared to they [83:04] were doing whatever 90 million and not [83:06] making any money or breaking even. [83:08] >> So [83:08] >> like what is the difference between how [83:09] you were running the business and how [83:10] they were? [83:11] >> We made a lot of changes. Uh we rebuilt [83:14] the org vastly. Um so it was uh roughly [83:18] 350 team members. Uh we um made it [83:23] substantially smaller. I think a year a [83:25] year and a half after the acquisition [83:27] closed, we were more around 50 to 60 [83:31] team members. [83:32] >> Hold on. So when you acquired it, they [83:34] had 350 people working on the product or [83:36] the company [83:36] >> approximately. [83:37] >> Okay. And you're cutting that down to, [83:39] let's say, 50 to 60. This is what we [83:41] were talking about before we started [83:42] recording, which I think is really [83:44] important. And this is why I like Adam [83:46] from Apploven too cuz you know his whole [83:48] thing is like the the if you factor in [83:51] either his cash flow to employee ratio [83:54] or market cap to employee ratio it's got [83:56] to like that's a very interesting metric [83:58] where his whole thing is like you know [84:00] I'm doing this with 400 employees. You [84:03] mentioned previously it's like to see [84:07] what can be done and how efficient a [84:10] business could be run. It's very similar [84:11] to like it's a good example for other [84:13] entrepreneurs in the same way that like [84:15] other runners didn't crack 4minute mile [84:18] for example until somebody did it and [84:20] then once they see somebody did it then [84:21] you just see it happen all the time. [84:22] >> Yeah. So why could you do this with [84:25] let's say 300 people less than they [84:26] could have? [84:27] >> I think there are different factors. One [84:29] is the access to talent. We we've been [84:31] able to build a an employer brand a [84:33] company that where some of the best [84:35] people want to work. Um, we got 800,000 [84:39] job applications last year. We hired [84:42] fewer than 300 people. So, and and [84:44] [snorts] if you're running Evernote, [84:46] even if you're Steve Jobs, [84:47] >> 800,000 people are not applying to [84:49] Evernote. [84:50] >> Yeah, exactly. And it's not it's not [84:52] anyone's fault. I mean, that executive [84:54] team was was doing the best they could [84:55] with the with the resources they [84:58] reasonably had available. So we we had [84:59] the we have good good fortune to be able [85:02] to to take advantage of a an arbitrage [85:05] in in access to talent. Um we also have [85:08] a massive advantage in that each of [85:10] these businesses matters to us but it's [85:11] not everything and so we can take risks. [85:14] For example, if you're running ever [85:15] Evernote uh and that's all you do that's [85:18] a standalone company and you you make [85:20] the change I just described in terms of [85:22] headcount. If something goes wrong, [85:24] you're out of a job and as a say a CEO [85:28] and realistically you're done because [85:31] [clears throat] that will be the the the [85:32] the blemish on your CV that you can [85:35] never clear pretty much. It's not that [85:37] we want something to go wrong if say [85:39] ever notice part of the bending spoon. [85:40] But we have we can make bets that are [85:43] the expected outcome is highly appealing [85:44] but maybe they're a little bit more [85:46] risky and so not so appealing if that's [85:48] all you do with your life. Uh the upside [85:50] of again it's almost like insurance. We [85:53] on average we get it right and it's a [85:55] great uh value. Occasionally maybe we we [85:58] make a mistake that would be painful if [86:00] it were uh if that business was was run [86:03] as a standalone company but you know net [86:06] net we we do so much better and we learn [86:08] so much faster and the good lessons we [86:09] learn from our business we can port them [86:11] and apply them as you know as relevant [86:13] to to all other businesses. Uh three we [86:16] had access to that technological [86:18] platform. So we have talent talent [86:20] levels the the ability to take some [86:23] risks that would be uncomfortable for [86:25] for for for that management team given [86:27] their boundary conditions. Three that [86:29] the technological platform I discussed [86:31] it just enables us to do so much more [86:34] with the with fewer people. Um and but [86:36] if you're running Evernote standalone [86:37] you're not going to have resources to [86:40] develop those technologies. But also you [86:42] don't have the the business case because [86:45] we can amortize those investments over [86:47] all of our businesses and increasingly [86:48] more businesses is required and it's [86:50] difficult to justify if all you do is is [86:52] evernote and then again those perverse [86:54] incentives uh I mentioned earlier [86:57] whereby if you're judged by Evernote and [87:00] Evernote alone making a change that [87:02] would for example result in a smaller [87:03] number of monthly active users or [87:05] subscribers will get you so much hail uh [87:07] even though it's maybe the right thing [87:08] to do for the business we we could make [87:10] some of those unpopular our decisions uh [87:13] more easily um and and take full [87:16] advantage because the business has been [87:17] thriving relative to the previous [87:19] trajectory and uh financially and in [87:23] terms of uh of customers. So these are [87:25] you know some of the big reasons but [87:27] yeah the changes were were sweeping I [87:28] mean the or we rebuilt it I mentioned [87:30] you know it was a lot smaller and by the [87:32] way today Evernote we ran it with about [87:34] 20 people so it [87:36] >> what yes because in the meantime you [87:38] keep improving so I'll describe the [87:40] improvements we made but uh some are [87:43] fundamental improvements in the [87:44] underlying technology codebase that [87:46] enable the team to do more with with [87:48] fewer resources because everything gets [87:50] a little bit cleaner and more [87:51] maintainable and manageable. Part of it [87:53] is our operating system of technologies [87:55] has gotten so much better in the [87:57] following two three years that we're so [88:00] much more productive. Uh we especially [88:02] with AI we have had some [88:05] close to breakthroughs in productivity. [88:07] Can you talk about that? [88:08] >> Yeah. Uh sure. Um [88:10] >> everybody's interested in this right [88:11] now. There's a bunch of founders that [88:12] have already been on the show that are [88:13] coming back on and we're just going to [88:14] do like an hour of how literally they're [88:16] redesigning their their entire [88:18] organizations with AI. Yeah, we've been [88:19] using AI pretty aggressively for as long [88:22] as I can remember. Certainly in 2018, [88:24] I'd like to say we're using it to [88:26] predict user lifetime basically to [88:29] inform our um AB testing. But uh but I [88:33] would say over the past two years [88:34] especially with very rapid progress in [88:36] LMS, we've been able to have some some [88:39] major breakthroughs in uh in various [88:41] areas, especially software engineering [88:43] and data analysis and and product [88:44] design. I'll give you a couple of [88:46] examples. So for for for design um we [88:50] recently actually deployed a a tool we [88:52] built in house called diagram. Whether [88:55] you're a designer, a product manager, a [88:57] growth manager, you go to this tool. It [88:58] looks a little bit like cloud design [89:00] just broadly speaking, but it's [89:02] specialized in our particular context [89:04] and fully integrated with everything [89:06] else at Benis. And you can just tell the [89:09] tool to pull up uh screens for the app [89:13] you're working on. say it's Evernote uh [89:15] uh for the relevant features and then [89:17] you guide it as it produces new versions [89:20] of those of those interfaces [89:23] uh and it will do so by automatically [89:26] following the design guidelines that the [89:28] say the head designer for that tool has [89:30] laid out in some document somewhere. You [89:32] don't have to know where they are like [89:33] the tool knows. So you just tell it what [89:35] what you need and it'll give you work [89:36] that the head designer would typically [89:38] approve. it will automatically look into [89:40] the codebase to know how the different [89:42] interfaces interact functionally. So it [89:45] will make proposals that make sense from [89:47] that point of view and then once you're [89:48] happy with your proposal it will develop [89:51] the code for you and then the say lead [89:54] engineer will be able to review and [89:55] approve it if it's fine and uh and then [89:57] automatically because it's in it's [89:58] integrated with our AB testing system [90:00] you'll have a new segment where you [90:02] you're going to test that new say on [90:04] boarding flow or whatever. If you had [90:05] the skills before, you're a product [90:06] designer, now you can do it sometimes in [90:08] maybe 1% of the time. In many cases, [90:11] it's actually a better result because [90:12] it's so precise and you just humans we [90:14] we tend to miss things. Uh but [90:17] interestingly, it uh it enables doing [90:21] design work for people before couldn't [90:23] like product managers, software [90:24] engineers, growth managers. So now a lot [90:26] of the inefficiencies that stem from I'm [90:29] a product manager. I have an idea, I [90:30] want to test something, but I need to [90:31] wait for the product designer to be [90:32] available. Then I need to explain to [90:34] them what I have in mind. I fail to [90:36] explain it properly. I get 3 days later [90:37] I get work back and it's not what I [90:40] meant. I need to explain [90:42] the inefficiencies [90:44] stemming from this exchange of [90:45] information. We humans are insanely [90:47] inefficient at exchanging information. [90:49] We're quite efficient at absorbing [90:51] information. But when we when we when we [90:53] have to articulate ideas, you know, [90:55] language is very very it's better than [90:57] not having language, but it's very [90:58] inefficient. Um and more so when it's [91:01] with another human uh with whom the [91:04] iteration cycle will be slow because [91:06] they may not be able to to to do the [91:08] task immediately. Even if they do will [91:10] take them time but with the machine you [91:12] can tell it it'll do it right away. Uh [91:14] and it will take them fraction of the [91:16] time. So you can actually iterate very [91:17] quickly. So overall you get to the [91:18] result in a tiny fraction of the time [91:20] but interestingly you can do it even if [91:21] you can't design. And so this overall [91:24] makes our teams a lot more are [91:26] efficient. This is an example. Another [91:27] example is something we called old [91:30] spooner um for it stands for alter ego [91:33] or alternative spooner and it's [91:35] basically an agent that lives in Slack. [91:38] We use Slack for communications that has [91:41] by design the very same access you do as [91:44] an individual in the company. So it has [91:46] access to the same tools to the same [91:49] degree. So if you have uh full access, [91:52] it's partial access. It's it mimics you. [91:54] It's meant to be you basically but [91:56] artificial [91:58] and you can instruct it to do pretty [91:59] much anything you could do. It could do [92:01] some some things it will do better, some [92:03] things it will do worse. So we have [92:05] Evernote, we we have a channel on Slack [92:07] where we can report feedback on things [92:08] think things could be improved bugs or [92:10] new features and I was there to to [92:11] provide input on something. I I was just [92:13] using to the tool and it failed at [92:15] something and I wanted to to to relay [92:17] that and I I saw live one of the best [92:20] users of Al Spooner by one of my [92:23] colleagues. She runs Evernote and and [92:25] she wrote uh uh you know she in this [92:28] channel she tagged old Spooner and said [92:29] I I noticed this bug could you please go [92:32] to Moros which is again back to the [92:34] integration of our tools. It's it's our [92:36] customer support tool that collects [92:38] feedback from users to check whether [92:40] it's a just I got unlucky or if it's a a [92:43] widespread phenomenon or issue and then [92:45] uh report back so we know how to [92:47] prioritize it and separately can you [92:49] look into the code base for root causes [92:51] for this issue and if you can find and [92:53] propose a fix and then ping um uh uh [92:57] Marco who's lead engineer for that [92:59] particular uh product uh so that he it [93:02] can review the code and push it to [93:04] production if it's fine And so she the [93:07] general manager for Evernote in maybe 3 [93:09] minutes essentially fixed identified and [93:12] fixed a b a bug something that would [93:14] have taken [93:15] >> if this is human to human coordination [93:17] >> yeah weeks maybe [93:18] >> forever like exactly there are many more [93:20] examples I mean I I had to [93:23] >> I wanted to to know the the [93:26] you know that the trajectory of uh [93:28] monthly active users on on on meetup not [93:31] one of our properties recently for an [93:33] analysis I was doing and and generally [93:35] would have to ask a data analyst uh and [93:38] they'd be busy. I would either interrupt [93:40] them or they would get back to me couple [93:42] days later. It would take them [93:43] presumably a couple hours to give me [93:46] that. Uh [93:47] >> I actually interacted with my old [93:49] Spooner went back and forth uh asking [93:52] for further cuts. Okay, just show me [93:54] just for the US just for for users on on [93:57] iPhone just on [93:58] >> and uh you know I got all the answers [94:01] all the graphs in a few minutes. [94:03] >> Perfect. done. [94:05] >> I need to go back to this cuz you just [94:06] blew my mind. So, I know you're not [94:08] telling us exact numbers, but Evernote's [94:10] doing probably a couple hundred million [94:11] or thereabouts in revenue. [94:13] >> Let's say more than 100, less than 200. [94:15] >> Okay, so there you go. Uh that's the [94:16] range of revenue. It h it is profitable [94:19] and you just gave a hint as to what like [94:21] the operating profit percentage might [94:23] look like, right? And you're doing this [94:25] to 20 people. [94:27] >> Yes, that's right. Yeah. Plus the help [94:31] slightly unquantifiable help of that [94:33] platform that keeps pumping out [94:35] technological improvements you know like [94:37] those automatically benefit everybody. [94:39] So it's you know you can allocate it uh [94:41] by dollars in revenue whatever but yes [94:44] people wake up in the morning and fix [94:45] bugs for Evernote launch features [94:47] optimize monetization that's about 20 [94:49] people right now. [94:50] >> Okay. So this kind of efficiency are you [94:52] seeing that in the rest of the [94:53] businesses that you own as well? Yeah, [94:55] for the most part I think that not all [94:56] functions are equally. [94:57] >> Are you optimizing for that? Like [94:59] >> no, I mean we we just try to make each [95:01] business as successful as possible. It's [95:02] not that we know we want to have like [95:04] the smallest number of people that we [95:06] can. I mean if more people create more [95:08] value for customers and for benefits we [95:10] would assuming we can we can hire enough [95:12] you know fast enough we would certainly [95:14] deploy them. Sometimes we have [95:15] situations where we would want to have [95:17] more people we just don't have them. So [95:18] okay that's a separate issue but uh we [95:21] don't aim to minimize number at all just [95:24] to run these businesses as well as [95:25] possible and we often find that some of [95:28] these resistances when you when we when [95:30] you take them back to startup mode so [95:32] they had been [95:34] large slightly bureaucratic sometimes [95:38] political organizations for a while [95:40] things tend to grind to a halt it's [95:41] difficult to be entrepreneurial [95:43] enthusiastic move fast work on what [95:46] matters if [95:48] bring them back to a much smaller size, [95:51] much higher talent density, we get rid [95:52] of a lot of red tape, then even though [95:54] the team is smaller or per perhaps [95:56] precisely because the team is smaller, a [95:59] product development and and optimization [96:01] of monetization pick up again and and [96:03] Evernote is a good example. If you look [96:05] at the it's difficult to precisely [96:07] quantify innovation but if you look at [96:08] the timeline of uh say product [96:10] improvements features before we acquired [96:12] it in the say 2 3 years before and after [96:15] it's night and day I I feel very com [96:18] comfortable saying it's at least three [96:20] times as fast under almost any frame of [96:23] measurement despite the team being much [96:25] smaller but it's really I think despite [96:27] is the wrong word in many ways it's [96:29] because it's a lot smaller and so they [96:30] these people are again startups [96:32] Instagram was built by I was I don't [96:34] know like 10 people I'm not sure [96:36] something like that [96:37] >> I think it was like 12 when they got [96:38] acquired some crazy proof that small [96:40] teams of very capable people with [96:43] extreme ownership who really care can do [96:47] can out work and [snorts] out produce [96:49] vast organizations where either not [96:52] enough people care or they do but there [96:54] are so many feet to step onto and you [96:58] know and so many hurdles to overcome to [97:00] get stuff done that they they they they [97:02] fail to do so and it's not I mean nobody [97:03] Nobody wants that to be the case. It's [97:05] more like frog in the boiling water kind [97:06] of phenomenon where you keep adding [97:08] teams and and and and processes and [97:11] rules and then at some point it's very [97:12] difficult to it's very difficult to care [97:14] and it's very difficult to get stuff [97:16] done. This is what I meant about like [97:18] goes back to how I've been describing to [97:19] other founders that just like this go [97:22] island of entrepreneurship because I [97:23] know you mentioned like being influenced [97:25] by Henry Singleton for example and he [97:26] would do this too. He was like over and [97:28] over again he says hey yeah I think at [97:29] one time he owned 130 different [97:31] businesses and 129 of them were [97:33] profitable but he wanted to break [97:35] business units into the smallest part [97:36] possible parts different was difference [97:38] was between you and him you like [97:40] breaking things down to smaller parts [97:42] less people more efficient but he didn't [97:44] uh he kept kept the business units [97:47] separate [97:48] >> um where you're actually studying them [97:49] all and then using insights and [97:51] spreading across your entire [97:52] organization very similar to like what [97:53] Mark Lender did with constellation is [97:56] there anybody else that you've been [97:58] influenced by or do you take like an [97:59] idea or two from? [98:01] >> Frankly, not a lot. Going back to what [98:03] we were discussing earlier that we we we [98:05] were growing up as a business in no [98:07] man's land in a way in Italy and [98:10] >> um and a purposefully chose to stay a [98:12] little bit isolated to try to [98:15] >> at the risk of reinventing the wheel [98:17] also coming up with some real powerful [98:19] innovation. So, not not so much. I'd say [98:21] maybe um Netflix I think uh I don't [98:25] really know Netflix from the inside. I [98:27] don't know anyone who works there but [98:28] you know their their famous deck uh [98:32] cultural deck and then there have been a [98:34] couple of books that have been written [98:35] about it. I think some of those ideas [98:36] that you want to beat complexity with [98:38] talent not process and keep rules to a [98:41] minimum. I think some of that I think [98:44] rubbed rubbed off on us. Um um but other [98:46] than that I think we've tried to be [98:48] quite autonomous in coming up with our [98:49] own idea. Do you describe bending spoons [98:51] as a conglomerate or no? [98:53] >> I mean it's a conglomerate to the extent [98:54] that [98:54] >> but do you actually use that word? [98:56] >> No, I've never used it. It doesn't it [98:58] doesn't bother me but I think it's a [99:00] it's more at least in my mind a [99:02] conglomerate is a set of relatively [99:05] distinct and separate parts. In our case [99:08] we try to make everything as homogeneous [99:10] and integrated as possible as I [99:12] described. [99:12] >> So Bergkshire and Teladine would be much [99:14] more conglomerate. [99:16] >> Exactly. Exactly. And I think what what [99:18] Singleton and Buffett did did well [99:21] better than almost anybody in history is [99:23] quite different from what I think we've [99:24] been doing really well. They were and in [99:27] the case of Buffett is still is [99:30] exceptional [99:32] at selecting at picking companies [99:35] management teams that should be worth [99:37] more than the market was valuing them. I [99:41] wouldn't say either certainly not [99:42] bircher I think by their own admission [99:44] would they most people wouldn't consider [99:46] them exceptional operators they [99:48] generally in fact would avoid buying [99:50] businesses where they thought a lot had [99:52] to be fixed they didn't like that they [99:53] like business for for the people [99:55] listening to this they haven't studied [99:56] singleton you can just go back I just [99:58] did an episode on him on my other [99:59] podcast it's remarkable how many ideas [100:01] that we've heard from Buffett and Mer [100:03] that singleton discovered like 20 years [100:05] before them and they both yeah Buffett [100:08] and Mer were both say it's like We these [100:10] are this is where we got these ideas [100:12] from. [100:12] >> If I have to think about all the people [100:14] who did exceptional things in business [100:16] and investing, if I had to take their uh [100:19] achievement [100:20] and and assuming we can quantify it and [100:22] divide it by their level of popularity [100:24] or just not meaning they are liked or [100:26] disliked but how well known they are he [100:29] would be at the top of the rank like [100:31] he's been one of the most successful [100:33] investors and business people ever by [100:35] any measure and yet very few people know [100:37] him actually. I think if you ask 100 [100:38] people even in business like 95 will not [100:41] know who it was. [100:42] >> One of the things that Buffett and [100:43] Singleton had in common is they [100:44] essentially primarily saw their job as [100:47] um they were the ones c allocate capital [100:50] like their main talent was capital [100:51] allocation right and Singleton I think [100:54] you might have said this and I could be [100:56] wrong but this is what I'm going to ask [100:56] you. It's like well after he started [100:59] acquiring stopped acquiring companies [101:00] right he bought like 160 in 10 years [101:03] something like that I forgot the number [101:04] exact number and then he's like oh now [101:06] we're going to reverse course we're not [101:07] going to he didn't make another [101:08] acquisition material acquisition for the [101:09] rest of his career and then he just [101:11] focused on capital allocation improving [101:13] the business unit he's had and then just [101:14] discovering like where's the best dollar [101:16] I could spend and is improving operation [101:18] of this company is it buying another [101:20] company and then he discovered it I was [101:22] actually buying back his own shares. [101:23] >> Yeah. So I think I heard you say before [101:26] that out of all the investment [101:27] opportunities you see in the future, it [101:29] might be buying back Benny Spoon's [101:31] shares. [101:32] >> Yeah. Not imminently. I think we see a [101:34] runway right now allocating capital [101:36] toward acquisitions is as as being and [101:38] expect to continue being way too uh like [101:41] returns [101:43] I think will be way too appealing for [101:44] that that not uh to be the the priority. [101:47] But I think if you ask me in the very [101:49] long run uh that could be an appealing [101:52] way of uh of creating um shareholder [101:54] value. I I think what what Singleton did [101:56] incredibly well it was acutely aware of [101:59] the circumstances and boundary [102:00] conditions and uh very creative and made [102:04] fully rational decisions. So for a [102:06] decade or more even the market was uh [102:09] you know affording uh his stock a good [102:12] multiple and it was uh aggregating a lot [102:14] of businesses. it was uh buying at a [102:17] lower multiple and it was on top of that [102:19] exploiting the arbitrage was also very [102:21] uh astute at selecting those businesses. [102:23] It kind of double dipped a business that [102:25] was undervalued uh regardless uh like [102:28] people didn't see the the potential in [102:30] the medium to long term and and add to [102:31] that the fact that it that business [102:33] would that them join a conglomerate with [102:36] a higher multiple. So double value [102:37] creation and then later the market uh [102:40] changed its uh preferences as the [102:42] markets often do. So you got to stay [102:44] open-minded about it and started [102:45] appreciating more vertical businesses. [102:47] And so he worked on improving those [102:50] businesses and spinning them off so they [102:51] could be maximally appreciated. So he [102:54] was never opinionated on opinionated on [102:56] the on the how they should be done. Uh [102:58] it just uh looked I believed I've never [103:01] met him of course but I believed it [103:02] looked at investing and running a [103:04] business as as a puzzle and try to find [103:07] the best solution. He was also a great [103:09] engineer like he could have been one of [103:11] the best engineers that he wanted to [103:12] pursue that and almost a grandmaster in [103:14] chess I believe or at least [103:16] >> he could play chess blindfolded. There [103:17] was a story in I think the episode I [103:19] just did where he's like playing with [103:20] his back turned and he goes hold on you [103:22] told me the wrong move. Three moves [103:24] again. I mean he's obviously genius of [103:26] IQ. Charlie Monger's on record saying it [103:28] was the smartest single human he ever [103:29] met in his entire life. And you imagine [103:30] all the people that Monger met in his [103:32] entire life. Singleton took Teladine [103:34] public almost immediately. Did you know [103:37] when you started bending spoons and you [103:39] okay we're not going to stick with one [103:40] company we're going to keep acquiring [103:41] you started having you did really small [103:43] acquisitions they were successful you [103:45] kept on that path was the plan for you [103:48] and your co-founders like this is going [103:49] to be a public company one day I'd say [103:51] when we talked about public versus [103:53] private I think more often than not we [103:54] thought this would at some point be a [103:56] public company there are advantages and [103:58] disadvantages in being a public company [103:59] I'd say for most companies the [104:01] advantages uh are greater than the [104:03] disadvantages and for a company like [104:05] Benny's phones that require requires [104:06] capital to grow fast. Uh the I think the [104:09] advantages are are way too large. I mean [104:11] it's not really a discussion that you [104:12] should be public but yes it's not all [104:15] roses obviously there are new pressures [104:18] and incentives and noise that you'd [104:21] [clears throat] be better off without [104:22] for sure. [104:23] >> How long it lasts when okay we know [104:25] we're going to go public to you actually [104:26] went public. [104:27] >> We like to make decisions as as late as [104:29] possible. I think that uh [104:33] procrastination is awesome if it doesn't [104:36] come from laziness because if you [104:38] postpone decisions, you often have more [104:40] information when you actually get to [104:41] make them. [104:42] >> Singleton said something like this where [104:43] he's like if you don't make a decisions [104:45] in many cases it resolves itself. [104:47] >> Yeah. [104:48] >> Removes the need to make a decision. [104:50] >> It's a slightly different thing. It's [104:52] it's another reason why uh it's a subset [104:54] of of what I said. meaning [104:57] so there are some decisions that there [104:59] is the only advantage of making them now [105:01] is that you can forget about them. So [105:03] there's something to say about that and [105:05] I think if the decision is not [105:06] particularly important sometimes the [105:08] moment you bring it up just make it so [105:10] that you know you can free up the your [105:12] RAM uh to to to tackle other tasks. But [105:15] if the decision is so critical as to [105:17] whether you should be a public company [105:18] or not or whether you want to buy a [105:20] company or another, uh generally [105:22] speaking, you're better off delaying it [105:24] as much as possible or at least there's [105:27] almost no cost to delaying it other than [105:28] the slight discomfort that you know it's [105:30] still on your shelf. You need still need [105:31] to make it. Uh, and there are advantages [105:33] or at least know I mean worst case [105:36] scenario you'll be just as well off when [105:38] you eventually make it as you were early [105:40] but often you have more information. [105:41] Maybe as as we're just saying the [105:45] boundary conditions shift and you just [105:47] don't need to make it any longer because [105:48] it's irrelevant. Maybe you would have [105:51] made a decision one way but then as the [105:53] world changes or you learn something [105:56] else that you had failed to to spot [105:58] earlier you you end up with a different [106:00] option. And so with the IPO, we decided [106:03] relatively early, probably something [106:05] like first half of 2025, that we would [106:08] want to prepare to go public in the [106:12] near to medium term. And so probably [106:15] late 25, mid 26, late 26, but we would [106:18] delay the decision as to whether [106:20] actually to pull the trigger to as late [106:21] as possible in the process. So we knew [106:24] that we were probably going to be a [106:25] public company at some point. I'd say [106:26] certainly by late 25, no doubt about it. [106:29] But we we didn't know if we would go [106:30] public necessarily in uh in early July [106:32] 26. We just said, "Okay, let's get ready [106:35] and then we'll see." I think the [106:37] definitive decision of okay uh we will [106:40] go public as soon as possible [106:43] definitively we made that decision in [106:45] the spring of 2016. [106:46] >> Earlier you said that assuming that the [106:49] company the first thing a company has to [106:50] do is have a strategy assuming that [106:52] strategy is good then the most important [106:53] thing is talent acquisition. How do you [106:55] articulate the strategy of bennings? [106:58] Well, basically we want to achieve the [107:01] maximum level of operational excellence [107:04] which means getting the most out of a [107:06] business possible [107:09] by any means necessary both in our case [107:11] structural means such as integrating [107:14] everything on the same platform so that [107:15] we eliminate all redundancies and we can [107:17] achieve all sorts of scale advantages [107:18] and network advantages uh and uh by [107:22] sheer investment in talent and and [107:24] technology by any means possible We want [107:28] to achieve the greatest advantage as an [107:30] operator. And once you have that, [107:32] meaning you are a business is better off [107:34] with you than with almost anybody else. [107:36] Once you have that for a sufficiently [107:38] large number of businesses, then you're [107:40] almost guaranteed to be able to compound [107:42] capital very efficiently through [107:43] acquisitions because by definition, by [107:45] definition, mathematically, you'll be if [107:47] a business is better off with you with [107:49] than with everybody else. If there is if [107:51] there are enough of those out there, you [107:52] should be the best, you know, the [107:53] highest bidder there when it's on sale. [107:55] and and so and the seller should still [107:58] get excellent returns from their sale [107:59] and you get excellent returns. So we [108:01] probably focus 99% of our resources and [108:03] and in efforts in in in being the best [108:06] operator building up that platform [108:08] unlocking as much of these structural [108:10] advantages as we possibly can and and [108:13] remarkably little uh in in actually in [108:17] the acquisition side of things. like we [108:18] we are very deliberate, highly [108:20] sophisticated, but but once we have a [108:22] powerful platform and and and these [108:24] structural advantages, then it actually [108:26] gets pretty easy to deliver very high [108:28] returns through acquisitions. It's not [108:30] that we see necessarily things in [108:31] businesses that nobody else saw. It's [108:33] just that we know those businesses are [108:35] going to do so much better with us than [108:36] with almost anybody else. So, we can [108:38] offer more [108:38] >> and that operational excellence allows [108:40] you to bid higher as well. you I think I [108:42] think you said like you're pretty sure [108:44] that you bid maybe 50% higher than the [108:46] next highest bid on like Evernote for [108:47] example. [108:48] >> Yeah. I I mean I don't know I cannot [108:50] never know for absolutely sure because [108:53] you don't you know obviously the sell [108:55] side only tells you so much but I I'm [108:57] pretty confident that our offer was uh [109:00] was way way higher than the which by the [109:02] way in hindsight we should have [109:03] negotiated better. Uh but [109:05] >> no but no but we talked about this at [109:06] lunch. Um, you I think you have a very [109:08] unique Let's talk about this now. Like [109:10] you had a very unique approach to [109:12] negotiation. Like I I'm pretty sure you [109:14] explicitly said like you don't want to [109:16] come in as like a like how most people [109:18] do. It's like let me just put a really [109:20] low number right now and then you say a [109:21] higher one and then we go back and forth [109:23] and back and forth and back and forth. I [109:25] think in time [109:27] uh you want to be known as a I want to [109:29] say generous because obviously nobody [109:31] buys a company as a generosity but uh [109:34] [laughter] [109:35] you you want to come across or establish [109:37] a reputation as someone who's trying to [109:39] you know get the last penny out of a [109:41] negotiation. You want to [109:44] help the seller get good value from the [109:47] transaction. But at the same time you [109:49] you you want to be known as pretty firm. [109:51] like I put a number that I think is [109:53] absolutely fair and highly competitive [109:55] on the table. I probably think I could [109:58] have gotten lower, but again, I'm not [110:00] trying to get all the value out of this [110:02] transaction at all. I wanted to to to [110:04] get a lot of the value, but at the same [110:06] time, I'm not going to be available for [110:08] a lot of back and forth. And uh [110:10] >> you tell them that upfront. [110:11] >> Generally, no. If they ask us, sure, but [110:14] hopefully they do their research. I I [110:16] can only think of one case in recent [110:19] memory which actually quite recent um [110:22] where we ended up raising our offer [110:24] substantially uh because and and [110:28] but the reason is s so in that case we [110:31] we were asked forced really by the [110:33] seller to put a number on the table [110:34] before we had the data we needed. Mhm. [110:37] >> And so, you know, between not even [110:40] participating or risking to having to [110:42] change it a lot later, we said, "Okay, [110:44] look, we don't know a whole lot. Based [110:46] on what we know, we think we'd be happy [110:48] to do this at uh between X and Y." [110:52] And then later as we progressed into, [110:54] you know, the sales sales process and [110:56] and we got more data and we finally [110:58] could form a somewhat complete opinion, [111:01] we found that we could offer a lot more. [111:03] And so we we increased that offer [111:06] substantially. [111:07] >> You increase the offer on your own or [111:09] they said that's way too low. [111:10] >> You know, I don't actually remember [111:13] exactly how it played out. [111:14] >> How would you do that today? [111:15] >> Probably a mix of the two. [111:16] >> How would you do that? Yeah. How would [111:17] you do that today? [111:18] >> I would imagine if you think it's an [111:20] >> I think I would, frankly, I think I [111:21] would do it similarly because we just [111:23] didn't we didn't have the data and be we [111:25] the data wouldn't be forthcoming unless [111:27] you put the number on the table. I don't [111:29] we're not trying to prove a point and be [111:31] dogmatic and say we only put a number on [111:32] the table if we have absolute certainty. [111:34] So we said look we are not highly [111:36] confident in this number because we [111:37] don't have a lot of data but this is the [111:38] number. [111:38] >> Okay. So let's say in a different [111:40] example you have the numbers that you [111:41] need. You put the number out. Is that [111:43] number pretty firm? [111:45] >> Yeah generally yes. I don't think we [111:47] increased it uh almost ever by more than [111:50] 5 or 10%. [111:51] >> So did you ever uh hear about the way [111:52] Buffett bid for Clayton Holmes? [111:54] >> No I don't think [111:55] >> the the founder of Clayton Holmes wrote [111:57] an autobiography. I can't remember I [111:58] can't remember um what it's called, but [112:00] I think his name is Jim Clayton and his [112:01] son was handling negotiations cuz he had [112:03] sat down and his son is CEO. And so his [112:05] son goes to Buffy. He's like, "Yeah, [112:06] we'd uh" they were they were debating on [112:09] like what price per share. He's like, [112:12] "Uh the board would entertain offer at [112:14] 17." And Buffett goes $12.50 [112:17] bid. And then the guy comes back, he's [112:18] like, "All right, we talked over we'll [112:20] we'll take 15." Buffet goes, "1250." He [112:23] goes, "All right." Went back. We're [112:25] going to do 14. And he goes, "11250 is [112:27] my final offer." And then his closer [112:29] was, I can assure you if every capital [112:31] market in the world closed tomorrow. And [112:34] when he's like, "You can still rely on [112:35] this offer." And they're like, "We'll [112:37] take 1250." [laughter] [112:40] I will not name names, but we we have [112:42] had uh one or two situations a little [112:44] bit like that. [laughter] [112:46] And look, I think it's easier to do [112:48] because we're so confident. I don't [112:51] believe to this day [112:54] that we have ever been outbid. I don't [112:56] remember a single case, at least not in [112:58] the last five years, in which we put [113:00] forth an offer and then the seller sold [113:02] to someone else. We have had cases where [113:04] they chose not to sell. Uh maybe they [113:06] thought it was too I suppose they [113:08] thought the offer was too low, but we [113:09] have never seen that business being sold [113:11] to someone else. A and that and we have [113:13] been able to deliver the extremely high [113:14] returns we have while winning uh [113:18] essentially all winnable sales processes [113:21] because of that massive advantage of an [113:23] as an operator where we can deliver such [113:26] a an improved performance visa be [113:28] private equities primarily and most [113:30] other people when you have that uh [113:32] ability to basically bid higher than I [113:34] mean it's I can't say everybody else [113:36] every single time of course there will [113:37] be exceptions but almost everybody else [113:39] almost every time then you can be [113:41] confident in your offer. And uh we have [113:43] seen that sometimes we put forth the [113:45] offer and the the sales party thought [113:46] they they could get more and they choose [113:49] not to engage further and then maybe we [113:51] hear back from them say 9 months later [113:52] or 6 months later and they then they're [113:54] willing to transact at that price cuz [113:55] they they needed to convince themselves [113:57] that actually that's what you can get. [113:59] It's harder to do if you think your [114:01] offer is weak. Then you need to be much [114:03] more persuasive and uh and try to get it [114:06] done before uh people shop it around. [114:08] But in our case we always say do you [114:09] want to shop it? You know, often people [114:11] try to look for exclusives. They're [114:12] like, "Okay, okay, this is my offer, but [114:14] unless I get exclusivity within five [114:17] days or the offer is gone because they [114:20] know that their best chance is to win on [114:22] timing, like I'm here now." They know [114:24] the offer is not that great. In our [114:26] case, when we're asked, we almost always [114:28] say, "Look, if you want to, we we [114:29] encourage you to go and shop it around [114:31] because in fact, once you convince [114:33] yourself that this is the best offer, [114:34] it'll be easier for us to like it'll be [114:36] a lot smoother from we'll sign faster. [114:38] will close more easily like we want it [114:40] to be fully satisfied that this is the [114:42] best value for for you and your [114:43] shareholders that you can get. So that's [114:45] being generally speaking our approach. [114:47] >> So I think from the outside [114:51] I I would ask like how much of your [114:53] business is run by numbers. Remember the [114:54] discussion we had on jiu-jitsu and MMA? [114:57] >> Uh vaguely. Okay. [114:59] >> Which part? You mentioned you ma named [115:02] some people that you were fans of in the [115:04] sport of jiu-jitsu and MMA. And then you [115:06] said the [115:07] >> her we discussed [115:08] >> but then you said one of the weirdest [115:09] anybody's ever said to me in my [115:11] life and you're like oh by the way I [115:12] don't know what they look like. [115:14] >> Oh yeah yeah. [115:15] >> And I'm like how can you be a fan? [115:16] >> I remember. Yes. Yeah. [115:17] >> How can you be a fan of a sport and not [115:19] know what the person looks like? [115:21] >> We don't say I'm a fan of the sport but [115:22] I know I I know something about the [115:24] sport. I'm a little bit of a a geek for [115:28] uh stats, numbers. Um um and so yeah, uh [115:34] same for for a lot of sports like for [115:37] example CrossFit. I don't practice [115:39] CrossFit. I barely ever I probably seen [115:43] >> You don't watch the sports study the [115:46] data that comes off the sport. That's [115:47] what I'm trying to get to. I would tell [115:49] you that Miss Tumi, she's the greatest [115:51] prostit. She probably won like eight [115:53] prostitute games. She only missed once [115:54] when she was pregnant, I think, uh, a [115:57] couple years ago, and then she came back [115:58] and won again. So, I don't know. I just [116:00] love to the stats, [116:01] >> but you don't know what she looks like. [116:03] >> I If she's walking out the street, would [116:04] you? [116:05] >> No, I I don't think I've ever seen her. [116:07] If I've seen her is maybe as I was [116:08] googling, I guess the picture, but [116:10] >> So, help help me understand this. So [116:12] like this part of you which is one of [116:13] the most memorable things you've ever [116:14] said to me right where you have like [116:15] this you had a bunch of knowledge about [116:17] these people like so clearly you retain [116:19] these numbers are you running your [116:22] business the same way I would say I'm a [116:25] strong believer in in logic rational I [116:29] think logic and rationality properly [116:30] defined are perfect they're always good [116:33] for you [snorts] I'm uh skeptical about [116:35] numbers actually meaning numbers can be [116:38] very dangerous because they are an an [116:42] approximation of reality and if you take [116:45] numbers at face value, if you're not [116:48] uh sufficiently skeptical and [116:50] inquisitive, you risk being misguided. [116:52] So [snorts] [116:53] numbers are wonderful and very useful, [116:55] but they need to be handled with care. [116:57] What we try to preach at Ben is there's [117:00] never a decision [117:02] that you have to make where being [117:04] logical and rational isn't the optimal [117:06] strategy ever. No matter how [117:09] quantifiable or unquantifiable the [117:11] matter hand is, you're going to be as [117:13] logical and as rational as you can. [117:15] Whether you should be datadriven, uh [117:18] let's see. I mean, uh some things are [117:20] very clearly well informed by numbers. [117:23] Other things are probably it's useful to [117:26] bring numbers to the table, but they [117:27] don't tell you everything. Some things [117:28] are somewhat dangerous. For example, [117:31] [snorts] [117:31] today we we generate well over four $4 [117:33] million in revenue per spooner. So we [117:36] >> $4 million per revenue per employee [117:38] >> per yeah per like core team employee [117:40] like um [117:42] >> so technically we and we pay some of the [117:47] highest compensation in the markets [117:48] where we operate [117:50] >> because we want to work with some of the [117:51] best people. Um, [117:54] and that's not the main thing, but we we [117:56] want to make sure it doesn't become a [117:57] thing. Like we want to want them to feel [117:58] that they're highly valued and and so [118:00] that we focus on on the the things that [118:02] are actually more exciting and and [118:04] motivating than, you know, the extra [118:06] dollar. So pay needs to be high enough [118:08] that nobody forgets about it, but it's [118:10] not front and center, let's say. Having [118:12] said that, we as as is only natural, we [118:14] don't want to waste money in composition [118:16] if it doesn't bring better talent, [118:18] right? I think I'm not saying anything [118:19] shocking here. And so I I remember [118:21] having this discussion with some of my [118:22] colleagues [118:24] whether we should uh raise salaries or [118:26] paying in general and I was firmly of [118:29] the opinion that we should and we have [118:31] by the way and we will further in the [118:33] future and and someone suggested that we [118:35] run a an experiment and that was we [118:37] would put out their job descriptions [118:40] with the higher salary number than we [118:42] would typically pay the company and see [118:43] whether that would get us more [118:45] applications, better applications, more [118:46] conversion rates. I was in favor of [118:48] running that experiment because had we [118:50] seen major uplifts that would have very [118:53] strongly supported the view that we [118:55] should be increasing salaries. But I [118:57] told the team before we ran the [118:58] experiment that I thought even if we [119:00] didn't see any uplift, I would still be [119:03] of the opinion that we should raise [119:04] salaries. And the reason why I I I [119:07] believe so is that I think that the [119:09] people who click on a job ad and [119:13] actually then decide what to do is [119:17] actually a fraction of the people [119:21] uh that could be clicking on that ad. [119:23] And uh and by the way, a lot of those [119:25] people will have already decided whether [119:27] they're inclined to apply or they're [119:28] just curious. And if you look at [119:30] conversion from a piece of information [119:32] you're you're changing so late in the [119:34] final essentially and running an [119:36] experiment that's going to last two [119:37] months, you're going to fail to observe [119:40] all the compounding effects of you [119:43] establishing a reputation as an [119:44] extremely highpaying company. Those will [119:47] never show up immediately. You need [119:49] people to spread the word at [119:51] universities in workplaces. You need to [119:53] start showing up in the job boards as [119:55] you know there are websites comparing [119:57] that will take many months at a minimum [119:59] probably multiple years. The same way as [120:01] today Benus is generally regarded as one [120:03] of the highest talent density best [120:05] places to go work. It's not something we [120:07] achieved overnight. It's a slow [120:08] investment in that. And by the way the [120:11] the test showed modest uplifts but but [120:14] not enough in and of itself to justify [120:16] maybe paying people 20% more. So I I I [120:19] cannot definitively prove that we were [120:21] right in ultimately increasing [120:22] compensation a lot, but I believe we [120:24] were. If you based your decisions on [120:26] numbers alone or primarily on numbers in [120:28] every case, you're very likely to miss [120:30] out on a lot of opportunity. I'm still [120:31] I'm sure Steve Jobs would have said that [120:34] numbers were occasionally interesting to [120:36] him, but but definitely not the guiding [120:38] deciding factor in many of the best [120:40] decision they made at Apple. So, we we [120:42] try to make the maximum possible use of [120:44] of numbers, but with uh with skepticism [120:47] and and and context, but logic and [120:49] rationality, they never they never fail [120:51] you. They're the best thing. [120:53] >> Luka, man, out of all the founders I [120:54] talked to, you're one of the least [120:56] predictable uh people that I have [120:59] conversations with. I really appreciate [121:00] you exist. I love what you're doing, [121:01] Brend Spoons, and I hope we have [121:02] multiple conversations in the future. [121:04] Thanks for taking the time, man. [121:05] >> Thank you, David. [121:06] >> I hope you enjoyed this episode. Please [121:07] remember to subscribe wherever you're [121:09] listening and leave a review. And make [121:11] sure you listen to my other podcast, [121:12] founders. For almost a decade, I've [121:14] obsessively read over 400 biographies of [121:16] history's greatest entrepreneurs, [121:18] searching for ideas that you can use in [121:20] your work. Most of the guests you hear [121:22] on this show first found me through [121:24] Founders.