The Algorithm of Success: How a Data Scientist Became a CEO With RappiCard Mexico

Ai4

AI-assisted summary of The Algorithm of Success: How a Data Scientist Became a CEO With RappiCard Mexico — original source. Check the original for full context.

Source published: 2025-08-27

Detailed notes

## Background Published by Ai4 on YouTube on August 27, 2025, this presentation features Jose Murillo, Chief Executive Officer of RappiCard Mexico. According to the page description, Murillo challenges the common narrative that confines data scientists to technical back-end support roles, illustrating how technical practitioners can transition into strategic executive leadership through the story of building one of Latin America's most successful FinTechs.

## TL;DR Jose Murillo details the core principles that guided his career progression from a data scientist to the CEO of RappiCard Mexico, leading the FinTech to profitability in under three years and orchestrating its pending acquisition by Banorte. He emphasizes that technical excellence must be coupled with corporate diplomacy, business alignment, and an uncompromising focus on bottom-line ROI, reflecting on how his data team previously generated $4 billion in net revenue for Banorte. Murillo highlights the necessity of embracing professional risk, making rapid, high-stakes executive decisions under incomplete information, and employing daily empirical experimentation to dismantle internal political friction.

## Atmosphere & style The presentation carries a pragmatic, reflective, and candid tone, seamlessly bridging analytical discipline with executive reality. Murillo speaks with calm, understated authority, sharing unvarnished war stories about navigating boardroom resistance, painful turnaround decisions, and regulatory hurdles. The session feels like an intimate, high-level masterclass in corporate leadership that replaces abstract theory with hard-earned, empirical grit.

## Key takeaways - [00:04](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=4s) RappiCard reached profitability in under three years while achieving the highest Net Promoter Score (NPS) in the Mexican credit card market, leading to an acquisition agreement by Banorte. - [01:45](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=105s) Technical data science mastery is only a baseline; practitioners must develop a genuine enthusiasm for core commercial problems to deliver strategic impact. - [02:30](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=150s) Corporate diplomacy and consensus-building are essential to overcome stakeholder skepticism and convert analytical models into tangible business returns. - [05:28](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=328s) Data science initiatives should avoid academic 'shiny objects' and be ruthlessly prioritized around direct, measurable bottom-line value. - [06:20](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=380s) At Banorte, an initial 10x ROI requirement pushed the data science department to generate $4 billion in net revenue over six years, scaling from a 46x to a 200x return. - [08:28](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=508s) Executive career progression requires actively embracing risk and leaving comfortable positions to tackle complex, distressed operational challenges. - [09:30](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=570s) Taking the helm at RappiCard involved turning around skyrocketing credit risk, inflated customer acquisition costs, and cultural rifts between founding stakeholders. - [13:15](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=795s) The primary shift from data scientist to CEO is transitioning from long, deliberate study to rapid, high-stakes decision-making under high uncertainty. - [13:50](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=830s) Halting immediate financial bleeding required decisive early moves, including terminating unsustainable cashback programs and replacing key leadership despite partner opposition. - [14:57](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=897s) Internal politics and subjective opinions can be depersonalized and resolved through continuous empirical experimentation, such as running daily champion-challenger tests. - [15:40](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=940s) Complex legal and regulatory transitions are best managed through forensic, fact-based analysis of statutes rather than speculative internal assumptions.

## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=4s) Bridging the Personal Possibilities Frontier: The RappiCard Story

Every individual operates with what can be described as a "personal possibilities frontier," and true professional success lies in pushing oneself as close to that frontier as possible. Sharing the trajectory of transitioning from a data scientist to a Chief Executive Officer, the speaker reflects on the growth of RappiCard, a digital credit card FinTech in Mexico. While it is exceedingly rare for FinTech startups in Latin America to reach profitability quickly, RappiCard achieved profitability in less than three years while securing the highest customer Net Promoter Score (NPS) in the Mexican credit card market.

This momentum culminated in a major milestone: three months prior, Banorte, one of Mexico's largest banking institutions, entered an agreement to acquire RappiCard. With the transaction navigating regulatory and antitrust approvals, the organization finds itself in the thick of administrative procedures following the initial excitement of the sale. While there is no universal formula for career advancement, specific foundational principles enabled this progression from technical modeling to executive leadership.

```mermaid flowchart TD accTitle: Data Science to Executive Leadership Evolution subgraph Mindset["Core Mindset Shift"] A["Master Technical Skills"] --> B["Understand Business Problems"] B --> C["Corporate Diplomacy & Trust"] end subgraph Execution["Bottom-Line Execution"] C --> D["Focus on ROI & Revenue Generation"] D --> E["Embrace Calculated Risks"] end subgraph Outcome["Executive Milestones"] E --> F["Turnaround of RappiCard"] F --> G["Profitability & Industry-Leading NPS"] G --> H["Acquisition by Banorte"] end ```

[01:45](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=105s) Beyond Technical Knowledge: Business Alignment and Corporate Diplomacy

Reaching one's potential requires first being an exceptionally capable data scientist, but pure technical mastery is only the baseline. A common blind spot among data practitioners is a lack of enthusiasm for the actual business dilemmas confronting company leadership. Data scientists must look past algorithms and mathematical formulas to grasp the underlying commercial stakes and operational friction experienced by business partners, a transition that demands continuous communication.

Beyond domain comprehension lies the critical dimension of corporate diplomacy, often colloquially referred to as organizational politics. Technical solutions cannot deliver an impact on return on investment (ROI) without overcoming entrenched human barriers. In large, highly complex corporations—as well as fast-moving startups—multiple stakeholders hold strong opinions and vested interests. Drawing from early career experience working with the board at a central bank, the most effective leaders were consistently those who took the time to converse with colleagues, establish rapport, and build consensus.

- **Overcoming skepticism**: Business leaders already have established priorities and approaches; they rarely expect a "new bright kid in the team" to arrive with instant solutions and often anticipate technical engagements will be a waste of time. - **Stepping outside comfort zones**: Data scientists naturally gravitate toward peers with similar backgrounds, yet the greatest professional growth occurs when engaging with stakeholders who think differently.

[05:28](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=328s) The Bottom-Line Discipline: Delivering $4 Billion in Value at Banorte

A frequent trap for data science teams is treating their work as an exercise in building "shiny objects" or operating within a quasi-academic vacuum. In commercial enterprises, technical initiatives are justified solely by their direct contribution to the bottom line. Prior to leading RappiCard, the speaker moved from central banking to Banorte to establish their data science department roughly a decade ago.

Upon signing, executive management instituted a strict condition: the newly funded data science team was required to generate at least ten times its operational cost. Although daunting, this mandate proved to be a "blessing in disguise," as it forced immediate, ruthless prioritization on high-value initiatives and eliminated wasteful pursuits where business partners lacked commitment.

```mermaid flowchart LR accTitle: Banorte Data Science ROI Trajectory subgraph Targets["Financial Milestones"] T1["Mandate: 10x ROI"] T2["Year 1: 46x ROI"] T3["Year 6: 200x ROI"] end subgraph Aggregate["Cumulative Outcome"] O1["$4 Billion Net Revenue"] O2["Harvard Business School Case Study"] end Targets --> Aggregate ```

- **Year 1 ROI**: The data science group yielded **46 times** its operating cost in its initial year. - **Final Year ROI**: By the sixth year of tenure, the team delivered **200 times** its cost. - **Total Financial Impact**: Generated **$4 billion in net revenue** for Banorte across a six-year period. - **Academic Recognition**: The transformation and bottom-line methodology became the subject of a Harvard Business School case study.

[08:28](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=508s) Embracing Risk: Rescuing and Scaling RappiCard

Following the success at Banorte and the publication of the Harvard business case, staying put would have been comfortable and prestigious. However, career growth necessitates an active willingness to "embrace risk." When Rappi—the prominent on-demand delivery platform and Latin America's first unicorn—launched RappiCard, the venture quickly encountered severe operational distress within its first year.

RappiCard was grappling with skyrocketing credit risk costs, customer acquisition costs (CAC) that were exceeding those of traditional incumbent banks rather than undercutting them, and disappointing customer growth targets. Deep cultural divides between the founding stakeholders created an impasse, prompting them to recruit a leader capable of translating between technology, risk, and commercial execution. Despite initial hesitation, the decision was made to accept the CEO position. While the reality on the ground was far more challenging than recruiter descriptions suggested, aggressive restructuring restored financial viability.

- **Turnaround timeline**: Achieved positive unit economics (variable income exceeding variable costs) in **less than 3 years**. - **Customer satisfaction**: Scaled to become the credit card with the **highest NPS in Mexico**. - **Deliberate discomfort**: Moving into a cultural integration with Banorte presents fresh challenges, reinforcing the principle that "if you're feeling too comfortable, then you should do something different."

[13:15](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=795s) Q&A: Decision-Making Velocity and High-Stakes Leadership

Responding to an audience question regarding the specific skills required to adapt from a technical scientist to an executive leader, the fundamental difference lies in the luxury of time versus the necessity of acting under uncertainty. Whereas data scientists typically have the latitude to study complex problems deeply before proposing a solution, a CEO must execute swift, consequential decisions with incomplete data.

Upon assuming the chief executive role, immediate intervention was necessary to halt severe financial bleeding. The previous administration had instituted aggressive, unsustainable cashback programs that were draining capital. In the very first week, the decision was made to terminate these cashback schemes. This directly triggered a second difficult decision: replacing the Chief Marketing Officer (CMO). One of the company's partners explicitly requested that the CMO remain in place, particularly because significant portions of the cashback expenditures were flowing directly to one of that partner's own businesses. Navigating this conflict required decisive action under high pressure and significant ambiguity.

[14:57](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=897s) Q&A: Navigating Internal Politics with Empirical Rigor and Evidence

Addressing a question on how to manage corporate politics, legal departments, and cross-functional friction, the solution lies in replacing subjective belief systems with rigorous, empirical experimentation. Stakeholders frequently hold strong convictions that are unsupported by hard data. Implementing champion-challenger testing frameworks helps de-escalate political friction by allowing empirical outcomes to settle debates.

During the ongoing acquisition and regulatory transition—moving from a standalone FinTech regulatory regime to the stricter framework governing traditional banks—debates arose regarding product permissions. While internal teams spent weeks debating potential legal constraints based on assumptions, a direct, forensic examination of the banking statutes combined with an external legal opinion clarified what was permissible under the law. Combining unassailable evidence with dedicated personal time to walk stakeholders through the findings provides the clearest path through corporate diplomacy.

- **Testing frequency**: Running approximately **one randomized controlled trial (RCT) / champion-challenger test per day** to validate business hypotheses. - **Evidence-based diplomacy**: Resolving complex regulatory and internal debates by anchoring discussions in granular analysis of the law rather than unverified opinions.

## Notable quotes > [00:04](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=4s) Every individual operates with what can be described as a "personal possibilities frontier," and true professional success lies in pushing oneself as close to that frontier as possible. — Jose Murillo

> [00:04](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=4s) While it is exceedingly rare for FinTech startups in Latin America to reach profitability quickly, RappiCard achieved profitability in less than three years while securing the highest customer Net Promoter Score (NPS) in the Mexican credit card market. — Jose Murillo

> [01:45](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=105s) Data scientists must look past algorithms and mathematical formulas to grasp the underlying commercial stakes and operational friction experienced by business partners, a transition that demands continuous communication. — Jose Murillo

> [01:45](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=105s) Technical solutions cannot deliver an impact on return on investment (ROI) without overcoming entrenched human barriers. — Jose Murillo

> [05:28](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=328s) A frequent trap for data science teams is treating their work as an exercise in building "shiny objects" or operating within a quasi-academic vacuum. In commercial enterprises, technical initiatives are justified solely by their direct contribution to the bottom line. — Jose Murillo

> [05:28](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=328s) Upon signing, executive management instituted a strict condition: the newly funded data science team was required to generate at least ten times its operational cost. — Jose Murillo

> [08:28](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=508s) if you're feeling too comfortable, then you should do something different. — Jose Murillo

> [13:15](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=795s) Whereas data scientists typically have the latitude to study complex problems deeply before proposing a solution, a CEO must execute swift, consequential decisions with incomplete data. — Jose Murillo

> [14:57](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=897s) Stakeholders frequently hold strong convictions that are unsupported by hard data. Implementing champion-challenger testing frameworks helps de-escalate political friction by allowing empirical outcomes to settle debates. — Jose Murillo

> [14:57](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=897s) Combining unassailable evidence with dedicated personal time to walk stakeholders through the findings provides the clearest path through corporate diplomacy. — Jose Murillo

## Mentioned Jose Murillo, RappiCard Mexico, Banorte, Rappi, Harvard Business School, Ai4

## Video description Jose Murillo, Chief Executive Officer, RappiCard Mexico

Abstracts

Despite being hailed as the architects of modern innovation, data scientists often find themselves confined to back-end roles—valued for their technical insight, yet rarely considered for strategic leadership. In many organizations, the path from building models to making executive decisions remains frustratingly narrow, with data science seen as a support function rather than a driver of vision. I will challenge that narrative as I share my journey building RappiCard, one of the most successful FinTechs in Latam.

[Source](https://www.youtube.com/watch?v=IeE8LcJOVi0)

中文简报

成功算法：一位数据科学家如何成为 RappiCard Mexico 的 CEO Jose Murillo 详述了引领其从数据科学家晋升为 RappiCard Mexico 首席执行官的核心原则；在他的带领下，这家金融科技企业在不到三年内实现盈利，并促成了由 Banorte 收购的相关交易。他强调，卓越的技术必须与企业外交手腕、业务目标对齐以及对底线投资回报率（ROI）的坚定关注相结合，并回顾了他此前带领的数据团队如何为 Banorte 创下 40 亿美元的净收入。Murillo 还着重指出，必须敢于承担职业风险，在信息不完备的情况下快速做出重大的管理决策，并借助日常的经验实验来化解内部政治阻力。 RappiCard 在不到三年内实现盈利并斩获墨西哥信用卡市场最高的 NPS（净推荐值），目前已与墨西哥最大银行之一 Banorte 达成收购协议。 (0:04) 数据科学项目必须摒弃学术化空谈并严苛聚焦底层商业回报；在 Banorte 期间，数据团队在 6 年内创造了 40 亿美元净营收，投资回报率从初期的 46 倍攀升至 200 倍。 (5:28) 高管职业突破需要主动拥抱风险，走出舒适区：接手第一年陷入信贷风险飙升、获客成本失控及文化割裂的 RappiCard，通过大刀阔斧的重组实现逆境翻盘。 (8:28) 从数据科学家向 CEO 跃迁的关键在于从容深入的研究思维转变为在高不确定性与高压力下的快速决断，例如在上任首周果断叫停不可持续的返现计划并撤换高管。 (13:15) 化解企业政治与跨职能部门分歧的最佳方式是用实证实验取代主观信念，包括每天运行冠军-挑战者测试（champion-challenger test）以及依托确凿法条事实推进监管合规。 (14:57)

Reading summary

Transitioning from technical modeling to executive leadership requires pairing data science rigor with commercial focus, corporate diplomacy, and the capacity to make rapid decisions under uncertainty. Drawing on experiences at Banorte and RappiCard Mexico, the speaker argues that technical teams succeed only when strictly measured by their bottom-line impact and empirical evidence. Bridging Technical Rigor and Corporate Diplomacy (1:45) Pure technical expertise serves merely as a baseline for data practitioners. The speaker notes that data scientists frequently stumble by isolating themselves in technical silos, overlooking the practical business dilemmas and operational frictions faced by commercial leadership. Bridging this gap requires continuous engagement with stakeholders holding divergent viewpoints and overcoming initial skepticism from partners who often anticipate that technical initiatives will waste time. Strict Bottom-Line Discipline and ROI Mandates (5:28) Technical teams must avoid pursuing academic exercises or "shiny objects" without clear revenue potential. When establishing Banorte's data science unit roughly a decade ago, management instituted a mandate requiring the department to generate at least ten times its operational cost. This constraint eliminated low-commitment projects and enforced strict commercial prioritization. The strategy drove substantial financial returns: the team generated a 46-fold return on cost in its first year, achieved a 200-fold return by its sixth year, and produced $4 billion in cumulative net revenue over that period, later becoming the subject of a Harvard Business School case study. Turnaround and Risk-Taking at RappiCard (8:28) Professional growth requires moving past comfortable positions to embrace operational risk. RappiCard, the digital credit card joint venture between delivery unicorn Rappi and financial partners, faced critical distress within its first year, including escalating credit risk costs, customer acquisition costs outstripping traditional banks, and cultural gridlock between stakeholders. Stepping in as CEO, the speaker implemented immediate structural interventions, including eliminating unsustainable cashback programs during the first week and replacing the Chief Marketing Officer despite partner resistance. These changes enabled RappiCard to attain positive unit economics in under three years, reach the highest Net Promoter Score (NPS) in Mexico's credit card market, and secure an acquisition agreement with Banorte. Empirical Experimentation to De-Escalate Internal Friction (14:57) Executive leadership contrasts with analytical roles because executives lack the luxury of prolonged analysis and must decide quickly amid incomplete data. To navigate high-stakes corporate politics and cross-functional friction, leadership relies on empirical frameworks rather than personal conviction. RappiCard implemented approximately one randomized controlled trial or champion-challenger test daily to validate commercial hypotheses. In addition, when internal teams stalled over regulatory constraints during the transition to Banorte's traditional banking regime, management resolved debates through forensic reviews of banking statutes and external legal counsel, using objective evidence to establish consensus. RappiCard's acquisition by Banorte remains subject to administrative and regulatory antitrust approvals, presenting continuing operational and cultural integration challenges as the venture migrates into a traditional banking regulatory environment.

精炼阅读摘要

演讲者结合自身从数据科学家转型为首席执行官（CEO）的历程指出，技术人员要突破职业潜能边界，不仅需要扎实的算法功底，更必须建立底线财务思维、主动承担高阶风险，并运用实证机制化解组织政治与监管摩擦。这一套方法论助力拉美数字信用卡初创企业RappiCard在三年内扭亏为盈，并在墨西哥信用卡市场取得最高客户净推荐值，最终促成其被大型银行机构Banorte收购。超越技术：商业对齐与组织政治协调 (1:45) 在演讲者看来，卓越的数学建模能力仅是数据科学家的基本盘，技术人员最常见的盲区在于对管理层直面的商业难题缺乏敏感度。许多技术专家倾向于停留在技术同侪的舒适区内，专注于打磨复杂的算法模型，却忽视了业务部门的运营痛点与商业诉求。要使技术成果切实转化为商业回报，数据科学家必须走出象牙塔，积极与不同思维方式的利益相关者沟通，逐步化解业务管理层对技术团队“空降”方案所持的天然怀疑。组织外交（即内部政治管理）同样是技术方案产生投资回报（ROI）不可或缺的支撑。演讲者结合早年在央行董事会工作的经历指出，无论是在大型复杂集团还是敏捷型初创企业中，不同利益方均持有深固的诉求与立场。最具成效的领导者往往不是单纯依赖技术权威，而是愿意投入时间与各方建立互信、展开对话并凝聚共识，从而打破阻碍业务推进的人际壁垒。底线导向：以严苛ROI创造40亿美元价值 (5:28) 商业环境中的数据团队极易陷入闭门造车或追求“闪亮新工具”的学术真空陷阱。演讲者强调，企业内一切技术项目的存在价值，完全取决于其对财务底线（bottom line）的直接推动。约十年前，演讲者从央行加入墨西哥大型银行Banorte组建数据科学部门时，管理层便设立了严苛的门槛：新建团队创造的商业价值必须达到其自身运营成本的至少10倍。这一考核硬指标倒逼团队实行了极度务实的业务优先级筛选，果断淘汰掉业务伙伴缺乏落地承诺的低价值项目。在该机制驱动下，该数据科学团队在第一年便实现了运营成本46倍的投资回报率，到第六年该比例提升至200倍。在六年的任期内，团队累计为Banorte创造了40亿美元的净营收，该团队将数据科学深度绑定财务产出的实践模式随后被哈佛商学院编写为教学案例。承担风险：RappiCard危机重组与扭亏 (8:28) 持续的职业跃升需要主动走出舒适区并拥抱风险。在Banorte取得巨大成功后，演讲者并未选择留在既有的声誉体系中，而是选择出任陷入困境的RappiCard首席执行官。作为拉美首家独角兽即时配送平台Rappi旗下的数字金融业务，RappiCard在上线首年便深陷多重危机：信用风险成本激增、客户获取成本（CAC）甚至高于传统老牌银行、获客目标未能达成，且初创合伙人之间存在严重的文化分歧，亟需一位兼通算法技术、风险控制与商业运营的复合型领导者主持大局。面对远比招聘阶段描述更为严峻的现实，管理层推行了强有力的重组措施以恢复财务可持续性。最终，RappiCard在不到三年时间内实现单位经济效益转正（即可变收入覆盖可变成本），扭亏为盈，这在拉美金融科技初创企业中极为罕见。同时，该产品攀升至墨西哥信用卡市场客户净推荐值（NPS）第一位，这一势头促成约三个月前Banorte签署协议收购RappiCard。决断速度与实证外交：化解政治阻力 (13:15) 从技术专家到CEO的跃迁，核心挑战在于从拥有充裕研究时间的分析模式切换为在信息不完备且高度不确定下的快速决断。数据科学家习惯深入穷尽变量后得出结论，而首席执行官则必须承受高压风险迅速行动。在接管RappiCard的第一周，面对严重的财务失血，演讲者顶住压力果断叫停了不可持续的激进现金返还（cashback）补贴政策，并直接撤换了首席营销官（CMO）。尽管有创始合伙人因返现支出涉及其自身关联业务而极力要求保留该高管，但管理层依然坚决推进了止血决策。针对内部政治分歧与复杂的合规摩擦，演讲者倡导以实证实验打破主观臆断。团队推行了每天约运行一次随机对照试验（RCT）或胜者挑战（champion-challenger）测试的机制，让客观数据成为平息争议的标准。此外，在RappiCard由独立金融科技监管转轨至传统银行严密合规体系的过程中，面对内部团队对产品权限基于主观假设的长达数周的争论，管理层通过深入研判银行法律条文并结合外部法律意见厘清了合规边界，以确凿的事实依据推动了组织外交与业务合规过渡。尽管RappiCard已达成收购协议并进入反垄断与监管审批流程，但在完成初创企业向受严密监管的传统银行体系过渡期间，仍面临行政与文化整合的实际挑战。演讲者总结认为，职业发展并不存在通用公式，但唯有警惕过度安逸、持续在不确定性中承担计算过的风险，并坚持以实证数据消除主观分歧，技术领导者才能不断逼近组织与个人的潜能边界。

Detailed notes — zh

## Background 该视频由频道 Ai4 于 2025 年 8 月 27 日发布在 YouTube 平台上。视频记录了 RappiCard Mexico 首席执行官何塞·穆里略（Jose Murillo）的主题演讲，分享他如何打破数据科学家常被局限于后台支持角色的困境，成功转型为战略决策者并带领 RappiCard 崛起为拉美顶尖金融科技公司的经历。 ## TL;DR 何塞·穆里略系统拆解了数据科学家跃升为企业 CEO 的核心路径：必须走出技术象牙塔，深度洞察业务痛点并展开组织“外交”，且始终以企业的底线利润（ROI）为唯一定位。在临危受命接管陷入获客与风控危机的 RappiCard 后，他在信息高度不完备的重压下快速做出叫停不可持续返现、调整高管等关键决断，最终在三年内实现单位经济效益转正并被墨西哥北方银行全资收购。面对复杂的职场政治与合规争议，他倡导通过每日高频的 RCT 科学实验与精读法条原文，用扎实的客观事实与耐心沟通来凝聚共识。 ## Atmosphere & style 整场演讲基调坦诚、务实且逻辑极其严密，既具备顶尖量化专家的严谨底色，又兼具成熟商业领袖的宏观格局。主讲人毫不掩饰接任初期面临的焦虑、人事博弈与自我怀疑，以真实的实战故事娓娓道来，赋予听众强烈的共鸣感。问答互动沉稳而犀利，生动展现了技术精英走出心理舒适圈、用数据理性直面商业不确定性的坚定力量。 ## Key takeaways - [00:04](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=4s) Jose Murillo 带领墨西哥数字信用卡金融科技公司 RappiCard 在不到三年内扭亏为盈，成为当地 NPS 最高的信用卡，并达成了被墨西哥头部银行 Banorte 全资收购的协议。 - [01:45](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=105s) 从技术骨干向高管跃升，过硬的数学和算法建模只是基本门槛，更需要深刻理解业务部门面临的真实商业痛点与挑战。 - [01:45](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=105s) 数据科学家必须学会“企业外交”，主动走出舒适区与不同背景的利益相关者建立互信，克服组织内的人性壁垒。 - [03:55](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=235s) 央行工作经验表明，最有效的领导者往往依赖高频的非正式沟通来推进关键议程并达成共识，而非闭门造车。 - [05:37](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=337s) 企业数据科学必须以底线效益（Bottom Line）为导向，坚决避免将资源浪费在无实际商业回报的自嗨型项目上。 - [05:37](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=337s) 在 Banorte 期间，Jose Murillo 团队实现了首年 46 倍、末年 200 倍于团队成本的收益，累计创造 40 亿美元净营收并入选哈佛商学院案例。 - [08:30](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=510s) 接任 RappiCard CEO 初期面临信贷风险失控、获客成本高于传统银行及创始团队文化割裂等严峻危机，需依靠算法优化风控与精细化单位经济效益管控力挽狂澜。 - [08:30](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=510s) 拥抱风险是一种刻意练习，当在原有岗位感到过于安逸舒适时，往往意味着应该迈出舒适区挑战全新领域。 - [13:29](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=809s) 数据科学家与 CEO 最大的思维差异在于：数据科学家拥有充足时间深入验证，而 CEO 必须学会在信息不全和高度不确定性下迅速拍板决断。 - [13:29](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=809s) 上任 CEO 第一周即果断终止不可持续的巨额返现补贴，并顶住合伙人压力撤换存在利益输送冲突的 CMO，迅速扭转财务危机。 - [14:57](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=897s) 化解组织政治与主观偏见的有效方法是将数据科学严谨性引入管理，通过高频运行 RCT（Champion-Challenger）实验用客观数据说话。 - [14:57](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=897s) 应对复杂的合规与监管争议时，应通过精读法律原文与引入外部专业意见来消除无休止的口头争论，并投入充分时间帮助各方理解证据逻辑。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=4s) 个人可能边界与 RappiCard 的崛起每个人心中或许都有一条“个人可能边界”（Personal Possibilities Frontier），而衡量成功的标准之一，就是看自己究竟能走得多靠近这条边界。何塞·穆里略（Jose Murillo）开宗明义地指出，他此次分享的核心并非提供一套普适的成功秘诀，而是回顾自己如何从一名纯粹的数据科学家逐步转型为金融科技公司首席执行官（CEO）的真实轨迹与心得。他所掌舵的 RappiCard 是墨西哥的一家数字信用卡金融科技公司。在拉美地区的 FinTech 行业中，绝大多数初创企业长期深陷亏损泥潭，而 RappiCard 却在短短不到三年的时间里实现了全面盈利，并一举斩获全墨西哥最高的客户净推荐值（NPS）。这种亮眼的业绩表现，促使墨西哥大型金融集团北方银行（Banorte）在三个月前达成了对 RappiCard 的全资收购协议。目前团队正处于度过收购狂热期之后、等待反垄断与竞争监管部门审批的繁复官僚流程之中。 - **核心成就**：不到 3 年实现盈利，成为墨西哥 NPS 最高的信用卡产品。 - **最新动态**：墨西哥头部银行 Banorte 已签署收购协议，目前正处于反垄断审批阶段。 --- ### [01:45](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=105s) 跳出技术象牙塔：理解业务与组织“外交” 要想从一名卓越的数据科学家跃升为企业高管，首先必须在技术领域达到顶尖水平，但这仅仅是基本门槛。数据科学家往往醉心于算法、模型与数学推导，却常常对企业面临的实际业务痛点缺乏热情。然而，真正的商业突破不仅需要过硬的数学功底，更需要深刻洞察业务伙伴每天面临的真实业务挑战，而这背后需要大量的跨界沟通。 ```mermaid flowchart TD accTitle: 数据科学家向 CEO 跃迁的四大支柱 subgraph S1["第一阶段：技术与业务融合"] A["技术专业能力 (数学与算法建模)"] --> B["业务痛点洞察 (理解真实商业需求)"] end subgraph S2["第二阶段：组织影响力"] B --> C["企业外交与信任构建 (跨部门沟通协作)"] C --> D["底线效益聚焦 (ROI 驱动与资本回报)"] end subgraph S3["第三阶段：领导力跃迁"] D --> E["拥抱未知风险 (走出舒适区接任 CEO)"] end ``` 除了业务认知，何塞强调的第二项核心能力是组织内的“政治”或“外交”（Diplomacy）。在复杂的现代企业架构中，数据科学要转化为实际的投资回报率（ROI），中间横亘着重重的人性壁垒。无论是大型机构还是初创团队，内部都存在大量利益相关者，大家对数据团队的动作各有看法。数据科学家必须主动走出办公室，投入时间与业务负责人建立互信，让他们相信数据算法不是空谈，而是能切实解决痛点的利器。 --- ### [03:55](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=235s) 跨越舒适圈：从央行经验看跨部门沟通何塞分享了自己职业生涯早期在中央银行工作的感悟。在央行董事会中，最卓有成效的董事往往不是闭门造车的人，而是愿意花大量时间与其他成员私下深入沟通、达成共识的人。这一法则同样适用于数据团队：业务条线的主管通常有自己的考核优先级和解决问题的既定思路，他们并不会理所当然地张开双臂欢迎一个“自以为聪明绝顶的新员工”来指手画脚，甚至往往会先入为主地认为技术方案只是在浪费大家的时间。因此，数据科学家必须主动跨出心理舒适区。人们天生喜欢与思维方式相似的人打交道，但真正的认知升级与商业突破，恰恰来自于与那些背景迥异、诉求不同的人展开深度对话。只有倾听并理解业务部门的种种细微顾虑，技术方案才有可能真正落地生根。 - **央行核心洞察**：最有效的领导者依靠高密度的非正式沟通来推进关键议程。 - **跨部门痛点**：业务端普遍怀疑算法团队会浪费时间，唯有靠持续的信任投资才能破局。 --- ### [05:37](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=337s) 聚焦底线效益：创造 40 亿美元净营收的哈佛案例数据科学家最容易陷入的误区，就是把精力耗费在打造“光鲜亮丽的摆件”（Shiny Objects）上，或是将企业研发当成准学术研究。但商业世界的终极检验标准永远只有一个——能否为企业的底线利润（Bottom Line）带来实质贡献。在创立金融科技公司之前，何塞曾在 Banorte 银行任职 6 年。刚从央行加盟这家大型银行时，管理层批准为他建立数据科学团队，但同时提出了一个极为严苛的前提：“你带来的收益必须达到团队成本的 10 倍以上。” > “在第一年，我们创造了相当于成本 46 倍的收益；而在我任职的最后一年，这个数字达到了 200 倍。” 这道看似苛刻的紧箍咒最终成了“塞翁失马式的幸运”（A blessing in disguise）。出于生存的本能，何塞被迫聚焦于每个项目的实际价值产出，将有限的资源精准投入到真正有丰厚利润、且业务伙伴愿意配合落地的场景中，坚决不把时间浪费在无意义的自嗨项目上。在 Banorte 的 6 年间，他的团队累计为银行创造了高达 40 亿美元的净营收，这一传奇成就后来被哈佛商学院（Harvard Business School）写成了经典教学案例。 --- ### [08:30](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=510s) 拥抱不确定性：临危受命出任 RappiCard CEO 当一个数据科学家坐拥哈佛案例、累计贡献数十亿美元净利润时，他在企业内部无异于享受着“摇滚巨星”般安逸舒适的待遇。然而，要真正实现从技术领袖向 CEO 的飞跃，就必须学会主动拥抱风险。彼时，拉美首家独角兽外卖巨头 Rappi 旗下的金融科技业务 RappiCard 刚上线不到一年便陷入了巨大困境：信贷风险成本飙升、获客成本不仅未能如愿低于传统银行，反而急剧恶化，用户增长停滞不前。创始合伙人们来自截然不同的背景与文化圈子，内部矛盾重重，急需一位既懂金融风控、又懂技术和商业的跨界领导者来力挽狂澜。 ```mermaid flowchart LR accTitle: RappiCard 从危机到盈利的逆转路径 subgraph Crisis["接任初期的严峻危机"] A["信贷风险成本失控"] B["获客成本高于传统银行"] C["创始团队背景割裂"] end subgraph Turnaround["战略重构与执行"] D["算法优化风控模型"] E["精细化单位经济效益管控"] end subgraph Success["三年实现突破"] F["边际利润为正 (Positive Unit Economics)"] G["登顶全墨 NPS 最高信用卡"] H["获 Banorte 银行全面收购"] end Crisis --> Turnaround --> Success ``` 面对这一挑战，何塞一开始十分抗拒放弃优渥平静的生活。尽管猎头已提前告知情况严峻，但当他真正接手 CEO 职位后才发现，现实远比描述的更为凶险。在无数个焦虑难眠的夜晚与反复的自我怀疑中，团队经过不到三年的攻坚，终于实现了正向的单位经济效益（Unit Economics，即边际收入大于边际成本），让 RappiCard 彻底绝地求生，不仅实现了商业盈利，更成为墨西哥客户口碑最佳的信用卡品牌。面对如今与传统银行 Banorte 融合带来的全新文化冲突与未知数，何塞坦言自己依然感到紧张，但他坚信： > “拥抱风险是一种刻意练习。如果你感到太舒适了，那就说明你应该去做些不一样的事情了。” - **接手危机**：RappiCard 运营不足 1 年即面临风控恶化、获客成本高于传统金融机构的致命危机。 - **逆风翻盘**：3 年内实现单位经济效益转正并扭亏为盈，最终完成高估值并购退出。 --- ### [13:15](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=795s) 现场问答：从数据科学家到 CEO 的关键能力跃变演讲结束后进入现场提问环节。观众席的一位听众首先发问：从一名纯粹的技术/数据科学家转型为执掌整家公司的首席执行官，在这个巨大的角色跃迁过程中，究竟必须学习和补齐哪些关键的核心技能？何塞随即展开了针对性解答。 ### [13:29](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=809s) 从数据科学家到 CEO：在高不确定性下快速决断作为数据科学家与担任首席执行官（CEO）之间最大的差异，在于决策的时间窗口与信息完备度。作为数据科学家，通常有充裕的时间深入钻研问题、建立模型并反复验证；然而作为 CEO，现实情况往往迫使管理者在缺乏足够数据支撑的情况下，迅速做出重大抉择。 ```mermaid flowchart LR accTitle: 角色转变与决策模式对比 subgraph DataScientist["数据科学家模式"] A["深入研究与验证"] --> B["充足时间与完备数据"] B --> C["高确定性决策"] end subgraph CEO["CEO 管理模式"] D["迅速决断"] --> E["信息不全与高度不确定性"] E --> F["业务止血与人事调整"] end ``` Jose Murillo 分享了自己刚接任 RappiCard Mexico CEO 第一周时面临的严峻考验。当时公司的财务状况非常困难，前任管理层发放了巨额且在财务上完全不可持续的返现（Cashbacks）。Jose Murillo 在上任第一周就果断叫停了这笔不可持续的补贴。紧接着是更具政治阻力的人事决策：某位合伙人曾明确要求不要撤换首席营销官（CMO），但背后原因竟是大量返现资金实际上流向了该合伙人旗下的关联公司。面对这种利益冲突，Jose Murillo 依然坚决做出了撤换 CMO 的决定。他坦言自己当时也希望能有更多时间去深思熟虑，但作为 CEO，必须学会在面临更高不确定性时快速拍板。 - 数据科学家拥有充足时间深入研究，而 CEO 必须在缺乏完整数据的情况下快速拍板。 - 上任首周即终止前任推行的巨额、不可持续的返现补贴，化解财务危机。 - 顶住合伙人施压，果断撤换将返现利益输送给关联公司的 CMO。 --- ### [14:57](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=897s) 应对组织政治：用科学实验与严谨证据说话在问答环节，现场观众提问数据科学家转型为管理者后，该如何应对组织内部复杂的职场政治，尤其是与法务等强监管部门的博弈。Jose Murillo 回应称，组织内部许多人往往持有非常强烈的个人观点，但深入探究会发现这些看法大多仅仅基于主观信念而非事实。为了打破这种主观偏见，他将数据科学的严谨性引入管理流程，公司每天会运行大约一次随机对照试验（RCT，即 Champion-Challenger 测试），通过快速的小规模实验和数据结果来验证假设，从而让业务决策建立在客观事实之上。 ```mermaid flowchart TD accTitle: 组织政治化解路径 subgraph Disagreements["组织分歧与政治阻力"] M1["主观信念与强烈偏见"] M2["法务合规争议与猜测"] end subgraph Solutions["严谨科学应对法"] S1["每日 RCT / A/B 实验验证"] S2["精读法律条文 + 外部专业意见"] S3["花时间沟通与解释证据"] end M1 --> S1 M2 --> S2 S1 & S2 --> S3 S3 --> Final["达成共识与高效推进"] ``` 在涉及并购与银行强监管合规问题时，证据同样是化解分歧的最有力武器。随着公司推进并购事宜，团队内部对于从 FinTech 牌照转入传统银行监管体系后“哪些业务合规、哪些会被禁止”争论了数周。Jose Murillo 发现各方其实并未真正仔细研读法条，于是他带领团队逐字逐句深入研究法律原文，并获取了权威的外部法律意见，从而彻底终结了无休止的口头争论。他强调，管理组织政治的核心在于两点：一是拿出扎实过硬的事实证据，二是愿意投入时间与相关方耐心沟通，确保大家真正理解证据背后的逻辑。 - **每日 RCT 实验**：通过高频运行随机对照试验（A/B 测试/Champion-Challenger），用小步快跑的数据反馈打破主观偏见。 - **依法求证终结内耗**：面对金融合规争议，拒绝空谈猜测，通过精读法条原文与引入第三方外部法律意见来厘清合规边界。 - **重在沟通与对齐**：仅有证据还不够，必须投入充足的时间与各方沟通，帮助团队和利益相关者理解证据，达成共识。 ## Notable quotes > [00:04](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=4s) 个人可能边界 — Jose Murillo > [03:55](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=235s) 自以为聪明绝顶的新员工 — Jose Murillo > [05:37](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=337s) 光鲜亮丽的摆件 — Jose Murillo > [05:37](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=337s) 你带来的收益必须达到团队成本的 10 倍以上。 — Banorte 管理层 > [05:37](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=337s) 在第一年，我们创造了相当于成本 46 倍的收益；而在我任职的最后一年，这个数字达到了 200 倍。 — Jose Murillo > [05:37](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=337s) 塞翁失马式的幸运 — Jose Murillo > [08:30](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=510s) 摇滚巨星 — Jose Murillo > [08:30](https://www.youtube.com/watch?v=IeE8LcJOVi0&t=510s) 拥抱风险是一种刻意练习。如果你感到太舒适了，那就说明你应该去做些不一样的事情了。 — Jose Murillo ## Mentioned Jose Murillo（何塞·穆里略）, RappiCard Mexico, RappiCard, Rappi, Banorte（墨西哥北方银行）, 哈佛商学院（Harvard Business School）, 墨西哥中央银行, Ai4 ## Video description Jose Murillo, Chief Executive Officer, RappiCard Mexico Abstracts Despite being hailed as the architects of modern innovation, data scientists often find themselves confined to back-end roles—valued for their technical insight, yet rarely considered for strategic leadership. In many organizations, the path from building models to making executive decisions remains frustratingly narrow, with data science seen as a support function rather than a driver of vision. I will challenge that narrative as I share my journey building RappiCard, one of the most successful FinTechs in Latam. [Source](https://www.youtube.com/watch?v=IeE8LcJOVi0)

Source quotes

[0:04] Jose Murillo: 个人可能边界 [3:55] Jose Murillo: 自以为聪明绝顶的新员工 [5:37] Jose Murillo: 光鲜亮丽的摆件 [5:37] Banorte 管理层: 你带来的收益必须达到团队成本的 10 倍以上。 [5:37] Jose Murillo: 在第一年，我们创造了相当于成本 46 倍的收益；而在我任职的最后一年，这个数字达到了 200 倍。 [5:37] Jose Murillo: 塞翁失马式的幸运 [8:30] Jose Murillo: 摇滚巨星 [8:30] Jose Murillo: 拥抱风险是一种刻意练习。如果你感到太舒适了，那就说明你应该去做些不一样的事情了。

Source quotes

[0:04] Jose Murillo: Every individual operates with what can be described as a "personal possibilities frontier," and true professional success lies in pushing oneself as close to that frontier as possible. [0:04] Jose Murillo: While it is exceedingly rare for FinTech startups in Latin America to reach profitability quickly, RappiCard achieved profitability in less than three years while securing the highest customer Net Promoter Score (NPS) in the Mexican credit card market. [1:45] Jose Murillo: Data scientists must look past algorithms and mathematical formulas to grasp the underlying commercial stakes and operational friction experienced by business partners, a transition that demands continuous communication. [1:45] Jose Murillo: Technical solutions cannot deliver an impact on return on investment (ROI) without overcoming entrenched human barriers. [5:28] Jose Murillo: A frequent trap for data science teams is treating their work as an exercise in building "shiny objects" or operating within a quasi-academic vacuum. In commercial enterprises, technical initiatives are justified solely by their direct contribution to the bottom line. [5:28] Jose Murillo: Upon signing, executive management instituted a strict condition: the newly funded data science team was required to generate at least ten times its operational cost. [8:28] Jose Murillo: if you're feeling too comfortable, then you should do something different. [13:15] Jose Murillo: Whereas data scientists typically have the latitude to study complex problems deeply before proposing a solution, a CEO must execute swift, consequential decisions with incomplete data. [14:57] Jose Murillo: Stakeholders frequently hold strong convictions that are unsupported by hard data. Implementing champion-challenger testing frameworks helps de-escalate political friction by allowing empirical outcomes to settle debates. [14:57] Jose Murillo: Combining unassailable evidence with dedicated personal time to walk stakeholders through the findings provides the clearest path through corporate diplomacy.

Original transcript

[0:04] I guess every one of us has kind of like [0:07] a personal possibilities frontier [0:11] and if you want to measure success in [0:13] some sense is how close are you to that [0:16] frontier [0:18] and what I'm pretending to u to talk [0:21] about today it's some of the things that [0:24] helped me on my way from being a data [0:26] scientist to become the CEO of a company [0:29] card. It's a fintech in Mexico which um [0:34] uh it's a digital uh credit card that uh [0:38] in the laps of u less than 3 years we [0:42] became profitable. Uh it's not it's not [0:45] that common in Latin America for fintex [0:47] to be profitable. It also has the [0:50] highest uh customer net promoter score [0:53] and that led for um a large bank Borte [0:58] to purchase the company uh 3 months ago [1:02] or we are in the process once the [1:04] competition authorities approve it. So [1:07] we're kind of in the midst midst of uh [1:10] of all the bureau bureaucratic processes [1:12] after all the hype of selling the [1:14] company. So uh [1:19] during this I basically in I have some [1:24] really simple slides. The idea is to [1:26] have a conversation more kind of with [1:28] you guys if you have some questions. [1:30] It's [1:32] it's kind of hard to think that uh a [1:36] recipe that would work for everyone. So [1:37] it's some of the things that helped me [1:39] in my way. So the first thing is uh [1:45] you need really to to uh [1:49] to move from being a successful first [1:52] you have to be a successful data [1:53] scientist to be something else and [1:55] really you really want to be on the on [1:58] that uh frontier of your own personal [2:02] possibilities but and for that you need [2:04] to have technical knowledge. But one [2:06] thing that usually data scientists are [2:09] more [2:11] are not that excited about sometimes [2:14] it's to learn about the business [2:15] problems that uh the companies are [2:18] facing. And so you you need to [2:20] understand not only the math to do the [2:22] technical part but also to understand [2:24] what are the [2:27] the business the true business problems [2:29] that um that uh your partners are your [2:35] business partners are facing and that [2:38] requires a lot of uh I guess a lot of [2:41] communication and so yeah it helps to [2:44] know the technique but it also helps to [2:46] know the business problems And the and I [2:49] I guess there's a third part in the next [2:51] slide what I'm talking about it's I call [2:53] it politics matter but it's more kind of [2:56] diplomacy or [2:58] you need to invest time with the with [3:01] the business leaders to understand their [3:03] problems and that they build the trust [3:06] and that you will bring them solutions [3:08] that they they will be helpful [3:11] and it's it's a process that there are [3:13] lots of human barriers into [3:17] making uh [3:21] um [3:23] data science have an impact in ROI. [3:26] That's a common theme that probably [3:27] you've heard in different uh talks today [3:31] and uh especially in in highly complex [3:36] corporations in a startup you think it's [3:38] easy. It's not usually that easy, but uh [3:43] but in in large complex organizations [3:46] what there are lots of people that have [3:47] a stake and they're talking about uh and [3:50] have an opinion on what you're doing. So [3:53] it's it's usually you need to invest [3:55] time and uh that's something that I [3:58] learned of working on my first [4:00] incarnation as a central banker. I work [4:02] with the board and and it usually the [4:05] most effective board members devoted [4:08] time to speak with other guys so that [4:11] they can make things work. And it's [4:14] something that that that I would [4:15] encourage that uh for any data scientist [4:18] to to get out of their their team and [4:20] talk to the business leaders and [4:22] understand their problems and build [4:24] trust with them and so that you you [4:27] understand all the nuances that they [4:29] they want to to work on. Usually it's a [4:33] there's there's a mixture of things that [4:36] happens in business that usually the [4:38] business leaders have their own [4:40] priorities. They have their own ideas on [4:42] how they are going to solve their their [4:45] problems and and it's not necessarily [4:47] that they have a a new bright kid in the [4:50] in the team that will come and [4:53] make [4:55] bring them all the solutions. usually [4:57] they think it's going to be a waste of [4:58] time and uh so you need to to invest in [5:02] those relationships with the with the [5:05] business with the business leaders and [5:07] and that forces you to get out of your [5:10] comfort zone because typically we think [5:13] that we are more comfortable talking to [5:16] people that are more like us than [5:18] different than us. And uh but all the [5:22] learning comes usually when you talk to [5:24] people that are kind of like different [5:26] from you. [5:28] So um [5:31] usually you need to understand the [5:33] business, you need to understand the [5:34] people and the third on the next slide [5:37] what I'm talking about it's um [5:43] at the end of the day you are and and [5:45] and I think that's a common confusion [5:47] that uh data scientists make. it's they [5:51] think that [5:53] it's about creating shiny objects or [5:56] about um [5:59] being in in a quasi academic environment [6:03] but at the end of the day it's uh it's [6:05] what's the impact on on the bottom line [6:09] why when I before this uh fintech that [6:14] I'm building I used to to work at Vanort [6:17] which is the company that purchased is [6:19] the my own fintech but make made it [6:22] easier. We knew each other from time [6:26] before [6:28] and and I moved to that bank after being [6:31] a central banker [6:34] and it was a highly complex organization [6:38] in which after I signed up they they [6:41] said well yeah we'll give you funds to [6:44] to make a [6:46] data science team. it was going to be [6:47] the that was about 10 years ago time [6:52] but it had a kind of like uh something [6:54] that I never never expected. They said, [6:56] "Well, yeah, you need to yield 10 times [6:58] your cost." [7:01] And that was kind of like a blessing in [7:03] this in disguise because at the end of [7:04] the day, it forced me to focus on [7:07] creating value on each data science [7:09] project. After six year tenure, I buil I [7:12] created $4 billion in net revenue for [7:15] for Banort. That was a Harvard business [7:18] case study. Uh and uh [7:23] and I guess it [7:26] it for me it was kind of like a matter [7:28] of survival because I needed to to yield [7:31] that uh [7:33] that kind of profit. On the first year I [7:35] yielded 46 times the cost and on on my [7:37] last year as a over there it was 200 [7:41] times the the cost. [7:44] And I guess when I when when it's [7:47] focusing on the bottom line, it it helps [7:49] you to optimize the resources that you [7:52] have and uh and to uh and really not to [7:57] waste a lot of resources on things that [8:00] either [8:02] the partners are not willing to to do, [8:05] they're only going to waste time. So [8:07] it's it's a question of where's where's [8:09] the the money [8:12] purse or the the price and who is [8:15] willing to work on it and and that's the [8:18] way that I decide on how to do that and [8:20] that built a lot of credibility and and [8:22] uh and uh in in in that I was able to do [8:26] this thing [8:28] and I guess the the the last thing that [8:30] really helped me on my way to become a [8:32] CEO after being a data scientist for a [8:35] long time it's [8:37] And it's on the on the last on the last [8:39] slide because I after having a Harvard [8:42] business case study and uh delivering $4 [8:46] billion in net revenue life was very [8:48] sweet [8:50] but uh [8:53] I on the on the next slide what I'm uh I [8:56] guess the main message is you need to [8:59] embrace risk and uh life was very sweet [9:02] at that time. I would never complain of [9:06] my previous life. You know, you would be [9:09] kind of like a rockstar in a any company [9:11] that you're delivering that kind of [9:13] profit. But suddenly there was this uh [9:16] fintech which was uh starting rapart. [9:20] Rapids [9:22] for uh it's a delivery app. It was the [9:24] first Latin American unicorn in uh but [9:27] it was almost uh [9:31] in after [9:33] less than a year it was in deep trouble [9:35] that company they were having a [9:38] skyrocketing uh risk costs the [9:42] acquisition cost that was supposed to be [9:45] uh [9:47] going downwards [9:49] it was uh or be significantly less than [9:52] a traditional bank. was above a [9:54] traditional bank and they were not [9:56] getting all the customers they wanted. [10:00] the founding um partners were very upset [10:05] about it and they they couldn't find u [10:07] and they were from different kind of [10:09] like worlds and [10:11] so they couldn't find someone who would [10:13] be able to translate [10:16] between between them and they offered me [10:19] the this position to be the CEO of of [10:22] the company and initially I was very [10:24] reluctant to take that I was leading a [10:27] very happy [10:31] And at the end of the day, I I I [10:34] took the risk. A lot of times I kind of [10:37] like regretted it. Uh it was much the [10:41] when I when when I got into the job, [10:44] they told me that it was in a in a tough [10:47] spot, but it once you get there, you [10:51] really understand that it's probably [10:52] much tougher than the recruiting people [10:55] were telling you. After [11:01] less than three years, we became uh we [11:04] had unit positive unit economics. So [11:06] that's variable cost minus uh variable [11:11] income minus variable costs. And so it [11:13] mean it makes it makes sense to have a [11:15] business that you have you're creating [11:18] marginal revenue in some sense. [11:22] and and it became a success story and I [11:25] was very happy with it and uh and uh and [11:29] I guess the the last part of it's uh [11:33] once I was starting to feel comfortable [11:36] again and uh and to feel that uh well uh [11:41] rapicard became a success story and and [11:44] uh that's it. I can uh be comfortable [11:48] for for some time. Well, now it's uh and [11:52] and as I said, it's the leading the [11:55] highest uh NPS uh credit card in Mexico. [11:59] And so it's, you know, it's very nice [12:02] story. [12:04] But uh [12:07] there was an opportunity to for to have [12:09] a uh a traditional bank to purchase uh [12:13] the fintech. The cultures are very [12:15] different different. It's not going to [12:18] be easy and uh and I'm again uh a bit [12:22] nervous on the change but uh I think [12:24] that uh embracing risk it's a deliberate [12:27] practice and uh so I'm not really sure [12:29] what will happen on the next on the next [12:32] uh months once the the the the company's [12:36] acquired completely but uh but I guess [12:39] uh the main thing is that uh if you're [12:42] feeling too comfortable then you should [12:44] do something different. Uh it's it's [12:46] easier uh said than done, but uh [12:50] evidently uh you kind of like uh [12:55] uh different stages. It uh you don't [12:58] sleep that well. But uh but I guess uh [13:02] that's the the fun part of uh of us [13:04] living. And I guess that's it. If you [13:07] have any questions, happy to address [13:09] them. [13:15] What skills did you have to learn in [13:17] order to adapt from your role as a tech [13:20] scientist and to new leadership role as [13:22] chief executive officer? [13:25] Uh I guess I guess one of the big skills [13:29] or one of the things that [13:32] was that uh as a as a data scientist you [13:35] have uh enough time to study things [13:38] deeply [13:40] and uh and as a a chief executive [13:43] officer there are a lot of of uh fast [13:46] decisions that you need to make with not [13:48] enough data. And so probably the first [13:51] thing that uh that I did when as as a [13:54] CEO [13:56] the the finances were having a tough [13:59] time and and my predecessor was granting [14:03] huge amounts of money cash backs that [14:06] were not financially sustainable [14:09] and uh and so that was a decision that I [14:13] took on on the first uh on the first [14:15] week. Uh the second one of the other [14:18] other things is that one of the partners [14:21] deliberately asked me not to remove the [14:25] CMO because [14:27] actually a lot of the cash back was [14:29] going to one of the companies of that of [14:32] that partner. So uh that was probably [14:34] the second decision to change the CMO [14:37] and I wish I had the luxury at that time [14:39] of having more [14:42] more time to to ponder and and uh and so [14:46] I guess that's the big difference [14:48] probably I have now to make more [14:51] decisions with much with a higher degree [14:55] of uncertainty. [14:57] >> Hello thank you. How do you manage [14:59] politics inside of the organizations [15:01] that you have been part of it? Because [15:04] like data scientists u we always deal [15:07] with analytics and all the stuff but we [15:10] also lead with lawyers and a lot of [15:13] politics inside of the organizations. [15:15] How do you manage those so you can come [15:17] forward and [15:20] you know in this success? [15:22] >> I guess I guess there are a lot of [15:23] things that have help and and it depends [15:25] on the case. For example, there are [15:27] things that people have strong opinions [15:30] and I find that uh having um that a lot [15:36] of those opinions are based on you know [15:38] beliefs and and uh I try to impose more [15:42] uh more rigor in the decision process [15:45] people make. And so we for example are [15:48] running about uh one RCT per day uh a [15:52] randomized control trial that's uh uh [15:56] champion challenger tests that you're [15:58] doing and uh and uh and you run little [16:02] experiments and that kind of proves [16:04] things and that that kind of helps with [16:06] business. Then there are now that we're [16:08] in the midst of a of a purchase there [16:10] are a lot of questions that are arising. [16:13] does the regulation will will uh allow [16:16] you to do this or that once you're a [16:18] fintech you have a different type of [16:20] regulation but in within the bank you'll [16:22] have a different type of regulation and [16:24] again again evidence helps a lot uh just [16:30] before this meeting there there was some [16:31] discussion if we were going to be able [16:33] to do something [16:36] once we were uh part of a regulated uh a [16:39] bank a traditional bank with different [16:41] regulation [16:43] And I found that they didn't that all [16:45] the discussions that they've been having [16:46] for a couple of weeks, they didn't [16:48] really study carefully what the law [16:51] said. And so at the end of the day, what [16:54] I did is uh uh we studied carefully the [16:58] law and and then we got an an external [17:02] opinion. But usually I guess uh there's [17:05] a lot of uh [17:09] you need evidence and the other thing is [17:12] investing time in talking to people and [17:15] and and so that they have time to [17:18] understand that evidence.