Lessons from Crystal Widjaja

Lessons from Crystal Widjaja

As Gojek's first data hire, Crystal Widjaja built the infrastructure that handled the company's jump from 30,000 to over 10 million daily transactions. She has since advised at Reforge, co-founded the STEM non-profit Generation Girl, and started the Semi-Technical newsletter. This profile collects her rules for building data teams, running growth experiments, and using AI tools effectively.

Part 1: Data Infrastructure & Engineering

1. On Data Integrity: "The most critical initial step for any data team is establishing what the real numbers are before attempting complex analytics." — Source: [Brave Southeast Asia Tech]
2. On Garbage In, Garbage Out: "If engineers aren't focused on reporting accuracy at the source, no amount of downstream modeling will fix the resulting bad data." — Source: [Brave Southeast Asia Tech]
3. On Database Scaling: "Transitioning from a MySQL monolith to PostgreSQL is often a necessary pain when query times for simple cohort analysis become unsustainable." — Source: [Lenny's Podcast]
4. On Democratizing Data: "Tools like Metabase are powerful because they allow self-serve access to reports across the company without requiring expensive per-seat licenses." — Source: [Lenny's Podcast]
5. On Analytical Bottlenecks: "If the data team spends their entire week pulling simple metrics for other departments, they have no time left to actually analyze user behavior." — Source: [Brave Southeast Asia Tech]
6. On Building Trust: "When people doubt the dashboard, they stop using it. Maintaining trust in the data pipeline is a non-negotiable requirement for data engineers." — Source: [Semi-Technical Substack]
7. On Tool Empowerment: "If you give teams access to a tool where they can help themselves, they will do it, and it empowers them to learn entirely new skills." — Source: [Malaka Cinematic Podcast]
8. On Problem-First Architecture: "Don't build infrastructure based on what resources you currently have; design it for the ideal solution to your specific business problem." — Source: [Data Magic Podcast]
9. On Defining Metrics: "A metric is only useful if everyone in the organization agrees on its exact definition and how it is calculated from raw events." — Source: [Brave Southeast Asia Tech]

Part 2: Product Management & Alignment

1. On Altitude Maps: "The goal with Altitude Maps is not to always be right, but it's to be less wrong by connecting daily tasks to overarching company goals." — Source: [CreativeMornings Talk]
2. On Unmeasurable Metrics: "Some of the most important product outcomes cannot be measured perfectly through A/B testing, requiring teams to rely on qualitative signals." — Source: [Reforge Unsolicited Feedback]
3. On Managerial Alignment: "A good product manager understands that if they win but their manager loses, everyone loses. Solutions must serve broader team goals." — Source: [CreativeMornings Talk]
4. On Strategic Decisions: "Strategy isn't just about what to build next; it's about explicitly deciding what not to build, even when customers ask for it." — Source: [Config APAC 2024]
5. On TARS Framework: "Evaluating product health requires looking at Target Audience, Adoption, Retention, and Satisfaction as interconnected metrics." — Source: [Reforge Unsolicited Feedback]
6. On Feature Bloat: "Shipping features constantly without measuring their usage leads to organizational debt and a complicated user experience." — Source: [Lenny's Podcast]
7. On Product Speculation: "You can spend months debating if a feature will work, or you can build a lightweight version in a week and let the users decide." — Source: [Config APAC 2024]
8. On User Feedback: "Users are great at identifying their pain points, but they are often terrible at designing the optimal technical solution for those problems." — Source: [Reforge Unsolicited Feedback]
9. On Cross-Functional Empathy: "Product managers must learn to speak the language of engineering, design, and marketing to effectively advocate for their roadmap." — Source: [Config APAC 2024]

Part 3: Growth & Experimentation

1. On Scrappy Growth: "Early-stage growth isn't about sophisticated machine learning models; it's about manually figuring out what drives your core loop." — Source: [Lenny's Podcast]
2. On Daily Experiments: "Running experiments daily allows a team to validate assumptions rapidly and adjust course before sinking resources into the wrong path." — Source: [Data Magic Podcast]
3. On False Correlation: "If you give everyone a voucher at 6:00 a.m., it won't work if morning users already have strong brand loyalty. You can't generalize that behavior." — Source: [Malaka Cinematic Podcast]
4. On Marketplace Supply: "In a two-sided marketplace, throwing money at user acquisition is useless if you don't have enough supply to handle the initial demand." — Source: [Lenny's Podcast]
5. On Growth Teams: "A successful growth team operates like a highly technical unit focused purely on quantitative decision-making and leverage." — Source: [Brave Southeast Asia Tech]
6. On Incentive Design: "Merchant incentives need to be carefully structured; otherwise you train your partners to only engage when they are being subsidized." — Source: [Lenny's Podcast]
7. On Retention Over Acquisition: "Acquiring ten users who churn immediately is far more expensive and damaging than acquiring two users who stick around for years." — Source: [Data Magic Podcast]
8. On High-Leverage Actions: "Growth is about identifying the single step in the funnel where a 1% improvement yields a 10% increase in overall revenue." — Source: [Lenny's Podcast]
9. On Failing Experiments: "An experiment that disproves a core hypothesis is just as valuable as one that increases conversion, provided you actually learn from it." — Source: [Data Magic Podcast]

Part 4: AI & Context Engineering

1. On Agent Infrastructure: "Product managers need to move beyond basic prompts and learn to build agent infrastructure that scales their analytical capabilities." — Source: [Semi-Technical Substack]
2. On Context Engineering: "The output of any AI model is strictly limited by the quality and structure of the context you provide it before asking a question." — Source: [Semi-Technical Substack]
3. On AI as Chief of Staff: "Delegating low-leverage tasks like drafting emails and organizing notes to AI frees up mental bandwidth for strategic thinking." — Source: [Semi-Technical Substack]
4. On Command-Line Literacy: "Learning to use the command-line interface allows non-engineers to interact directly with AI developer tools and automate their own workflows." — Source: [Malaka Cinematic Podcast]
5. On Productivity Traps: "It is easy to spend more time organizing your AI tools than actually doing the work those tools were meant to accelerate." — Source: [Semi-Technical Substack]
6. On Human Evaluation: "AI can generate a hundred ideas in seconds, but you still need human judgment to determine which one actually aligns with your brand." — Source: [Config APAC 2024]
7. On the AI Era: "We are shifting from an era of writing code from scratch to an era of editing and reviewing code generated by autonomous systems." — Source: [Malaka Cinematic Podcast]
8. On Automating Data Pulls: "Using language models to convert natural language questions into SQL queries is the next step in fully democratizing data access." — Source: [Semi-Technical Substack]
9. On Technical Intuition: "You don't need a computer science degree to use modern AI tools, but you do need an intuition for how systems pass data to one another." — Source: [Semi-Technical Substack]
10. On Iterative Prompting: "The first response from an AI is rarely the best one; the skill lies in iteratively refining the prompt until the output matches the intent." — Source: [Malaka Cinematic Podcast]

Part 5: Women in STEM & Education

1. On Early Exposure: "Providing young girls with early exposure to coding and robotics fundamentally shifts their perspective on what career paths are available to them." — Source: [KrASIA Interview]
2. On Mentorship: "Being a minority in tech means that finding mentors who look like you is difficult, making intentional mentorship programs a necessity." — Source: [KrASIA Interview]
3. On Building Confidence: "Generation Girl isn't just about teaching syntax; it's about giving young women the confidence to tackle hard technical problems without fear." — Source: [KrASIA Interview]
4. On Bridging the Gap: "Closing the gender gap in technology requires systemic intervention long before women enter the workforce, starting in secondary education." — Source: [Brave Southeast Asia Tech]
5. On Normalizing STEM: "We have to normalize the idea that women are just as capable of being lead engineers and data scientists as their male peers." — Source: [KrASIA Interview]
6. On Safe Environments: "Holiday bootcamps provide a safe environment where girls can experiment, fail, and learn without the pressure of a male-dominated classroom." — Source: [KrASIA Interview]
7. On Future Leaders: "Investing in STEM education for girls today is how we guarantee diverse technical leadership in the tech industry a decade from now." — Source: [CreativeMornings Talk]
8. On Imposter Syndrome: "Many women exit technical roles not because they lack skill, but because persistent imposter syndrome makes them feel they don't belong." — Source: [KrASIA Interview]
9. On Community Support: "A strong community of female peers provides the emotional resilience needed to survive the attrition rates commonly seen in tech careers." — Source: [CreativeMornings Talk]
10. On Giving Back: "Those of us who have succeeded in this industry have an obligation to build ladders for the women coming up behind us." — Source: [KrASIA Interview]

Part 6: Career Trajectory & Angel Investing

1. On Starting at the Bottom: "The reason I was great at Gojek is because I was willing to start at the bottom and do the unglamorous data entry work." — Source: [CreativeMornings Talk]
2. On Parallel Lives: "The math that says parallel lines don't intersect doesn't apply to life; you can be an investor, a non-profit founder, and a tech leader simultaneously." — Source: [CreativeMornings Talk]
3. On Investing in Others: "We should rethink investing entirely in ourselves and instead focus on providing others with the sponsorship we wish we had when we were younger." — Source: [CreativeMornings Talk]
4. On Angel Investing: "Evaluating a startup as an angel investor is less about the current product and entirely about the founder's resilience and ability to pivot." — Source: [Brave Southeast Asia Tech]
5. On Career Progression: "Moving from an individual contributor to a management role requires completely abandoning the metrics that made you successful in the first place." — Source: [Brave Southeast Asia Tech]
6. On Choosing Companies: "When evaluating an opportunity, look at the velocity of the team's shipping cycle rather than the polish of their current application." — Source: [Lenny's Podcast]
7. On Chief of Staff Roles: "Being a Chief of Staff is an exercise in managing influence without direct authority, requiring deep empathy for the executive's blind spots." — Source: [CreativeMornings Talk]
8. On Leaving Comfort Zones: "Staying in a role where you have mastered every process is the fastest way to stunt your long-term career trajectory." — Source: [Brave Southeast Asia Tech]
9. On Portfolio Careers: "Having multiple professional identities prevents you from tying your entire self-worth to the success or failure of a single company." — Source: [CreativeMornings Talk]

Part 7: Leadership & Organizational Debt

1. On Scaling Teams: "Growing a team from thirty to two hundred people requires formalizing communication channels that used to happen naturally over lunch." — Source: [Lenny's Podcast]
2. On Organizational Debt: "Just like technical debt, companies accumulate organizational debt when they implement temporary management structures and forget to dismantle them." — Source: [Brave Southeast Asia Tech]
3. On Hiring Scrappily: "Sometimes the best hire isn't the person with a pristine resume, but the candidate who has a proven history of figuring things out." — Source: [Lenny's Podcast]
4. On Information Silos: "When teams stop sharing data with each other, it is usually a symptom of misaligned incentives rather than a technical limitation." — Source: [Brave Southeast Asia Tech]
5. On Meeting Culture: "A recurring meeting without a clear agenda and a designated decision-maker is a drain on the entire organization's momentum." — Source: [Reforge Unsolicited Feedback]
6. On Firing Fast: "Keeping a toxic high-performer on the team signals to everyone else that company values are secondary to individual output." — Source: [Brave Southeast Asia Tech]
7. On Promoting Internally: "Whenever possible, promoting from within preserves institutional knowledge and demonstrates to the rest of the company that hard work is rewarded." — Source: [Lenny's Podcast]
8. On Managing Up: "Your manager cannot support you if they don't know what you are doing; proactive communication is your responsibility, not theirs." — Source: [CreativeMornings Talk]
9. On Restructuring: "Re-orgs are painful but necessary; fighting against a reorganization often means fighting against the company's natural evolution." — Source: [Brave Southeast Asia Tech]

Part 8: Philosophy & Personal Development

1. On Rationality: "Maintaining a rational mindset in a chaotic startup environment requires aggressively filtering out noise and focusing only on controllable inputs." — Source: [Malaka Cinematic Podcast]
2. On Productivity Systems: "The best productivity system is the one you actually use every day, regardless of how simple or unoptimized it appears to others." — Source: [Malaka Cinematic Podcast]
3. On Continuous Learning: "The moment you believe you have nothing left to learn from junior employees is the moment your expertise begins to decay." — Source: [CreativeMornings Talk]
4. On Embracing Failure: "If none of your initiatives are failing, you are operating too safely and leaving massive opportunities on the table." — Source: [Reforge Unsolicited Feedback]
5. On Setting Boundaries: "Burnout is not a badge of honor; establishing firm boundaries around your time is a requirement for long-term endurance." — Source: [Malaka Cinematic Podcast]
6. On Evaluating Advice: "Advice is heavily influenced by the specific context in which the advisor succeeded; you have to extract the principle, not the exact tactic." — Source: [CreativeMornings Talk]
7. On Handling Criticism: "Constructive criticism hurts the ego but accelerates growth; learning to detach your self-worth from your work makes feedback easier to process." — Source: [Brave Southeast Asia Tech]
8. On Deep Work: "Blocking out uninterrupted time for deep thought is the only way to solve complex architectural problems that can't be rushed." — Source: [Semi-Technical Substack]
9. On Finding Purpose: "Purpose isn't usually found in a grand epiphany; it is built slowly by consistently engaging with problems you find interesting." — Source: [CreativeMornings Talk]
10. On Legacy: "Your legacy in a company isn't the code you wrote, but the frameworks and culture you leave behind for the next team." — Source: [Brave Southeast Asia Tech]