Building a world-class data org

Jessica Lachs, VP of Analytics and Data Science at DoorDash, shares how she built and scaled one of tech's most respected data organizations. The conversation covers why a centralized analytics model works, how to hire and prioritize data work, and how to pick metrics that align incentives across a complex, multi-sided marketplace.

01Key learnings

• Treat analytics as a business-impact function with a seat at the table, answering 'so what do we do now?' rather than only building dashboards or answering tickets.
• A central analytics org with pods mapped to partner teams keeps talent bars consistent, creates growth paths, and avoids duplicated models and metric definitions.
• Carve out intentional time for self-directed deep dives, such as hackathons, because exploratory work is the first thing to slip when inbound requests pile up.
• When pushed to take on new asks, make the trade-offs explicit with stakeholders and let shared goals drive which work gets prioritized.
• Hire for curiosity and self-motivation, testing with a real business case that has an intentional flaw to see if candidates notice and pursue it.
• Choose short-term proxy metrics that drive long-term outcomes, since metrics like retention are hard to move quickly; keep composite metrics simple enough that teams understand them.
• Build a common currency across levers like price, selection, and delivery time so teams can compare trade-offs and allocate investment quickly.
• Track edge-case failure states such as never-delivered orders, which are rare but cause churn and high cost, and set explicit goals to eliminate them.

02Frameworks mentioned

Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.