Building a world-class data org | Jessica Lachs (VP of Analytics and Data Science at DoorDash)

Jessica’s background (0s)

• Built one of the largest and most respected data teams in tech.
• Believes analytics is a business impact driving function, not just a service function.
• Focuses on answering "what do we do now that we know this?" instead of just "why".
• Former colleague described her as incredibly good at defining metrics.
• Believes retention is a terrible thing to goal on and prefers short-term metrics that drive long-term output.
• Encourages data scientists to take ownership and do whatever it takes to understand what's happening, even if it means picking up the phone and calling customers.
• Advocates for a centralized org model, where all data-related functions are under one roof.
• When hiring data people, looks for those who are curious, humble, and have a strong desire to learn.
• Believes in picking the right metrics for teams to align incentives and drive the right outcomes.
• Shared examples of how the data team at DoorDash has helped the business make better decisions.
• Emphasizes the importance of experimentation and learning from failures.

Centralized vs. embedded analytics teams (4m59s)

• Analytics should be a business impact driving function, not just a service function, and should have a seat at the table.
• A central Model Center of Excellence with pods that map to different functions is superior to embedded analytics teams within business units.
• A central analytics team should be business partners and thought partners with other teams, sharing the same goals and initiatives.
• The data team should provide data-driven insights and solutions to business problems, identify opportunities for growth, and make recommendations on the most efficient ways to achieve it.

The benefits of a centralized analytics team (10m52s)

• Consistent and high talent bar:
  • Same talent evaluation criteria for all candidates.
  • Higher talent acquisition and retention.
• Growth opportunities:
  • Opportunities to work on new problems in different areas of the company.
  • Potential for career advancement and people management roles.
• Consistency of methodologies and metrics:
  • Standardized definitions and methodologies across teams.
  • Improved methodologies with input from multiple teams.
  • Avoidance of duplicating efforts.
• Scalability:
  • Identification of common problems across teams.
  • Prioritization of automation and improvement efforts.
• A Team culture brand:
  • Strong team culture focused on learning and sharing.
  • Peer support and review.
  • A sense of belonging and pride in being part of the team.

Balancing proactive and reactive work (15m10s)

• To ensure a proactive data team, allocate time for exploratory work and deep dives, and set goals for finding insights through self-directed work.
• Hackathons can be an effective way to carve out time for deep dives and encourage self-directed work.
• A deep dive into the referral channel at DoorDash revealed fraudulent behavior and misleading average metrics, leading to recommendations for better fraud checks and caps on referrals.
• To build a world-class data organization, think long-term, look for opportunities to find big ideas, and align data initiatives with business goals.

Advice on how to push back effectively (20m45s)

• Establish a culture and operating model where junior data folks aren't forced to always say no.
• Set clear goals that are aligned with business partners to make it easier to prioritize tasks.
• Communicate tradeoffs and re-evaluate prioritization regularly with business partners.
• Prioritize and communicate priorities, and align on the tradeoffs of shifting priorities.
• Sometimes do quick tasks to build goodwill, but ruthless prioritization is usually necessary.
• Show your worth by providing value in longer-term projects to justify spending time on them.

Hiring for curiosity and problem solving (24m20s)

• Focuses on curiosity and problem-solving skills beyond technical skills.
• Curiosity is essential and difficult to teach.
• Looks for candidates who are self-motivated to investigate and explore anomalies.
• Tests for curiosity by including something incorrect in case studies and observing if candidates notice and investigate.
• Uses real-world problem-solving cases from DoorDash history to assess problem-solving abilities.
• Values candidates' reactions to being proven wrong and their ability to adapt and make decisions with imperfect information.
• Observes how candidates demonstrate softer skills during case interviews, even if not directly related to the specific problem being solved.

Coming from a non-traditional background (28m57s)

• Jessica Lachs, VP of Analytics and Data Science at DoorDash, has a non-traditional background in data science and became a data scientist out of necessity.
• Lachs' background in finance and focus on driving business impact have helped her successfully lead a team of highly skilled data scientists.
• She emphasizes solving problems and focusing on immediate needs rather than getting overwhelmed by the bigger picture.
• Lachs' motivation to make DoorDash succeed and her willingness to go above and beyond have been key factors in her success.
• The early days of DoorDash required employees to take on tasks outside of their job descriptions, such as taking out the garbage on Saturday nights, to ensure the company's success.
• This culture of rolling up one's sleeves and doing whatever it takes to win has been ingrained in the company by its founder and CEO, Tony Shu, and resonates with Jessica Lachs' own work ethic.

The early days and culture at DoorDash (34m40s)

• DoorDash's early marketing efforts involved grassroots initiatives like distributing promo codes in Boston during winter 2014.
• The company culture emphasizes extreme dream ownership, with everyone participating in customer-centric activities like customer support.
• During a significant outage, the entire company, including new employees, provided customer support to ensure customer satisfaction.
• The "We Dash" program encourages employees to experience the customer perspective by dashing or providing customer support four times a year.
• Jessica Lachs, VP of Analytics and Data Science at DoorDash, emphasizes the importance of catching bugs in the product and ensuring data quality in building a world-class data organization.
• ATO is a powerful, easily configured, and intuitive new type of CRM designed for the next era of companies.
• ATO syncs with data sources, configures to unique structures, and works for any go-to-market motion.
• ATO automatically enriches contacts, syncs email and calendar, provides powerful reports, and allows for quick building of Zapier-style automations.

Encouraging cross-functional roles (40m39s)

• Expects team members to have extreme ownership over outcomes rather than focusing solely on their specific roles.
• Team members are encouraged to go beyond their traditional roles and engage in product management, engineering, and other areas to gain a holistic understanding of the business.
• This approach has led to team members transitioning to different roles within the organization based on their interests and experiences.
• Skilled at defining metrics that align with business objectives, especially in complex environments like DoorDash.
• Emphasizes the importance of finding the right metric to drive the desired incentive, particularly in messy business situations.

Defining effective metrics (44m39s)

• Focus on short-term metrics that drive long-term output.
• Experiment and iterate to find the right inputs that drive retention.
• Keep metrics simple and easy to understand.
• Avoid composite metrics that are hard to interpret.
• Choose metrics that people across the company can talk about and understand.

Simplifying metrics for better outcomes (46m30s)

• Quantify all business levers in common terms (e.g., gross order value and volume) to facilitate tradeoffs between teams and make quicker decisions.
• Prioritize metrics based on their impact on the business and focus on improving the most important ones first.
• Use simple, understandable metrics that are close enough to the ideal perfect metric, even if it means having multiple metrics.
• Avoid overly complicated metrics and composite metrics that are difficult to understand and move.
• Stay focused on a metric until there are no more opportunities for improvement before moving on to another one.

Focusing on edge cases and fail states (55m28s)

• Look at the average, but also focus on edge cases and fail states.
• Set goals and create metrics around edge cases.
• Prioritize eliminating fail states as they can have a significant impact on the business.
• Understand why fail states happen and find ways to prevent or fix them.
• Don't ignore fail states because they happen infrequently.
• Fail states can lead to churn and lost revenue.
• Some fail states may not be visible in the data, such as login errors.

Managing a global data organization (1h0m12s)

• Similarities between data scientists, consumers, Dashers, and careers across different countries.
• Differences in managing data people in different countries include:
  • Different currencies and languages add complexity.
  • Different regulations in EU countries vs. non-EU countries in Europe.
• Despite these complexities, many of the problems are the same.
• The speaker draws on experiences from Volt Analytics to approach problems in new countries.
• Excitement in exploring new ideas that might work in one country but not in another.
• Focus on similarities with pleasant surprises when finding differences.

Leveraging AI for productivity (1h2m31s)

• DoorDash has an initiative called "office hours" where the analytics team provides support to other teams.
• They are developing AI tools to help non-technical users edit queries and access data independently, reducing the workload of the analytics team.
• The chatbot that provides this assistance is called "ask data AI," named after their internal Slack channel for data-related questions.
• Jessica Lachs, VP of Analytics and Data Science at DoorDash, discusses the importance of building a world-class data organization.
• Key elements include:
  • Having a clear vision and mission for the data organization.
  • Building a strong team with diverse skills and backgrounds.
  • Creating a culture of collaboration and innovation.
  • Investing in data infrastructure and technology.
  • Ensuring data literacy across the organization.
  • Measuring success and continuously iterating.

Building diverse and skilled data teams (1h5m25s)

• Jessica Lachs emphasizes the importance of diverse backgrounds and skill sets when building a data team.
• She highlights that formal training is not always necessary and values individuals with transferable skills who are eager to transition into analytics.
• Lachs appreciates the unique perspectives and expertise that team members with different backgrounds bring, fostering a collaborative learning environment.
• She seeks individuals with experience at different sized companies, valuing both the hustle of startup employees and the insights of those who have witnessed large-scale operations.
• Lachs believes that a diverse team, encompassing a variety of skills, backgrounds, and prior company experiences, leads to a well-rounded and effective data organization.

Lightning round (1h8m40s)

• Jessica Lachs recommends the Libby app and supporting local public libraries.
• She prefers rewatching classic TV shows like "The West Wing" and "Alias" to recent releases.
• Jessica highly recommends Korean sunscreens, especially the Beauty of Joseon brand.
• Her favorite life motto is a quote by John Steinbeck about finding solutions after a good night's sleep.
• Jessica acknowledges the influence of several remarkable women in her career, including Vanessa Roberts, Gina Terrone, Tia Sheringham, and Liz Jarvisen.
• She joined DoorDash for the interesting challenges and people, without much thought about its potential success.
• Two significant moments for her were DoorDash becoming the number one player in the third-party market share and seeing widespread recognition of the company.
• Jessica encourages people to follow her on LinkedIn for insights into building a global analytics organization.
• She appreciates honest feedback but requests that it be delivered with kindness.
• Jessica Lachs emphasizes the importance of seeking and speaking the truth in a world filled with misinformation.
• The podcast episode is available on Apple Podcasts, Spotify, and other platforms.