Principal Data Scientist, Upstart

• 8+ years of experience in data science or a related quantitative role, much of it spent measuring marketing or growth spend
• Hands-on experience designing and analyzing incrementality tests, and a solid grasp of causal inference beyond standard A/B testing
• Working knowledge of marketing mix modeling and multi-touch attribution, including where each one breaks down
• Strong Python and SQL skills
• A track record of measurement work that changed how a marketing budget was spent
• Clear communication with technical and non-technical audiences, and the judgment to push back when the evidence doesn't support the plan
• Good judgment about when a result is solid enough to act on and when it needs another test
• Ability to work independently, scope your own projects, and drive them end to end in ambiguous situations

Desirable

• As our Marketing Measurement lead, you'll own how Upstart answers some key questions about its marketing spend.
• What did each channel actually cause? How should credit be shared across the touches a borrower saw before applying? Where should next quarter's budget go?
• The answers decide whether a dollar of spend buys a funded loan or takes credit for one that would have arrived anyway
• Guide incrementality measurement across our marketing channels, working with the analysts who run tests in each channel on design and interpretation, so the team knows what each channel causes rather than what it gets credit for
• Apply causal inference methods when a clean test isn't feasible, so budget decisions still rest on causal evidence
• Build and maintain the attribution approach that splits credit across direct mail, digital, partnerships, and email, including channels where a borrower never clicks anything
• Turn measurement results into spend allocation recommendations that marketing and finance use to plan budgets
• Use marketing mix models alongside experiments and attribution, and work out which to trust when they disagree
• Automate recurring measurement reporting, so your time goes to interpretation and recommendations
• Mentor analysts across the team and review their measurement work