Uber — Brand Science — Product Sense & Metrics

Role context: Applied Scientist, Brand Science · Est. study time: 35 min · 4 practice questions

How to prepare for this role

Uber Brand Science is judged on measuring marketing incrementality (not attribution) with experiments as ground truth, building MMMs calibrated by those experiments, connecting brand KPIs to trips causally, and allocating budget on marginal ROI — all while respecting that Uber is a marketplace where marketing drives demand that supply must fulfill. The interview stresses experimentation, causal inference, and the multi-sided marketplace (switchback, synthetic control).

The day-to-day is proving and improving how a large brand budget (TV, billboards, digital) drives incremental trips and gross bookings, via MMM, geo/incrementality experiments, and budget optimization. The instinct that separates strong candidates: refusing to judge brand marketing on brand-lift surveys or attributed conversions, and always asking whether the demand a campaign drives can actually be fulfilled by the supply side.

Where to spend your prep time:

  • Marketplace + marketing product sense (this article) — the multi-sided marketplace, gross bookings, and why incremental beats attributed.
  • Incrementality and causal inference — geo experiments, synthetic control, MMM calibrated by experiments, switchback for marketplace features. See the A/B Test & Causal Inference section.
  • SQL, Python, and communication — be fluent, and practice explaining an incrementality result and a budget recommendation to marketing and leadership.

The through-line: measure incremental gross bookings per marketing dollar, calibrate MMM with experiments, and respect the marketplace supply constraint.

What Uber Brand Science actually is

Uber is a multi-sided marketplace: Rides (riders ↔ drivers) and Eats (eaters ↔ couriers ↔ restaurants), matched by Uber and monetized by a take rate (plus Eats ads). The identity to carry into every answer: a marketplace whose North Star is gross bookings, where Brand Science's job is to prove and improve how a large brand budget drives incremental trips — while respecting that marketing drives demand the supply side must fulfill.

Two things make this different from a normal marketing-analytics job:

  • Marketing value is incremental, not attributed. Brand marketing is upper-funnel and easy to over-credit — many trips would have happened anyway. The job is measuring the incremental trips/gross bookings marketing caused, using experiments as the ground truth (geo lift, synthetic control), not last-touch attribution or a brand-lift survey alone.
  • It's a marketplace, so demand isn't free. Brand marketing drives demand (riders/eaters), but a trip only happens if a driver/courier fulfills it. In a supply-constrained market, more demand raises surge and wait times rather than net trips — so marketing's true effect is a marketplace effect, not a demand lift.