Estimate infrastructure or operational capacity needed for Ads

Problem Statement Description

Product context: Uber is a mobility and delivery platform; its products include rides, Uber Eats, grocery and retail delivery, freight, driver and courier tools, and marketplace pricing.

Uber is exploring the infrastructure or operational capacity required to support Ads across its marketplace, with particular attention to driver-facing surfaces and ad delivery within a global mobility, delivery, and logistics ecosystem. In this guesstimate, you are expected to size the capacity needed to run an Ads product reliably at scale, not to design the ad product itself.

Assume Ads could appear in relevant Uber experiences where drivers interact with the platform, such as in-app screens, trip-related workflows, earnings or opportunity surfaces, or other marketplace-adjacent placements. The estimate should account for Uber’s large, geographically distributed user base, variable demand by city and time of day, and the need to avoid disrupting core marketplace reliability, safety, and driver productivity.

You may define the scope of “capacity” clearly: for example, ad impressions served, requests per second, storage, compute, moderation or sales operations, campaign support capacity, or a combination of technical and operational needs. The interviewer is looking for a structured, transparent estimation approach with reasonable assumptions, not a precise factual number.

The estimate should consider:

- The user population in scope, such as active drivers, delivery couriers, or regional subsets

- The relevant unit of capacity, such as ad requests, impressions, campaigns, support tickets, moderation reviews, or compute/storage needs

- Frequency assumptions, including driver app sessions, screen views, trips, delivery tasks, and peak-hour behavior

- Adoption and rollout assumptions, such as pilot cities, global launch, or phased marketplace coverage

- Segmentation by geography, business line, driver activity level, and marketplace maturity

- Peak-versus-average load, seasonality, special events, and city-level demand spikes

- Operational dependencies, including advertiser onboarding, campaign review, fraud prevention, support, and local compliance

- Sanity checks against marketplace scale, driver experience constraints, and reliability expectations

Your goal is to produce a defensible capacity estimate with clear assumptions, simple math, sensitivity ranges, and a final order-of-magnitude conclusion that could help Uber decide what infrastructure or operational readiness is needed before scaling Ads.