How would you improve reliability and latency 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 Ads operates inside time-sensitive marketplace flows where riders, eaters, couriers, drivers, merchants, and advertisers expect fast, dependable experiences. In this interview, you are asked to think like a Technical PM responsible for improving the reliability and latency of Ads delivery, especially where ad serving can affect driver-facing workflows, trip flow, marketplace liquidity, and overall app responsiveness.

Assume Ads may be shown across high-traffic surfaces such as ride booking, delivery ordering, post-trip, in-app driver surfaces, or sponsored marketplace placements. The system must make decisions quickly, fetch eligible campaigns, respect targeting and privacy constraints, track impressions/clicks/conversions, and fail gracefully without disrupting core Uber tasks like accepting trips, navigation, earnings visibility, or order fulfillment.

Your task is to frame how you would approach improving both reliability and latency for this product area. The focus is not on designing an ad auction from scratch, but on defining the technical product requirements, trade-offs, system behaviors, rollout approach, and measurement plan needed to make Ads performant and trustworthy at Uber scale.

The experience should consider:

- User and business-critical workflows affected by Ads latency, including driver app responsiveness, trip acceptance, navigation, rider/eater conversion flows, and advertiser delivery expectations.

- Reliability requirements such as uptime, graceful degradation, fallback behavior, retry strategy, dependency isolation, and avoiding impact on core marketplace operations.

- Latency requirements across client, network, backend services, targeting, ranking, auction, creative retrieval, tracking, and rendering.

- APIs and data dependencies needed for ad eligibility, targeting, pacing, frequency capping, campaign budget enforcement, and event logging.

- Privacy, safety, and policy constraints around user data, location, driver context, and advertiser targeting.

- Observability needs, including service-level indicators, tracing, alerting, dashboarding, error budgets, and incident diagnosis.

- Rollout and experimentation considerations across geographies, app versions, surfaces, network conditions, and marketplace states.

- Product trade-offs between monetization, user experience, driver trust, marketplace reliability, advertiser outcomes, and engineering complexity.

The goal is to demonstrate how you would translate a broad reliability and latency problem into a concrete Technical PM plan: what you would measure, where you would investigate, what requirements you would set, how you would partner with engineering and data teams, and how you would ensure Ads improve without compromising Uber’s core marketplace experience.