Conversion in Marketplace fell after a redesign. How would you investigate

Problem Statement Description

Product context: Meta is a social technology company; its products include Facebook, Instagram, WhatsApp, Messenger, Threads, Quest, creator tools, and ads.

Meta Marketplace has recently launched a redesigned experience, and the team is seeing a decline in conversion after the change. Assume this is a meaningful product concern for Marketplace users, especially families who may use Marketplace to discover, compare, message about, and purchase everyday items such as furniture, baby products, household goods, vehicles, or local services.

Your task is to walk through how you would investigate the conversion drop as the product manager responsible for understanding what happened. The focus is not to jump to a redesign rollback or propose a new feature immediately, but to diagnose whether the issue is real, where in the funnel it is happening, which users or surfaces are affected, and what evidence would help isolate the root cause.

You should consider the full Marketplace workflow: entering Marketplace, searching or browsing listings, viewing item details, evaluating seller trust, saving or sharing items, messaging sellers, arranging pickup or payment, and completing the intended transaction. The redesign may have affected user comprehension, trust, ranking, navigation, performance, messaging intent, or measurement itself.

The investigation should consider:

- How to define “conversion” precisely, including numerator, denominator, funnel stage, time window, and whether it refers to buyer intent, messages sent, transactions completed, or another Marketplace outcome.

- How to validate that the drop is real, including experiment setup, logging changes, tracking gaps, seasonality, traffic mix shifts, and statistical significance.

- Which segments to compare, such as new vs. returning users, families, geographies, device types, app versions, acquisition sources, listing categories, price bands, and buyer vs. seller behavior.

- Where the funnel changed most after the redesign, including impressions, clicks, listing views, saves, seller profile views, message starts, message replies, and completed exchanges.

- What hypotheses could explain the decline, such as reduced listing discoverability, lower trust signals, confusing UI, slower load times, ranking changes, fewer relevant listings, safety concerns, or messaging friction.

- What evidence you would seek from product analytics, experiment data, user research, session replays, customer support, seller feedback, performance monitoring, and integrity/safety systems.

- How you would prioritize mitigation options while balancing user experience, seller liquidity, safety at scale, ads or commerce objectives, and the risks of reverting or partially rolling back the redesign.

- How you would prevent recurrence through better launch monitoring, guardrail metrics, phased rollout criteria, instrumentation reviews, and post-launch decision gates.

The goal is to demonstrate a structured root-cause investigation that separates measurement issues from true user behavior changes, identifies the most likely drivers of the conversion decline, and leads the team toward an evidence-based product decision.