Company Context

Andreessen Horowitz operates a portfolio support platform for founders, including a16z speedrun, network programming, and software surfaces that help portfolio companies access talent, experts, and operator resources. One internal product team recently launched a new founder-facing feature inside the a16z portfolio platform to improve engagement with expert matching and resource discovery.

Problem

Six weeks after launch, the dashboard suggests the feature is performing well: weekly active usage among exposed founders is up 18%, repeat usage is up 11%, and session depth on the feature has increased from 2.1 to 3.4 actions per visit. Based on these numbers, the team is preparing to expand the feature to all eligible users.

However, qualitative feedback is negative. In founder interviews and support tickets, users describe the feature as "confusing," "too pushy," and "not actually helping me get to the right expert faster." Several high-value founders say they are using it because they feel they have to, not because it solves their problem well. The PM asks you, as a Product Growth Analyst candidate, how you would respond.

Deliverables

1. Diagnose how both signals could be true at the same time, and identify the most likely explanations.
2. Define how you would segment users, behavior, and feedback to determine whether the feature is truly creating value.
3. Recommend what action you would take next: scale, iterate, limit rollout, or roll back.
4. Propose the additional research and analysis you would run over the next 2-4 weeks.
5. Define the decision criteria you would use to resolve the conflict between quantitative and qualitative evidence.

Constraints

• You have one PM, one data analyst, one designer, and limited engineering support for the next month.
• The feature is already live to 35% of eligible founders.
• Leadership wants a recommendation within 2 weeks on whether to expand rollout.
• You cannot rely on building a brand-new measurement system before making the decision.