Facebook · Metrics · Senior · 45 to 60 minutes

Should Facebook test richer post feedback?

Facebook is a social network where people share posts, photos, and videos and interact with content from friends, creators, and pages. Answer this product management case and get it graded line by line against a weighted rubric.

Experiment design
Metrics strategy
Ecosystem thinking
Tradeoff judgment
Executive communication

The situation

Facebook leadership asks you to recommend how to evaluate a possible change to the main feedback control on feed posts. Users often say the Like button is too limited, especially for sad, frustrating, or upsetting updates, and some people inside the company want a simple Dislike option while others want a broader set of lightweight responses. The idea could increase expression and make feed interactions feel more honest, but it could also change the tone of the platform, affect how comfortable people feel posting, and create downstream ranking or moderation issues. Because this control sits on one of Facebook's highest traffic surfaces, a bad launch could have company wide consequences. You are the PM responsible for defining the test plan and the criteria for a launch decision before any experiment starts. Your job is not to pick the final UI in this exercise. It is to specify how Facebook should test the idea, what metrics matter most, what guardrails must be in place, and how leadership should interpret mixed results.

What to produce

A 1 to 2 page experiment strategy memo.

Your answer must

State a clear product hypothesis and at least one counter risk hypothesis before listing metrics.
Define one primary success metric and 3 to 5 guardrail metrics, including at least one creator side metric tied to posting or sharing behavior.
Outline the randomized experiment design, including variants, target population, segmentation, run length, and how you would account for novelty effects.
Explain how you would interpret at least two result patterns, including one where viewer engagement improves but creator activity worsens.
Give a launch recommendation rule and a follow up plan if the first test is inconclusive.

Constraints

Scope only the Facebook feed post feedback control, not comments, messaging, Groups, or other apps.
Base the decision on metrics that can be observed in a short product experiment, not on a long policy, brand, or societal study.
Do not assume new moderation systems, new ranking models, or custom survey programs beyond normal product analytics.

Format

Use sections: hypothesis, metrics, experiment design, readout logic, recommendation rule.

How Facebook actually played it

Uses the story of Facebook declining a Dislike button and shipping Reactions to teach experiment design. It builds a hypothesis, picks primary and guardrail metrics such as content creation rates, and discusses how to read results.

Write the hypothesis and the guardrails before launching the test
Watch creator-side metrics when you change what viewers can do
A long-requested feature can still carry ecosystem risks that a test must cover

Built from A/B Testing PM Interview: The Facebook Reactions Rollout Decision by Ajitesh, Tough Tongue AI (PM Interview Prep Weekly).