Metrics question

Facebook is running a test where they hide the like/comment/share bar of each Newsfeed post in the ellipses menu (3 dots in the corner) to increase the density of the feed. What metrics do you expect to move? How do you decide to fully roll out the feature?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Tests anticipating metric movement from a specific UI change and defining a clear rollout decision rule ahead of running the test.

Predict likely movements: feed density (posts seen per session) should rise since less vertical space is used per post, and overall like/comment/share rate could drop since the affordance is less visible.
Predict a secondary effect: session length or posts-per-session may rise from higher density, but meaningful interactions per session could fall if engagement requires an extra tap to find the bar.
Segment the expected impact: power users familiar with the UI may adapt quickly, while casual or older users may interact less simply from reduced discoverability.
Propose the rollout decision rule upfront: fully roll out only if meaningful interactions per session stay flat or improve, even if raw density rises, since interaction quality is the real goal, not density alone.
Add a guardrail: monitor for a drop in comment or share rate specifically, since those carry more product value than passive likes and are more likely hurt by reduced visibility.

Predicts a nuanced, sometimes counter-intuitive tradeoff (density up, but interaction rate potentially down) rather than assuming all metrics simply improve.
Sets the actual rollout decision criteria before the test even runs, meaningful interactions per session, not raw feed density, must hold or improve.
Uses an illustrative number, e.g. assume density in-session rises 10% but interaction rate per post drops 8%, netting roughly flat total interactions, which would be a borderline call.
Distinguishes between high-value interactions (comments, shares) and low-value ones (likes), recognizing they shouldn't be weighted equally in the rollout decision.

Assuming the change is purely positive without predicting the likely interaction-rate tradeoff from reduced visibility.
Not defining a clear go/no-go decision rule before running the test, leaving the rollout call subjective.

What would you do if density rose but comments specifically dropped sharply?
How would you test whether the drop in interaction is from reduced visibility versus reduced genuine intent to engage?