Facebook · Metrics · Mid · 45 to 60 minutes

Pick Facebook's activation metric

Facebook is an online social network where people connect with friends, share updates, and communicate. Answer this product management case and get it graded line by line against a weighted rubric.

Metric design
Retention analysis
Experimentation
Strategic communication
Trade off judgment

The situation

Facebook's growth leadership wants one simple metric that tells the company when a new user has found enough value to be likely to come back. Right now, different teams are pointing to different early behaviors, like adding friends, sending messages, or simply logging in often, and there is no shared target for onboarding or growth work. That is creating confusion about what to optimize and how to judge success. The VP of Growth asks you to recommend how Facebook should identify and validate a single activation target for new users. The hard part is that the most predictive behavior may be too rare to guide product work, while the most common behavior may be too weak to mean much. Assume you have access to recent signup cohorts, first 14 day event logs for core actions such as friendship creation, messages sent, and logins, plus a 30 day retention label for each user. You need to decide what analysis to run, how to choose between candidate actions, and how to test whether a chosen behavior truly drives retention rather than just tagging users who were already likely to stay.

What to produce

Write a one page product memo, 700 to 900 words.

Your answer must

Define the retention outcome, cohort, and early usage window you would use.
Name the first set of user actions you would analyze and explain why they are plausible signals of value for Facebook.
Describe an analysis method that compares retained and lost users and avoids picking a threshold that is highly predictive but too rare.
Recommend how leadership should choose a simple activation target from the analysis, including whether to use a rounded heuristic or a more exact threshold.
Design one experiment that tests whether the chosen action causes higher retention, including primary metrics, guardrails, and what result would falsify the hypothesis.

Constraints

Use only data Facebook already logs in user and event tables.
Your recommendation must result in one shared metric that non analysts can remember and use this quarter.
Do not assume a full product redesign, only modest prompts or onboarding changes can be tested first.

Format

Use sections titled Goal, Analysis plan, Metric choice, Experiment, and Risks.

How Facebook actually played it

Argues that famous activation numbers are rounded rallying points, not exact laws. It shows how to find the actions that retained users take and test whether pushing them raises retention.

Compare retained and lost users to find candidate actions
Test causation by nudging users toward the action
Treat a round target as a heuristic

Built from Facebook's Aha Moment Was Simpler Than You Think by Benn Stancil.