Select and prioritize metrics with guardrails

The new Groups Stories feature is intended to drive deeper, more meaningful engagement while avoiding damage to the wider ecosystem. Design a measurement framework and choose the smallest set of launch gates:

1. identify three primary goal metrics (success metrics) and explain why each is normalized by DAU, viewer, or poster;
2. list metrics that exist only in the feature (treatment-only CTR or creation rate) to estimate the opportunity if it ships;
3. include ecosystem metrics to detect downstream effects, such as sessions, retention, and revenue;
4. add explicit cannibalization checks, specifically the impact on Feed, Marketplace, and Pages;
5. set guardrails (latency, crash rate, report rate, cancellations) and sanity checks;
6. choose a multiple-testing correction (for example, Holm-Bonferroni versus BH) and state when you would favor each;
7. demonstrate a rough, back-of-the-envelope estimate of the LTV and ARPDAU impact from a 1% increase in DAU. Rank the three metrics that must pass before a 100% rollout, and explain why the remaining metrics are secondary. Overview: This question tests whether a candidate can build a sound metrics framework for product experiments, including choosing and normalizing success metrics, sizing with feature-only metrics, tracking ecosystem effects, reading cannibalization explicitly, setting guardrails, controlling for multiple testing, and estimating LTV/ARPDAU impact with rough arithmetic.