On analytics, inherited thinking, and what gets measured
Most product teams inherit metrics like they inherit codebases: with respect, acceptance, and the assumption that they're right. A metric that's been in place for years, that leadership trusts, that everyone has optimized around — it carries the weight of institutional knowledge. But sometimes, that weight is just inertia.
A metric that measures frequency can hide a product that's creating compulsion, not value. A metric that measures reach can hide a product that's lost trust. A metric that measures usage can hide that users have no better alternative. The danger of inherited metrics isn't that they're wrong — it's that they're specific. They measure something real. But they measure one thing, and the one thing they measure creates a tunnel vision around what success means.
Ask yourself: If this metric disappeared tomorrow, would our product get better or worse? If the honest answer is "better," then you've found a metric that's optimizing for something that isn't actually serving your users. It's optimizing for you. Second question: Can I explain to a user — in one sentence, without jargon — why this metric matters to them? If I can't, then I'm not measuring something that actually creates value in their life. I'm measuring something that looks good in a dashboard.
The hardest work in product isn't building features. It's asking whether you're measuring the right thing. Because if your metric is wrong, then every optimization you make is moving you further away from what actually matters. When the metric is the problem →