Vanity Metrics Are Comfort Food

There’s a particular kind of number that makes a bad week feel survivable. Total signups. Registered users. Impressions. App-store ranking. You open the dashboard, the line is up and to the right, and something in your chest unclenches. That relief is the tell. A metric that exists mainly to make you feel better is doing the job of comfort food — it’s warm, it goes down easy, and it quietly leaves you starving.

The trouble isn’t that these numbers are fake. Signups are real people who really typed a password. The trouble is what they can’t do: they can’t go down. Cumulative counts only rise, so they always tell a story of progress, even during the exact weeks your product is dying. A number that’s structurally incapable of delivering bad news isn’t measuring your business. It’s flattering it.

So how do you tell a comfort-food metric from one with nutrition in it? I’ve found three questions that cut through almost every time.

First: can this number get worse? A metric that can only climb — total downloads, all-time users, lifetime pageviews — is a trophy, not an instrument. Real instruments move in both directions. Weekly active users can fall. Retention can slip. Conversion can crater. The willingness to go down is what makes a number honest, because it means the number is actually coupled to reality instead of just accumulating history.

Second: if this goes up, does it force a decision? Useful metrics are annoying. They point at something and demand you do a specific thing next. If your week-four retention drops, you know roughly where to look — onboarding, the core loop, the promise you made versus the thing you delivered. But if “total impressions” doubles, what do you do? Nothing. There’s no action on the other end of the number, which means it was never really information. It was scenery.

Third: would this survive being turned into a rate? This is the sharpest one. Almost every vanity metric is a raw count that looks impressive because it’s absolute. Convert it into a ratio — per user, per week, per dollar spent — and watch what happens. “One million signups” feels like a company. “One million signups, of whom eight thousand came back this month” feels like a warning. The absolute number was hiding the denominator, and the denominator was where the truth lived.

The reason this matters isn’t accounting hygiene. It’s that the metrics you watch slowly become the product you build. Stare at signups long enough and you’ll start optimizing the top of the thing — louder launches, more channels, a landing page that converts strangers a hair better — while the reason people don’t stick goes unexamined, because your dashboard was structurally unable to mention it. You’ll be busy, the graph will be climbing, and the ground will be washing out beneath you the whole time.

None of this means counts are useless. Absolute numbers are fine for context and terrible as a compass. The move is to keep one or two honest metrics on the wall — ones that can fall, that force a decision, that survive becoming a rate — and treat the comfort-food numbers as exactly what they are: pleasant to look at, safe to ignore, and no substitute for a meal.

The question I’d sit with isn’t “are our numbers good?” Good numbers are easy to find; you can always locate a graph that’s going up. The question is “which of our numbers is telling us something we don’t want to hear?” That one’s worth more than all the trophies on the shelf.

The idea, to look up: “vanity metrics” (popularized in the Lean Startup literature), and the general distinction between vanity and actionable metrics.