Vanity metrics are a tax on bad decisions

A few years ago I inherited an account where every number on the dashboard was green. Impressions up. Clicks up. ROAS up. The client was happy. And the business was quietly losing money on every new customer it acquired.

That gap between a dashboard that looks healthy and a business that isn’t is what vanity metrics cost you. They flatter, and they also actively misdirect. Every hour spent optimizing toward a number that doesn’t connect to revenue is an hour of budget pointed slightly wrong, and those degrees compound.

Here’s the problem with the usual headline metrics.

Impressions and reach measure how many people the platform was willing to show your ad to, which mostly reflects your budget rather than your performance. Clicks measure curiosity more than intent. And ROAS, the one most people trust, is the sneakiest, because it’s a ratio that hides its own denominator. A 5x ROAS on a campaign that only ever reaches people who’d have bought anyway is worse than a 3x ROAS that brings in genuinely new customers. On its own, ROAS rewards you for taking credit even when you created no demand.

So what do I actually optimize to?

CPA and CAC come first. That means the real cost to acquire a customer, loaded with the full spend, instead of the platform’s self-reported version. Then pipeline contribution for B2B, meaning leads that became opportunities that became revenue, traced back to the channel that sourced them. And blended efficiency across the whole account, because a single channel can look brilliant while cannibalizing another that was already going to convert.

The test I use is simple, almost annoyingly so. Would this number change a decision? If a metric goes up and I’d do nothing differently, it’s decoration. If it goes up and I’d confidently move budget toward it, it’s real. Most dashboards are 80% decoration and 20% signal, and the work is knowing which is which.

None of this means ignoring the top-of-funnel numbers entirely. They’re useful as diagnostics, the way a fever tells you something’s wrong without telling you what. The mistake is promoting a diagnostic to a goal. Clicks can tell you your creative is working, but they can’t tell you your business is.

The discipline that follows from this is unglamorous. You report fewer numbers. You sometimes show a client a dashboard with a metric that went down and explain why that’s good. You make decisions that look slower because they wait for a revenue signal instead of a click signal. But the account performs better where it counts, and “where it counts” is the only place that survives a CFO asking what the marketing budget actually returned.

Measurement should be built for decisions. It’s the line I keep coming back to, because the alternative, measurement built to feel good, is one of the most expensive habits in this industry.