Every company I've seen struggle with marketing has the same problem: they're measuring too many things.

They have dashboards with thirty graphs. They track impressions, clicks, CTR, bounce rate, time on page, scroll depth, email open rates, social engagement, MQLs, SQLs, pipeline velocity, and customer acquisition cost. They have a weekly meeting where someone presents slides about all of them. Nothing changes.

The companies that grow fast do something different. They pick one number and obsess over it.

At Facebook in the early days, the number was monthly active users. Not revenue. Not engagement time. Not number of photos shared. Monthly active users. Every decision ran through that filter. Should we launch in another country? Does it increase MAU? Should we build a photo tagging feature? Does it increase MAU?

This seems reductive. It is reductive. That's the point.

When you have thirty metrics, you can always find one that went up this week. That feels like progress. It's not. It's the numerical equivalent of rearranging deck chairs. (This is also why you shouldn't build a funnel — funnels tempt you into optimizing stages instead of outcomes.) Real progress means one important number going up consistently over months.

The hard part isn't picking the number. It's committing to it. Because once you pick one metric, you have to ignore the others. And ignoring data feels irresponsible. What if bounce rate is going up? What if our social engagement is dropping?

It doesn't matter. Not because those things are irrelevant, but because you can't optimize for everything at once. A company that tries to improve thirty metrics improves none of them. A company that focuses on one metric actually moves it.

So which metric should you pick? The one that most directly measures whether people are getting value from your product. For a marketplace, it's transactions. For a SaaS tool, it's weekly active users. For a media company, it's subscribers. Not visitors. Not signups. Not "awareness." The number that means someone used your thing and got something out of it.

Revenue is usually the wrong choice, by the way. Revenue is a trailing indicator. By the time revenue drops, the real problem happened months ago. I've seen this with retention — companies stare at churn numbers for months before realizing the issue was upstream. Usage metrics are leading indicators. They tell you about tomorrow.

I worked with a startup that tracked everything. Gorgeous dashboards. Terrible growth. We deleted all but one metric: weekly users who completed their core action (creating a report). Within three months, the whole company was aligned. Engineers built features that helped people create reports. Marketing targeted people who needed reports. Sales led with the reporting demo.

Before, everyone was busy. After, everyone was effective. The difference was focus.

One metric. Pin it to the wall. Talk about it every day. Make it go up.

Sasha Volkov ran a useful experiment applying this framework to AI tool selection. Her finding — that a three-tool stack focused on one metric outperformed a seven-tool stack optimizing for many — is exactly what I'd expect. The tools don't matter until the focus does.