AI-Native GTM

What Gets Measured Gets Mismanaged

The metric that was supposed to create discipline quietly becomes the thing everyone games. Here's how activity targets eat the outcomes they were meant to produce.

Kevin OLeary··5 min read

Every revenue leader inherits a dashboard. Calls made, emails sent, meetings booked, activities logged. The numbers feel like control. You can see them move, you can set targets against them, and when they go up you feel like something is working. That feeling is the trap.

Here's what actually happens when you put a rep on sixty calls a day. They make sixty calls. Not sixty good calls, not sixty calls to the right accounts, sixty calls. The target was a proxy for the thing you wanted, which was pipeline, and the moment you made the proxy the goal, the rep started optimizing the proxy instead of the pipeline.

Goodhart's Law, on your sales floor

When a measure becomes a target, it stops being a good measure. Economists have a name for this. On a sales floor it looks like reps sandbagging easy calls to hit a number, logging activities that never happened, or burning your best accounts on rushed outreach because the clock said dial now. The metric didn't measure discipline. It manufactured theater.

The target was a proxy for the thing you wanted. The moment you made the proxy the goal, everyone started optimizing the proxy.

What we did instead

At a company I ran demand gen for, I inherited ten reps living under activity quotas. The first thing I did was delete them. The new goal was two qualified meetings a week, and I did not care how anyone got there. Two emails that turned into two conversations was a perfect week. The reps who used to hide behind their call count had nowhere left to hide, and the ones who were actually good got the freedom to be good.

That quarter was the best the team had ever had. When leadership eventually reinstated activity metrics, production dropped. The lesson was permanent: measure the outcome you actually want, and trust people to find the path. If you measure the path, you will get the path, and the outcome will quietly walk out the back door.

The AI-native version

This gets sharper with AI in the loop. When a machine can generate a thousand activities an hour, activity is worthless as a signal. The only things worth measuring are the ones that resist gaming: qualified conversations, new pipeline generated against last period, deals that close. Build your system to watch those, let the machine handle the volume beneath them, and never again confuse motion for progress.

Kevin OLeary

Written by

Kevin OLeary

Founder of Nicl. Twenty years in go-to-market. Writes about founder-led sales, AI-native GTM, and the referral co-op.

More from Kevin →