Doubling the Channel Halved It

At one company, marketing spend rose 40% over two quarters while pipeline rose 60%. The growth team's QBR called it a doubled channel. The CEO cited it in three board meetings. The marketing lead got promoted.

Six months later, the cohort converted at a third of the previous baseline. The channel hadn't doubled. It had doubled the wrong customers.

This pattern is common at growth-stage companies, and it survives because pipeline leads conversion by a quarter or two. Marketing shows pipeline growth in Q1. Sales works the new pipeline in Q2. The conversion data lands in Q3. By the time the number is honest, the team has been scaling spend for three quarters against a metric that was already wrong.

The logic is plain. A channel performing well at a given spend has a profile of who responds. Doubling the spend rarely doubles that profile. It pulls in lower-intent traffic: broader keywords, looser targeting, weaker creative against weaker audiences. The dashboard records the lead. It doesn't yet know the lead won't close.

What catches it is segmentation at intake. Score every new lead on ICP fit against a defined set: company size, role, industry, observed intent. Report pipeline in two columns, in-ICP and out-of-ICP. Grade the growth team on in-ICP growth, not the total. Show the out-of-ICP column for honesty, but don't celebrate it. This is the same drift behind the customers you're winning not matching the ones you're selling to.

That one habit kills the doubled-channel story before it gets a quarter to scale. If in-ICP pipeline is flat while out-of-ICP grew sixty percent, the channel didn't double. It broadened. Broadening is sometimes right; lower-intent leads can nurture into customers over a longer cycle. But it's a different operating decision than scaling a high-converting channel, and the dashboard should say which one happened. Without attribution tying leads to closed revenue you can't tell, which is its own failure mode.

Most growth teams resist this because the in-ICP number is smaller and looks less impressive in a board update. The smaller number is also accurate. Run for a year, it produces a channel mix where each line does what the team thinks it does. The flattering number, run for a year, produces a mix where one or two lines are quietly poisoned.

A team that celebrates pipeline growth without conversion segmentation is celebrating the first half of a transaction. The channel doubled. The transaction is still open. The close rate, which decides whether the channel worked, lands in six months. By then the budget is already raised and the cohort decay gets explained away as seasonality.