TheIncrementalityTestNobodyWantstoRun

Here's a simple question: if you turned off your Google Ads tomorrow, how much revenue would you lose?

Not "what does Google attribute to your ads." How much revenue would genuinely disappear?

Nobody wants to answer this question. And that's exactly why it matters.

An incrementality test - a properly controlled geo-split or holdout experiment - reveals the true causal impact of your ad spend. Not the correlated impact. Not the attributed impact. The actual, would-not-have-happened-otherwise impact.

Agencies avoid this test because the answer is usually smaller than the attributed number. Sometimes significantly smaller. Brand campaigns often show 20-40% incrementality. Meaning 60-80% of those "conversions" would have happened anyway.

That's not a comfortable slide to put in a monthly report.

But it's the truth. And truth is cheaper than wasted spend.

We run incrementality tests in the first quarter of every engagement where brand spend exceeds 20% of total budget. Not because we want to find problems, but because knowing the real number is the only way to allocate capital efficiently.

The brands that run this test save money. The ones that refuse keep paying for conversions they already owned.