Incrementality Is the Only Number That Matters. Almost Nobody Measures It

Here is the test. Pause your Meta budget for a month. Watch what happens to revenue.

Most marketers will not do this. The reason they give is that it is too risky. The real reason is that they are worried the answer will be nothing.

Attribution vs Incrementality

Attribution answers the question of who gets credit when a conversion happens. The customer clicked four ads before they bought. Which channel deserves the thank-you note?

Incrementality answers a completely different question. If that channel did not exist, would the conversion still have happened?

These are not the same thing. They are not even close to the same thing. And most marketing budgets are allocated as if they were.

Why The Difference Matters

Imagine a customer who was already going to buy from you. They googled your brand, clicked a branded search ad, and purchased. The ads platform gives your paid search campaign full credit.

The incremental value of that click was zero. The customer was going to find you anyway. You paid Google to take credit for a conversion that belonged to your brand, your product, or your organic search traffic.

Scale this across thousands of conversions and you start to understand why some companies cut 30% of their paid budget and see no revenue impact.

How To Actually Measure It

There are three serious methods.

Geo holdouts. Run the campaign in some geographies and not in others. Compare the difference. Works well for TV, radio, and broad reach digital. Harder for tightly targeted search.

Time-based pauses. Turn a channel off for two to four weeks. Watch what happens to total revenue, not just attributed revenue. The gap between what attribution predicted would stop and what actually stopped is the incremental contribution.

PSA / ghost ads. Show a placebo ad to a control group. Compare conversion rates between the group that saw the real ad and the group that saw the placebo. This is the cleanest method and also the most operationally difficult.

All three require the willingness to turn something off and see what happens. Which is where most teams bail.

Why Teams Avoid It

Because the results are often uncomfortable.

The first time you run a serious incrementality test, you often discover that 20 to 40% of what your attribution tool credits to paid media would have happened anyway. Some channels perform worse than expected. Some perform better.

Either way, the result is a document that makes last quarter's budget decisions look wrong. That is politically inconvenient. So most teams stick with attribution, which never makes anyone look wrong because it assigns credit whether the credit is deserved or not.

What To Actually Do

If you run meaningful paid budget (say, over €20k a month), you should be running incrementality tests at least twice a year. Pick your two biggest channels. Run a proper test. Accept whatever the result tells you.

The output will change how you allocate budget. Sometimes dramatically. Always for the better, because for the first time you will be making decisions based on causation rather than correlation dressed up in tracking pixels.

Attribution tells you how you got credit. Incrementality tells you whether you earned it.