4.6.5.2 - How to Measure Incrementality in Ads (Difficulty: Advanced | Path: Scale)

Lesson Summary

How to Measure Incrementality (Advanced)

What is it?

Incrementality is the measure of sales that were *caused* by your marketing, which would *not* have happened otherwise. It's the 'true lift' we identified in the holdout test. This is the single most important metric for an advanced advertiser.

Why is it important?

This number tells you your true Return on Ad Spend (ROAS). If your ad platform reports a 5x ROAS, but a lift test shows that half of those sales would have happened anyway (from the control group), your *incremental* ROAS is only 2.5x. This tells you exactly how much 'new' money your ads are generating, allowing you to set your budgets with real-world data.

How to Calculate Basic Incremental Lift:

Let's use our 90/10 test example:

1. Test Group (90%): 9,000 people, 180 sales. (Purchase Rate = 2.0%)
2. Holdout Group (10%): 1,000 people, 10 sales.
3. Scale the Holdout: To compare, what would 9,000 people in the holdout group have done? (10 sales * 9) = 90 'baseline' sales.
4. Calculate Lift: Your ads generated 180 sales, but 90 of those would have happened anyway (the baseline).
   Incremental Sales = 180 (Test) - 90 (Baseline) = 90 sales.
5. Your True Impact: Your ads *caused* 90 brand new sales that would not have happened otherwise.

Common Pitfall

Believing the ad platform's built-in 'Conversion Lift' studies without question. While these are based on the same principle, the platform has a vested interest in showing a positive result. Running your own manual holdout test, where you control the audiences, is a more robust (though complex) way to verify the results for yourself.