Short answer: In our experience, Meta accounts spending more than $300,000 to $400,000 a month usually waste 25% or more of their budget on sales that would have happened anyway. Ads Manager can't show you this, because it credits every sale your ads were near, not just the sales they caused. A conversion lift test can. For big accounts that have tapped out on growth, the win from incrementality often isn't more sales. It's the same sales for 25% to 30% less.
Why can't you believe the numbers in Ads Manager?
Because attribution measures what happened near your ads, not what your ads caused. If someone was already going to buy, saw your retargeting ad, and bought, Meta counts that sale. So does your attribution tool. So does Google, if they searched your brand first.
Meta's own measurement team says it plainly in their incrementality training: the biggest challenge in advertising is "paying for customers who would have bought anyway." The same deck cites research from Analytic Partners estimating that 35% of spend is wasted on non incremental sales when budgets are set by last click.
Academic work backs this up. Researchers at Northwestern's Kellogg School, working with Meta, compared hundreds of randomized Facebook experiments against standard measurement methods. The standard methods often missed the true effect by a wide margin, sometimes overestimating and sometimes underestimating. In a later paper covering 2,226 Meta experiments, 7 day last click attribution explained only a small share of the real incremental effect.
So our rule is simple: you can't believe any dashboard on its own. Not Ads Manager, not your attribution tool, not GA4. You have to test.
Why do big accounts waste more?
At small budgets, Meta spends most of your money finding new people. At big budgets, it runs out of easy new people and starts spending more on people who already know you: past buyers, site visitors, email subscribers, people who follow you. Those people convert at a high rate, so the reported ROAS looks great. But many of them were going to buy anyway.
That's why we see the pattern most at $300,000 to $400,000 a month and up. The account looks efficient, growth has stalled, and a quarter or more of the budget is paying Meta for sales it didn't cause. This is our experience across the accounts we've audited and tested, not a published benchmark. Your number could be higher or lower, which is exactly why you test.
What is incrementality?
Incrementality is the share of results your ads actually caused. Meta's definition is the cleanest: incrementality answers "what outcomes did this ad or campaign actually cause that wouldn't have happened otherwise?"
The only way to measure it properly is a randomized experiment. On Meta, that's a conversion lift test: Meta splits your audience at random into a test group that can see your ads and a control group that can't. The difference in conversions between the two groups is your incremental lift.
How much can you save?
Here's a made up, rounded example of what a lift test can reveal on a big account.
| Before the test | After acting on it | |
|---|---|---|
| Monthly spend | $400,000 | $300,000 |
| Purchases Ads Manager reports | 8,000 | 6,600 |
| Incremental purchases (from lift test) | 4,800 | 4,700 |
| Cost per incremental purchase | $83 | $64 |
The account cut $100,000 a month and lost almost no incremental sales. Ads Manager would show a drop in reported purchases and make the cut look like a mistake. The business would see the same sales and a lot more profit.
That's the point for tapped out accounts: there may not be more growth to buy, but there's often 25% to 30% to save.
How often should you run lift tests?
All year, not once. Incrementality changes as you scale, change creative, enter new seasons and add channels. Our approach:
- Every quarter: a lift test on the account or your biggest campaign.
- Before a big budget increase: test first, so you know what the next dollar buys.
- After a big structural change: new campaign setup, Advantage+ changes, new exclusions.
- Before and after peak season: Black Friday behaves differently from March.
The Kellogg researchers make a similar point: you don't need to test every campaign, but you need to test regularly and use the results to adjust how you read everything else.
Running the test is the easy part
A lift test only tells you something is wrong. A great media buyer is the person who fixes it. That usually means excluding warm audiences as well as Meta allows, shifting budget toward ads that reach new people at low frequency, and cutting campaigns that mostly re-reach buyers. We cover it step by step in what to change after a lift test.
Key takeaways
- Attribution credits sales your ads were near, not sales they caused.
- In our experience, accounts over $300,000 to $400,000 a month often waste 25% or more.
- A conversion lift test is the only clean way to measure it on Meta.
- For tapped out accounts, the win is often 25% to 30% savings, not more sales.
- Test every quarter, and act on the results.
FAQ
Is 25% wasted spend a proven number? It's our experience from big accounts, not a published benchmark. Meta's own training cites an Analytic Partners estimate of 35% waste when budgets follow last click. Your account needs its own test.
Does a lift test cost me sales? A small share of your audience doesn't see ads during the test. That's the price of knowing the truth, and it's usually far smaller than the waste it uncovers.
Can smaller accounts waste spend too? Yes, especially with heavy retargeting. It just tends to grow with budget.
What's the first thing to change if my test shows low lift? Look at who your spend is reaching. If most of it goes to past buyers and site visitors, start with exclusions and low frequency, new audience ads.
Want to know how much of your Meta spend is incremental? We do a free Meta audit: a full account audit plus net new reach and cumulative reach reports. Reach out on our contact page.
Sources
- Meta, "Measure What Matters with Incrementality" practitioner deck (citing Analytic Partners ROI Genome 2023 and eMarketer 2024)
- Gordon, Moakler and Zettelmeyer, "Predicted Incrementality by Experimentation (PIE) for Ad Measurement," arXiv 2304.06828
- Kellogg Insight, "Here's a cost effective way to tell if your digital ads are working."
Check what you learned
Tap an answer to see if you got it right.
1. What does attribution credit?
2. What is the only clean way to measure wasted spend on Meta?
3. For tapped out accounts, what is often the win?
Ready for the full test? Take the Measurement exam and get certified. Want this run on your account? Get my free audit.
Module 3: Measurement · article 7 of 15
Previous: When Meta Blocks Your Pixel: How We Rebuilt Lead Tracking for a Healthcare Client
About the author
Jake Mehani, founder of Creative Analytics
Jake has spent 15 years running Meta and paid social ads for consumer brands. Meta published his RBX Active Advantage+ case study, he taught the DMAnc Meta and TikTok ads masterclass, and he spoke at Motion's 2024 Creative Strategy Summit. His senior team runs media buying, creative and measurement like an in house team.