Standard vs Incremental Attribution on Meta: Which Deserves Your Budget?

Meta hands you a ROAS figure, you trust it, and you move budget because of it. That figure rests on a choice most advertisers never revisit: how Meta decides which purchases your ads deserve credit for. Two models now sit side by side in Ads Manager, and they can tell very different stories about the same campaign.

What Is Incremental Attribution on Meta?

Incremental attribution is Meta’s model for crediting only the purchases your ads actually caused. It relies on Meta’s own modeling instead of a fixed window, so it ignores the click, engage and view settings entirely. Standard attribution works the other way. It credits any purchase that lands inside the window you set, which by default for sales campaigns is 7-day click, 1-day engage and 1-day view.

Because the incremental model is stricter, reported results tend to be lower. That is how it is designed, not a sign something broke.

What Does the Independent Evidence Say?

The most useful data comes from Haus, which compared the two models in geo tests. Across July 2025 to June 2026, campaigns on incremental attribution returned 1.26 times the incremental ROAS of standard attribution on a pooled basis. In the twelve months before, they trailed at 0.80 times. Across the full period the figure was 1.11 times.

Haus describes the sample as modest, and Jon Loomer has reported little difference in his own accounts. My reading is simple: incremental attribution has earned a test, not an automatic switch.

Why a Lower ROAS Is Not Bad News

If you switch and your reported ROAS falls, resist the urge to switch back. The two models use different definitions of a conversion, so their ROAS figures are not comparable. Judge both against what the store records, especially new customer revenue. Our guide to new customer ROAS explains how to set that benchmark.

How to Run the Test Properly

The attribution setting shapes what an ad set optimizes toward, so treat a change as a new test, not a reporting tweak. A clean split looks like this:

• Use the same campaign structure, creative and budget on each model
• Run both for several weeks, not several days
• Judge by store-side new customer revenue, not either model’s reported results
• Give the budget to whichever model moves more new customers

Three Mistakes to Avoid

• Switching the whole account at once. You lose any fair comparison.
• Reading in-platform ROAS across a definition change. Meta’s January and March 2026 updates already changed what a conversion means, so mark both dates on every trend chart.
• Testing during peak season. Black Friday traffic distorts everything, so test before it or well after.

Conclusion

Incremental attribution may deliver more real sales, but the evidence is still early, so prove it on your own store. Run a split test, judge it on new customer revenue, and keep a record of when settings changed. For the full breakdown of Meta’s windows and defaults, read our complete guide to the Meta attribution window. If you’d like help checking your Meta numbers against Shopify, our tracking and attribution service does exactly that.