All answers

The Prevailing
Answers for
Business model:
Business model, pick up to two

Should we switch to Meta's Incremental Attribution setting?

The PrevailingUpdated Oct 27, 2026

Our answer

The answer depends on spend.

Switch your volume campaigns to Incremental Attribution and expect Meta's reported ROAS to drop. Switch back if a lift test shows no gain after 4 weeks.

Test Incremental Attribution before moving to it. Split the volume campaign or run a lift study for 2 to 4 weeks, and make it the default if the test agrees.

Expect about parity with standard attribution. Switch to Incremental Attribution only if your own test shows a gain.

Why this is the answer

3 reasons and our judgment

Incremental Attribution delivered 1.26 times the incremental return of standard attribution in tests by Haus, an incrementality testing platform, from July 2025 to June 2026. For brands that sell only on their own site, Haus measured 1.38 times.

Client accounts improved after the switch. Ecommerce growth consultant Andrew Faris moved 10 to 12 accounts to it and saw higher spend and performance, and he says no client has done worse.

The case for Incremental Attribution isn't settled. Haus hasn't disclosed its 2026 test count, and Olivia Kory of Haus says results range from crushing to tanked.

Our judgment

The 4-week cutoff is ours. It draws on Brad Ploch's 3-week head-to-head test and his second account, where two strong weeks were followed by weeks he called miserable (both on Scalability School, a podcast run by agency hosts). Judging the switch on new-customer revenue and lift tests rather than reported ROAS follows Ben Heath's warning that reported ROAS may drop, Meta's note that the results can't be compared in one table with standard attribution, and Kory's advice to test. Applying Haus's result for brands that sell only on their own site to subscription brands is an extension.

Our judgment, built from Brad Ploch (Scalability School), Ben Heath, Meta, Olivia Kory and Haus: our reading of these sources, not a finding from any one of them.

What the sources say

21 statements from 8 sources
  • MetaUndatedSupports

    Meta defines incremental attribution as optimizing delivery for conversions its models predict the ad caused. It warns results cannot be compared in one table with standard-attribution ad sets.

    “Results cannot be compared in the Campaign Overview table across ad sets with different attribution models.”

    Read it on facebook.com

  • Barry HottNov 6, 2026Caveat

    Barry Hott predicts Meta's on-device conversion measurement will make Meta credit more sales, most for 7-day click 1-day view users, and widen gaps with outside trackers. He still advises not opting out.

    “I wouldn't recommend opting out of this change.”

    Read it on x.com

  • Barry HottOct 23, 2026Supports

    Barry Hott says an account chasing 1-day view conversions can rack up cheap credited conversions from recent visitors and low-quality placements, perhaps 1% of them truly new sales; incremental attribution avoids that.

    “Incremental attribution doesn't care to get "conversions" like that.”

    Read it on x.com

  • Olivia KoryOct 9, 2026Supports

    Haus's holdout tests found Meta's Incremental Attribution tracked measured lift more closely than standard windows, and optimizing to it drove about 1.2x the incremental return. Sample size not given.

    Read it on x.com

  • Barry HottOct 9, 2026Caveat

    Barry Hott says incremental attribution does not stop Meta targeting existing customers or warm audiences. It stops wasting budget on people who were going to buy anyway.

    “Meta still targets them, it just stops wasting your budget on the ones who were already going to buy anyway.”

    Read it on x.com

  • Ben HeathOct 7, 2026Caveat

    Ben Heath says Incremental Attribution steers Meta spend toward sales you wouldn't otherwise get, though Meta's reported ROAS may drop.

    Read it on youtube.com

  • AnyTrackSep 22, 2026Caveat

    A tool vendor relaying Meta's rules says incremental attribution works for Sales, Engagement and Leads objectives, cannot be used with bid controls, and cannot be changed after the campaign is published.

    “You cannot use bid controls with incremental attribution.”

    Read it on readme.anytrack.io

  • Andrew FarisAug 17, 2026Supports

    Andrew Faris says all his clients now run manual bids with Meta's Incremental Attribution, mostly cost caps, and none has gotten worse.

    Read it on youtube.com

  • HausAug 17, 2026Disagrees

    Haus's Olivia Kory says its 2025 read on Incremental Attribution rested on about a dozen tests; Andrew Faris notes the 2026 sample is still not large, and Kory says brands should test it rather than switch blind.

    Read it on youtube.com

  • Andrew FarisAug 17, 2026Caveat

    Andrew Faris says Incremental Attribution gave a much smaller gain on highest-value (ROAS goal) campaigns than on volume campaigns, and sometimes a slightly worse result, for his agency's clients.

    Read it on youtube.com

  • HausAug 17, 2026Supports

    Pushes back on running dual-cell tests: Andrew Faris says brands spending under $5K a day should just switch to incremental attribution.

    “If you're sub-5 grand a day in spend, just switch, just switch”

    Read it on youtube.com

  • HausAug 14, 2026Disagrees

    Haus's Olivia Kory says brands' results with Meta's Incremental Attribution vary widely: some find it crushing, others say it tanked their business.

    “this tanked my business”

    Read it on youtube.com

  • Andrew FarisAug 10, 2026Supports

    Andrew Faris says brands spending below about $10K a day on Meta should move their highest-volume spend straight to Incremental Attribution, and below $5K a day should just switch rather than test.

    Read it on youtube.com

  • Andrew FarisAug 10, 2026Supports

    Andrew Faris says brands that sell only on their own site should not test Incremental Attribution first; they should simply switch to it.

    Read it on youtube.com

  • HausAug 10, 2026Supports

    Andrew Faris, a consultant, says all of his clients run manual bids or cost caps with incremental attribution on Meta and none has seen performance go backward. He gives no sample size.

    “100% of my clients are running manual bids with incremental attribution”

    Read it on youtube.com

  • HausAug 10, 2026Supports

    Andrew Faris of Haus says that if you spend under $5,000 a day on Meta, a two-cell test is not worth running: just make the switch.

    “if you're sub-5 grand a day in spend, just switch. Just switch.”

    Read it on youtube.com

  • Olivia KoryAug 10, 2026Supports

    Haus's Olivia Kory says the ratio of Meta's incremental attribution ROAS to standard ROAS moved from 0.8 times in 2024 to 1.26 times in July 2026, giving incremental attribution a slight edge.

    “you see that number jump to 1.26x, which means now incremental attribution has the slight edge.”

    Read it on podcasters.spotify.com

  • HausJul 16, 2026Supports

    Haus found Meta's Incremental Attribution setting delivered 1.26 times the incremental return of standard attribution from July 2025 to June 2026, up from 0.80, and 1.38 for brands selling only on their own site.

    “IA iROAS is beating standard iROAS by a pooled geo-mean of 1.26x, compared to 0.80x in the July 2024-June 2025 period we originally reported on.”

    Read it on haus.io

  • Andrew FarisJul 15, 2026Supports

    Andrew Faris says he changed approach after a client's bid caps would not spend. Spend then rose 3-4x, and he later saw higher spend and performance across the 10-12 accounts he moved.

    Read it on youtube.com

  • Andrew FarisJul 15, 2026Supports

    Andrew Faris says advertisers should choose incremental attribution at the ad set level in Meta when given the option over standard optimization.

    “We are fully embracing incremental attribution.”

    Read it on youtube.com

  • Scalability SchoolJul 2, 2026Disagrees

    Brad Ploch says for one apparel brand at $150-200K a month, Incremental Attribution passed a lift test but drifted after Meta's March click change, so they went back to one-day click and plan to retest.

    Read it on scalabilityschool.com

In the full answer

For readers who sign in
  • What to do this week3 steps
  • The numbers behind it3 figures from cited sources
  • Where each source standsA chart placing each source on the question
  • The strongest disagreementThe best case against the answer, from a cited source
  • How sure we areWhat its confidence rests on, and what would change the answer

Read next

More in Measurement