How-to

Do Advantage+ creative enhancements improve performance? What Meta published, what practitioners report, and how to test it yourself

Every guide on this topic quotes a number, and most of those numbers have no source. Meta's one figure with a methodology is 4 percent, from 2022; what counts for your account is a holdout test you run yourself.

Adlio team5 min read

The short answer

Meta's evidence is one number: 4 percent lower cost per result, from 154,000 ad accounts in May 2022, on the old bundled feature set. The 12 percent CTR and 22 percent ROAS figures in 2026 blogs have no primary source. One agency measured a 26 percent ROAS gain after switching Text improvements off, in one account. Ads Manager reporting does not split enhanced from original variants, so the only number that applies to your account comes from a two-ad holdout test.

What Meta has actually published

Sorted by how much weight it can bear. All fetched or cross-checked on 31 August 2026.

The 4 percent (2022). Meta's Newsroom post of 10 August 2022 introduced Advantage+ creative with a test across 154,000 ad accounts (2 to 30 May 2022) that produced "4% lower cost per result". Meta's developer post of 14 April 2023 restated it: ads "opted-in to standard enhancements delivered a 4% reduction in cost-per-result in campaigns optimizing for link clicks, landing page views, and offsite conversions." Note what that is: one month, the old bundled feature set, three objectives, and an average across accounts that had already opted in.

Product-page figures (undated). Meta's Advantage+ creative page lists "2-3%" more conversions for background generation in catalogue ads, "2%" more conversions on Facebook Reels with video expansion, and "+13%" more conversions using related media. The footnotes with the methodology could not be retrieved from the page.

Case studies (2023 to 2024). An Italian retailer with "56% lower cost per purchase". Fresh with "5x incremental return of ads spend" from a combination of shopping campaigns, Shops ads and text variations. Casetify with "a 13% increase in our return on ad spend" from background generation. Case studies are selected, not sampled.

Campaign-level figures that are not about enhancements. "9% improvement in cost per conversion" for Advantage+ sales campaigns and "10% lower cost per qualified lead" for leads campaigns are about campaign automation, not creative enhancements. Enhancement articles quote them anyway.

The numbers you should stop quoting

  • "12% higher CTR for campaigns using Advantage+ creative", attributed to "Meta Marketing Science 2025" or "Meta's own research". We can find no Meta page with that number. A benchmark aggregator uses 12 percent for AI-generated ads versus human-made ads in general, which is a different claim.
  • "22% increase in ROAS, 17% more conversions at 16% lower cost." Circulates in search snippets and 2026 blog posts. No primary source.
  • "Up to 44% lower cost per result." Refers to the whole Advantage+ suite in the source that cites it, not to creative enhancements.

An article that quotes one of these without a link to a Meta page is repeating another article.

What practitioners report

  • Metalla (agency, August 2025): "In one account, we tested disabling text improvements, which resulted in a 26% increase in ROAS when A/B testing the same ad with and without the enhancement." One account. Others repeat it as if it were general.
  • Jon Loomer (2022 to 2026): the most consistent independent voice, and he publishes no lift numbers. His 2022 take: "slow to blindly trust some of these automatic adjustments", but worth testing. His 2024 per-enhancement guide recommends leaving most on and switching off overlays, and says outright that no test results are provided.
  • HyperFX (May 2026), a pattern rather than data: "Accounts with strong creative production typically see neutral or negative impact from enhancements" while "Accounts with limited creative resources often see positive impact."
  • Ads Uploader (June 2026): "Meta claims a 4% lower cost-per-result on average, but individual results vary significantly by enhancement and account."

The honest reading

Three things are true at once:

  1. Meta's evidence is one 2022 average over a bundle that no longer exists in that form. It is positive, and it is small.
  2. Enhancements do the most for advertisers with the least creative resources. A crop or a brighter image improves a weak file and degrades a finished one. That pattern is reported, not measured.
  3. Nobody can measure it from standard reporting. Ads Manager does not split enhanced from original variants: performance is reported "including all variations" (Flighted, July 2026), and Meta's own note on background generation says results "cannot be broken down by creative variation".

So the only useful number is the one from your own account.

How to test it yourself, cleanly

  1. Pick one creative that is finished to your standard, exported in the ratios each placement needs. Then expansion and crops have nothing to do.
  2. Create two identical ads in the same ad set, same copy, same file: one with every enhancement on, one with every enhancement off. Name them so the split is visible (see naming conventions) and treat the pair as a single-variable batch (how to structure a creative test batch).
  3. Verify the settings on both before they run: the Enhancements panel, or the API read in how to audit enhancements via the API. "Off" has to be explicit.
  4. Run for at least seven days, and until each ad has a meaningful number of results for your objective. Do not read it at 20 conversions.
  5. Compare cost per result, not CTR. Enhancements often raise CTR (a moving image, a text bar) without moving the result you pay for.
  6. Repeat with a different concept before you generalise. One test tells you about one creative.

If enhanced wins, turn that enhancement on for that account. If original wins, you have your answer, and it cost one week. Either way you replace a quoted number with a measured one.

Adlio's position

Off by default, per account and per creative type. A bulk tool's job is to run the ad you approved, and the published evidence for enhancements is small, old and averaged over accounts unlike yours. Every toggle is one click if your own test says otherwise. How the opt-out works, and where it cannot reach, is in how to turn off Advantage+ creative enhancements.

The launch flow that applies that default on every batch, folder in and paused ads out, is on bulk upload Facebook ads.

Frequently asked questions

What does Meta claim enhancements are worth?+

The headline figure is 4 percent lower cost per result, from a test across 154,000 ad accounts between 2 and 30 May 2022, on campaigns optimising for link clicks, landing page views and offsite conversions (Meta Newsroom, August 2022; restated in Meta's April 2023 developer post). Newer per-enhancement claims on Meta's product page (2 to 3 percent, 2 percent, 13 percent) come without retrievable methodology.

Where do the "12 percent higher CTR" and "22 percent ROAS" figures come from?+

We could not find a primary source for either. The 12 percent is attributed in 2026 blogs to "Meta Marketing Science 2025" or "Meta's own research" without a link; the 22 percent circulates in search snippets with no origin. Treat both as unverified.

Is there any independent measurement?+

One published A/B result: Metalla, an agency, reported a 26 percent ROAS increase in one account after disabling Text improvements (August 2025). Jon Loomer, who has tested enhancements since 2022, publishes recommendations without numbers. Everything else is anecdote or vendor claim.

Why can I not just compare enhanced and non-enhanced versions in reporting?+

Because Ads Manager reporting does not split them. Performance is reported for the ad as a whole, "including all variations" (Flighted, July 2026), and Meta's own note on background generation says results cannot be broken down by creative variation. A holdout test with two ads is the only clean measurement.

Sources

Adlio team

Built by media buyers who launch Meta ads in volume. Questions or corrections: support@adlio.ai.

Topic cluster

Advantage+ control

Keeping Meta's automatic creative enhancements off, on every ad, and proving it.

Start here

How to turn off Advantage+ creative enhancements

Related guides