Should we follow Meta's opportunity score recommendations, and which ones are safe to accept?
Our answer
Don't chase the opportunity score. It counts adopted recommendations, not results. Accept tracking fixes and structural changes other answers back, review every AI creative change before it runs, and test the rest first.
- Confidence
- Medium
- Strongest proof
- 2 Meta rules
- 4 practitioners, 10 opinions or claims
- Sources
- 10 sources
- 16 statements
- M
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentThe score rises when you apply recommendations. Meta's developer page says applying them raises the score, and its Help Center says a high score does not reflect actual or future performance (reported by RedTrack). Elevarus, a paid media agency, calls it an adoption score.
Meta's case for its recommendations is its own figure: 5% more conversions at the same cost per conversion for advertisers who follow its assistant. That compares adopters with non-adopters, and Meta gives no comparison group, period or verticals (Relevant Audience).
Meta's AI creative edits need checking. Eight advertisers and agency executives told Business Insider that cleaning up after Meta's AI had become routine (reported by The Next Web and Common Thread Collective). Meta says reviewing AI output is the advertiser's job.
Our judgment
- Treat each recommendation as the setting change it is. Follow the answer for that setting, because Meta's developer page lists each recommendation as a specific change, Elevarus calls them predictions rather than instructions, and RedTrack sorts them into accept, test and ignore.
- Accept tracking and error fixes and the structural changes other answers back. The exceptions are a change that crosses a real boundary or breaks up a deliberate test. This combines RedTrack's usually-accept list, Elevarus's keep-on list with its boundary rule for audiences and its exception for small test cells, and Common Thread Collective's view that placement optimization and audience broadening have shown lift for many accounts. RedTrack would test placements and broad audiences first.
- Test budget, goal, catalog and Shop recommendations before applying them account-wide. Meta's figures compare adopters with non-adopters, and Elevarus says only a holdout settles platform-reported gains against incrementality results. RedTrack often ignores irrelevant catalog suggestions, and Testing a new Meta setting sets out how to test.
- Review every AI creative change before it runs, rather than turning all of them off. This combines the error reports from Business Insider and Webreality with Common Thread Collective's audit steps and Jon Loomer's advice neither to accept AI output blindly nor to turn enhancements off reflexively.
Our judgment, built from Meta, Elevarus, RedTrack, Common Thread Collective, The Next Web, Webreality and Jon Loomer: our reading of these sources, not a finding from any one of them.
What the sources say
16 statements from 10 sourcesMeta estimates advertisers who adopted its recommendations drove almost 5% more conversions at the same cost per conversion and raised spend 9% versus similar non-adopters. No other method is given.
“drove almost 5% more conversions at the same cost per conversion”
PPC Land relays a platform claim that adopters saw almost 5% more conversions at the same cost per conversion and spent 9% more than similar non-adopters. PPC Land calls it a correlation among self-selected advertisers.
“That design measures a correlation among self-selected accounts.”
Meta says advertisers who adopt its AI assistant's recommendations typically see 5% more conversions at the same cost per conversion. The article notes Meta does not explain the comparison group or period.
“typically seeing 5% more conversions at the same cost per conversion”
Meta, reporting on its own product, says advertisers adopting its Meta AI business assistant see 5% more conversions at the same cost per conversion. No test period, sample or method is given.
“advertisers adopting the Meta AI business assistant are seeing 5% more conversions at the same cost per conversion than those who don't.”
Meta says advertisers who adopt its AI business assistant's recommendations typically see 5% more conversions at the same cost per conversion than those who do not. No period, sample or method is given.
“Advertisers who adopt its recommendations are typically seeing 5% more conversions at the same cost per conversion than those who don't.”
Jon Loomer says to use all five primary text and headline options, make each genuinely different, put key text early since it truncates after one to three lines, and not reflexively turn off enhancements.
“don't simply provide minor variations of the first text. Make them all uniquely different.”
Meta's opportunity score runs 0 to 100 and rises as you adopt its recommendations, which Meta says were rigorously tested but publishes no tests for. Advertisers apply them themselves.
“assigned point values based on how much each is expected to improve your campaign performance”
A practitioner says Advantage+ creative began rewriting headline text inside uploaded images on July 27, 2026, up to 8 variants per image, on by default. Check active campaigns and review variants for misstated offers.
“On July 27, 2026, Meta activated a default-on feature inside Advantage+ Creative that automatically rewrites headline text embedded in your uploaded ad images.”
Elevarus says Meta's opportunity score is an adoption score, not a performance score: you can reach 100 whatever your cost per result is doing. It treats recommendations as predictions, not instructions.
“It is not a performance score. It is an adoption score.”
Jon Loomer fills five primary texts and five headlines and keeps Meta's AI text variations in rotation, but says Meta's branding settings are not always followed. No performance data.
Common Thread Collective, summarizing a Business Insider probe, advises auditing Advantage+ creative enhancements, comparing live ads to source files, turning off unintended AI changes, and rechecking in 48 hours.
“multiple agencies confirmed that after turning AI settings off inside Ads Manager, the settings toggled back on.”
Business Insider reports eight advertisers and agency executives say cleaning up after Meta's AI has become routine. Meta says it is the advertiser's responsibility to review AI outputs. No performance data.
“AI can make mistakes and that it is the advertiser's responsibility to review the AI outputs.”
Andrew Foxwell expects a low opportunity score in Meta's account audit to hurt delivery in the auction. This is his expectation, not a Meta statement.
“I think you're going to want to get familiar with this because that is going to have implications in the auction”
Webreality, an agency, says Advantage+ creative enhancements made off-brand, misleading versions of a client's ads, so it turned them off before launch. It advises checking every default-on setting.
“every "enhancement" Meta tried to force on us misrepresented the business and broke the brand guidelines.”
RedTrack, which sells tracking software, says Meta's opportunity score measures readiness, not results. Accept fixes such as Conversions API errors, test automation carefully, and skip changes that may reset learning.
“Opportunity score (including a high score) itself does not reflect your actual or future performance.”
Sydney Chocolos says Meta's ad setup warned her of a 10% lower click-through rate if she skipped its AI-generated images and post copy.
“it was telling me things like, I'm going to get a 10% lower clickthrough rate without using their AI generated images and post copy”
In the full answer
For readers who sign in- What to do this week3 steps
- 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