Open Ads Manager and there is a number between 0 and 100 above your campaigns table, with a list of suggestions promising to raise it. That number is the opportunity score. It is easy to read it as a grade on your account, and it isn't one. It measures how many of Meta's recommendations you have applied, weighted by how valuable Meta thinks each one is for you. This guide covers what the score measures, where it shows up, how recommendations earn points, what Meta says about their predicted impact, what happens when you dismiss one, and what auto-apply does. The last part is practitioner advice on which recommendations deserve a closer look when you run many campaigns or accounts. Everything except that section comes from Meta's Help Center, its opportunity score page and its developer documentation.
What the opportunity score measures and where it appears
Meta describes the opportunity score as a measure, on a 0-100 point scale, of how optimized your campaigns, ad sets and ads are. The score is based on how many Ads Manager recommendations you apply, and Meta ranks those recommendations by estimated performance impact. Recommendations include best practices as well as errors and warnings, such as incorrect formatting, missing information or policy violations.
The scope is the ad account. Meta's FAQ calls opportunity score account-level, covering your entire ad account, while the recommendations behind it can target a campaign, an ad set or an ad. Meta's Marketing API exposes recommendations at the ad, ad set, campaign and ad account levels, and returns the score itself as a field on the ad account.
Meta lists three places. In Account Overview there is an Opportunity score section with your score, your recommendations, Available and Dismissed tabs, and instructions for applying each one. In the Campaigns tab the score sits at the top of the page above the table; click it to see and apply recommendations, and you may also see an Actions column listing recommendations that could raise your score. The score also appears when you create an ad.
Two situations hide or freeze the number. If you have an account-level error, Meta says the score will not show. And if a campaign has multiple errors, you will not receive points until all of them are resolved.
Meta is explicit about what the number does not mean. Its Help Center says the score does not reflect your actual or future performance, even if it is high, and that applying recommendations can help improve performance over time but is not guaranteed. Meta also says you should not turn off campaigns with a low score, because the score shows how optimized a campaign is, not whether it is performing poorly. Judge performance against your business goals. Meta labels the feature as in development and says it may evolve.
How recommendations raise the score
Each recommendation carries points. Meta says points are personalized by how valuable the recommendation may be for your account, based on your business type, campaign objective and characteristics, budget allocation, and the performance improvements seen for that recommendation in Meta's experiments. Errors and warnings are assigned points the same way.
Meta's own example: an account at 60 could fix audience fragmentation, where similar audiences in different ad sets compete with each other, for 35 points, and turn on Advantage+ placements for 5 points. Applying both could take the score to 100. Meta adds that fragmented ad sets would be the place to start if you want to prioritize by points.
The score moves for more reasons than applying recommendations. Meta says it updates in near real time when you apply recommendations or create, edit or turn off campaigns, ad sets or ads, and it can drop when the system identifies new recommendations, for example when creative fatigues and Meta suggests adding new creatives. A score of 100 means you have applied all current recommendations or none are available right now. Being in the learning phase or learning limited does not currently affect the score, according to Meta's FAQ.
One more detail: Meta calls the score a weighted measure of recommendations adopted relative to the maximum opportunity. A 10-point rise means you applied more of what Meta suggested, not that performance rose by the same amount. Meta says there is no guarantee a higher score brings higher performance.
The types of recommendations you will see
Meta's Help Center groups recommendations into categories, each with an example. Meta says the list will evolve with its latest best practices.
- Automated campaigns: moving to campaigns with end-to-end automation, such as an Advantage+ app campaign.
- Objective and goals: for example, switching to the Maximize number of conversions performance goal.
- Audience: audience expansion, audience fragmentation and auction overlap, such as turning on Advantage+ audience. Meta notes you can still set limits on who can and can't see your ads.
- Creative and placements: improving and diversifying creative, including creative fatigue alerts that suggest adding another image or video.
- Signals: data quality and signal strength, such as setting up Conversions API alongside the Meta Pixel.
- Budget and bidding: for example, Scale your high-performing campaign, which suggests increasing budgets on campaigns with stable delivery and better performance than your others with the same goals.
- Other: important errors and warnings, plus partnership and non-campaign items such as replying to customer messages within 5 hours.
What Meta says about predicted impact
Meta says every recommendation is backed by a study showing how it helped improve advertiser results on average, and that recommendations are ranked by highest estimated performance impact. On its opportunity score page, Meta states that advertisers who adopted the recommendations saw an over 5% median decrease in cost per result on average. That is Meta's own aggregate figure, not a forecast for any single account.
In the Marketing API, each recommendation can carry an estimated lift, which Meta's developer blog describes as the average performance improvement seen for that recommendation in Meta's experiments, plus the number of score points it could add. Meta's developer documentation says the API may return fewer recommendations than Ads Manager shows, because new types are tested in Ads Manager first.
Read the predicted impact as an average across advertisers. The same pages repeat that a high score does not guarantee performance and that results depend on factors the score does not account for, such as market dynamics.
Applying, dismissing and auto-apply
Applying a recommendation makes the change and adds its points. Many can be applied in a few clicks from Account Overview or the Campaigns tab. Through the Marketing API, supported types can be applied with a POST request, and for several of them Meta's documentation says that if you don't pass an object selection, the change applies to all objects listed in the recommendation. If you script this, select objects explicitly.
Dismissing does not change the score. Meta says a dismissed recommendation moves to the Dismissed tab in Account Overview, and if business constraints stop you from applying some recommendations, your score will always be below 100. A dismissal applies only to that campaign, so the same recommendation can appear again on another campaign. Meta's FAQ also says your ads will not be penalized for a lower score or for dismissing recommendations.
Auto-apply exists. Meta's opportunity score page includes a tip to use auto-apply in Account Overview to select recommendations and apply them automatically when they become available. In the Meta Help Center and developer pages we reviewed, we found no step-by-step instructions for turning auto-apply off. Check the auto-apply option in your Account Overview, and review which recommendations are selected.
Practitioner advice: which recommendations to question at scale
This section is our practitioner view, not Meta guidance. Meta builds recommendations from averages across advertisers and from your account's settings. At scale, your structure often has reasons Meta can't see: separate client budgets, compliance limits, deliberate tests or spend spread across several ad accounts. Treat these as a review queue, not a to-do list. Errors and warnings are the exception: fix those.
- Budget increases: Scale your high-performing campaign compares a campaign with your other campaigns that share its goal, and budget-limited suggestions say budget may be limiting results. Neither knows your margin, cash flow or account spending limits. Check the cost per result against your own target before raising anything, and raise in steps. See our budget increase rules.
- Audience fragmentation and consolidation: these can carry the most points, as in Meta's 35-point example. If ad sets are split on purpose, by offer, country, client or test cell, merging them can wipe out the comparison you built. Consolidate true duplicates, and dismiss the rest at campaign level.
- Broad audience changes: Advantage+ audience can work, but on an account with exclusions or regulated offers, test it in a separate ad set first and confirm the limits you need are still set. Our broad targeting guide covers how it behaves.
- Automatic creative changes: suggestions such as music, text improvements, overlays or generated image expansion change what people see. If you have brand or compliance review, preview each ad before you accept, especially when applying in bulk.
- Placement expansion: Advantage+ placements is often a reasonable default, but check placement reporting afterwards if certain placements didn't work for you before.
- Auto-apply: on accounts with large or shared budgets, we'd keep budget and audience recommendations out of auto-apply so every change has an owner. If you want automation that follows your own thresholds, automated rules are easier to audit.
- Signals: Conversions API and data quality recommendations are usually worth acting on, since they improve measurement rather than spend. See our Conversions API setup guide.
Using the score across many accounts
When you run campaigns across several ad accounts, each account has its own score and its own list of recommendations, and a dismissal on one campaign doesn't carry over to another. A short weekly pass through Account Overview on each account is usually enough: fix errors, apply what fits your plan and dismiss the rest so the Available tab stays useful. Power Ads supplies Meta agency ad accounts at 4% per top-up; we don't manage campaigns, so these decisions stay with you.
Key takeaways
- The opportunity score is a 0-100 account-level measure of how many of Meta's recommendations you have applied, weighted by estimated value. It is not a performance grade.
- Points are personalized by business type, objective, campaign characteristics, budget allocation and Meta's experiment results.
- Dismissing a recommendation doesn't change the score, applies only to that campaign, and Meta says your ads are not penalized for it.
- Meta's 5% median cost per result figure is its own aggregate figure, not a promise for your account.
- Fix errors and warnings, but review budget, consolidation, broad audience and auto-applied changes against your own goals before accepting them.
FAQ
Does a low opportunity score hurt my ad delivery?
Meta says no. Its FAQ states your ads will not be penalized for a lower opportunity score or for dismissing recommendations, and that you should not turn off campaigns with a low score. Judge campaigns on your business goals.
Why did my score drop when I didn't change anything?
Meta says the system may have identified new recommendations, for example when creative fatigues and it suggests adding new creatives. Errors and warnings also change the score.
If I dismiss a recommendation, will my score go up?
No. Meta says the score does not change when you dismiss a recommendation. It moves to the Dismissed tab in Account Overview, and if you can't apply some recommendations because of business constraints, your score will stay below 100.
Does the learning phase affect my opportunity score?
Not currently. Meta's FAQ says being in the learning phase or learning limited doesn't affect your score, although applying some recommendations may help you exit learning faster.
Is campaign score the same as opportunity score?
No. Campaign score covers a single campaign while you draft it and predicts how optimized it is before publishing. Meta says it does not affect your account's opportunity score.
Can I turn off auto-apply?
Meta's opportunity score page describes auto-apply in Account Overview as a way to select recommendations to apply automatically when they become available. In the Meta pages we reviewed, we found no step-by-step instructions for turning it off, so check the auto-apply option in Account Overview and review which recommendations are selected.
Sources
- facebook.com/business/help/804913634782260
- facebook.com/business/help/2086509315182746
- facebook.com/business/help/1729106371347026
- facebook.com/business/help/3864826443789572
- facebook.com/business/help/149770628855079
- facebook.com/business/tools/opportunity-score
- developers.facebook.com/documentation/ads-commerce/marketing-api/overview/performance-recommendations
- developers.facebook.com/blog/post/2025/11/10/unlock-peak-performance-with-new-opportunity-score-features-in-the-marketing-api
