You open Ads Manager to check on the week and there is a dial at the top of the campaign table. It says 64. Hover over it and a little panel slides out with Meta's suggestions for your account. Combine these ad sets. Turn on this placement. Scale this campaign.
Sixty-four out of a hundred reads like a D-minus. Nobody likes a D-minus, so the natural move is to start clicking Apply until the needle climbs.
Before you do, it helps to know one detail about how that number is built, because it changes what the number means. On Meta, there is exactly one way to raise your score: do what Meta suggested. Saying no does not count.
Saying No Costs You Points. Permanently.
Meta calls the dial your opportunity score. It is scored out of 100, and Meta's own definition is plain about what feeds it: your score is based on how many Ads Manager recommendations you apply. It sits at the top of the Campaigns tab and again, in more detail, on the Account Overview page, where the recommendations are listed underneath it.
Each recommendation carries a point value, and the values are personalized. Meta weighs them by your business type, your campaign objective, how your budget is spread, and how much that change helped other advertisers in its experiments. So the same suggestion might be worth 20 points in your account and 6 in someone else's. Meta's help page gives its own worked example: an account sitting at 60 could add 35 points by fixing something called audience fragmentation, and another 5 by switching on Advantage+ placements. Apply both and the account reads 100.
Now the part that matters. Every recommendation has a dismiss option, and here is what Meta says happens when you use it: "If you dismiss a recommendation, your score will not change." The recommendation moves to a Dismissed tab and the points stay missing. Meta then spells out the consequence itself. If you cannot apply some recommendations, "for example, due to business constraints, your score will always be lower than 100 points."
That is a different design from Google's. Over on Google Ads, dismissing a recommendation raises your optimization score, so a perfect score there can mean you accepted everything or rejected everything. We took that number apart in A 100% Optimization Score Just Means You Agreed With Google. Meta closed that loophole. On Meta, a 100 can only mean one thing: somebody said yes to every card.
Two more details worth knowing. Dismissals work one campaign at a time, so declining a suggestion on one campaign does not stop it reappearing on the next. And the score does not touch delivery. Meta's own FAQ says your ads will not be penalized for a lower score or for dismissing recommendations, and a Meta product marketing manager has said publicly that the score is not used in the auction. Meta even tells advertisers not to turn off a campaign because its score is low, since the score reflects how a campaign is set up, not whether it is working.
So the dial is not a grade on your results. It is a count of how many of Meta's suggestions you have agreed to, weighted by how much Meta thinks each one matters.
Meta's Own Number Is a Modest One
Meta does make a performance claim for all this, and it is worth reading closely. The opportunity score page says advertisers who adopted the recommendations saw "over 5%" median decrease in cost per result. The same page said 5% when it was archived in March 2025.
Take that for what it is. Five percent is a real improvement, and it is a median, which means plenty of accounts did better and plenty did worse. It is also measured across advertisers who chose to take the advice, which tells you nothing about the cards you looked at and decided against. Meta's page carries the honest caveat right underneath it: a high score "does not reflect your actual or future performance."
None of that makes the list useless. A modest average gain from a pile of mostly sensible fixes is exactly what you would expect. It just means the dial deserves attention, not obedience.
Most of Meta's Advice Is What We Would Tell You Anyway
Here is where Meta's version earns more credit than its reputation suggests. A good share of what shows up in that panel is the same thing we tell clients, and in several cases it is the same thing we have written whole articles arguing for. If one of these appears in your account, the default answer is yes.
Fix the data connection. Meta files these under "signals," and the most common one asks you to set up the Conversions API alongside your pixel. The pixel reports sales from the visitor's browser, which ad blockers and privacy settings interrupt. The Conversions API sends the same events from your website's server, so fewer sales go missing. Meta can only learn from the sales it can see, which is the whole argument of Meta's AI Can Only Be as Smart as the Data You Feed It. On Shopify, WordPress and most booking platforms this is an integration you switch on, not a developer project.
Replace tired creative. When the same people have seen an ad too many times, Meta flags it. Its own example warns that cost per result for that ad set "may be at least twice as much as ads you ran in the past." That is ad fatigue, and it is real. The fix is a new image or video, not a rebuilt campaign. We explained how to spot it early in Your Best Meta Ad Has a Shelf Life.
Add a vertical version. If your ads only exist as squares, Meta will suggest a 9:16 version for Reels and Stories. Take it. A square ad stretched into a full-screen placement gets cropped or boxed, and it looks like an ad that was not made for where it is running. Your Story Ad Is Not Your Feed Ad Resized covers what that version should actually look like.
Let Meta pick placements. If you trimmed the placement list by hand, expect a card asking you to turn Advantage+ placements back on. For most small accounts that is the right call, for reasons laid out in Meta Wants to Choose Where Your Ad Runs. Let It.
Combine ad sets that are doing the same job. Three ad sets with the same offer, the same creative and slightly different interest targeting split one small budget three ways and compete against each other in the same auctions. Meta will call that fragmentation or auction overlap, and when the ad sets genuinely are doing the same job, it is right. Fewer, simpler campaigns learn faster, which is the case we made in Stop Building Complicated Meta Campaigns.
Fix errors and warnings. These count toward the score too, and they are always worth clearing. A rejected ad or a broken link is money not working.
Three Cards That Need Your No
The cards to watch share a trait. They do not improve what Meta learns from. They change how much you spend, which customers get lumped together, or what your ad says, and each of those depends on things only you know.
"Scale Your High-Performing Campaign"
Meta's wording for this one is gentle: the campaign "has had stable delivery and better performance" than others with the same goal, so consider increasing the budget. Every account sees it eventually.
Sometimes it is right. If a campaign is profitable at its current spend and you can handle more work, scaling is the job. But the answer turns on three facts that live outside Ads Manager. Do you have room on the schedule? A plumber booked solid for a month gains nothing from twenty more leads except twenty disappointed callers. Does the margin hold? "Better performance" in Meta's terms means cheaper results, which is not the same as profitable ones. And is the offer proven yet? A business still working out which ad and which offer sells is better off learning at a small budget than paying more to learn the same lesson faster.
Even when the answer is yes, go up gradually. A campaign that works at $30 a day can stumble when it suddenly gets $150, because Meta has to go find a much larger pool of buyers who are, on average, less ready than the first ones. We covered why in Your Meta Ads Work at $30 a Day. Here's Why They Break at $150. Let each step settle before the next.
The Fragmentation Merge, When the Ad Sets Do Different Jobs
This is the card carrying the most points in Meta's own example, and it is the one most likely to be wrong in a small account that was split on purpose.
When you click into it, Meta shows you what it plans to do: keep the best-performing ad set running, switch the others off, and pool all of their audiences into the survivor. That is fine when the ad sets were duplicates. It is a problem when they were separate for a reason.
One agency walked through this on camera in a live client account. Meta wanted to merge three ad sets: a broad prospecting ad set, an interest-targeted prospecting ad set, and a remarketing ad set aimed at people who had already visited the site. The remarketing ad set had the best numbers, as remarketing nearly always does, so it was the one Meta planned to keep. Apply the card and new-customer prospecting would have been folded into an ad set built for warm visitors.
That matters because warm audiences always look like the winner. They convert more easily, so a campaign that leans toward them reports a lovely cost per result while the stream of new customers quietly dries up. We wrote about exactly that drift in Meta Removed the Cap on Ads to Your Existing Customers. Build It Back.
The test is simple. Open the merge preview and look at what each ad set is for. Same job, different interests? Merge them. Different jobs, like finding strangers versus reminding past visitors? Dismiss it, and accept the points you lose.
Cards That Rewrite or Restyle Your Ad
The creative category mixes the excellent (a fresh image for a fatigued ad) with suggestions that change your ad without you seeing the result. Cards to opt more ads into Advantage+ creative enhancements, to add text variations, or to add music automatically all fall here. Some of those changes are harmless. Some put cropped images, generated backgrounds or rewritten lines in front of customers under your name. Preview every one before applying it, and see You Approved One Ad. Meta Is Running Several. for what to look for.
The same caution applies to Advantage+ audience when the campaign has a specific reason to be narrow, such as a campaign set up to test creative against a fixed audience. And if you run a health practice, audience suggestions need extra care for privacy reasons we covered in Your Health Business Cannot Retarget People the Normal Way.
The Good Cards Feed Meta. The Risky Ones Decide for You.
Meta's list changes over time, so a memorized list of good and bad cards goes stale. One question sorts nearly all of them.
Does this card improve what Meta learns from, or does it change a decision about money, customers or your message?
Cards that improve the inputs are usually a yes: better data, fresh creative, the right formats, fewer duplicate ad sets, fixed errors. They make the system smarter without taking anything away from you.
Cards that change a decision are yours to make: how much to spend, which customers get grouped together, what the ad says. Meta might be right about any of them. It is just guessing from inside the account, and you know things the account does not.
Automatic Adjustments Say Yes for You, Every Day
On the Account Overview page there is a setting called automatic adjustments. Switch it on and Meta applies recommendations for you, on a schedule that defaults to daily, by category. Meta's own opportunity score page promotes it as a "pro tip."
The trouble is that the categories bundle the good cards with the ones that need your judgment, and the changes happen without you looking. Merging ad sets and raising budgets both sit inside those categories, so a change you would have argued with can go live overnight and you only notice when the numbers move. If you want the easy wins, apply them by hand. It takes minutes.
A cousin of the opportunity score shows up when you build a new campaign. It is called campaign score, and it grades a single campaign before you publish it. Accept Meta's defaults and it reads 100. Switch off Advantage+ audience, trim a placement or turn off creative enhancements, and it drops. That is useful information, but the same rule applies: a lower score on a deliberate choice is not a mistake.
This also tells you something about reporting. If whoever runs your ads reports the opportunity score as a sign the account is healthy, ask what was dismissed and why. On Meta, a score in the high 90s often means most suggestions were accepted without much argument. What you want to hear about is cost per lead, booked jobs and sales, and that is what a good Facebook ads agency should be putting in front of you each month.
What a Healthy Account Looks Like
A well-run small account on Meta usually carries a score somewhere in the 70s or 80s, with a Dismissed tab full of decisions that each have a reason behind them. That is not an account falling behind. It is an account where somebody is reading the cards.
Here is a routine that takes about fifteen minutes a week:
- Open Account Overview and read every new card, starting with the highest point values.
- Apply the input cards: data, fresh creative, vertical versions, placements, duplicate ad sets, errors.
- For anything about budget, merging ad sets or changing the ad itself, open the preview and ask what each piece is for before deciding.
- Dismiss what does not fit, and keep a one-line note of why somewhere you will see it next month. The Dismissed tab remembers what you declined. It does not remember your reason.
- Confirm automatic adjustments are still off.
Then judge the account the way Meta itself says to, on your business goals: what you spent, how many real leads or sales came in, and what each one cost. The dial is a to-do list written from inside Ads Manager, and your business does not live there. Read it every week. Apply the cards that make Meta smarter. Make the money decisions yourself.
If you would like a second set of eyes on which cards in your account are worth taking, that is a conversation we have with small business owners every week, and we are happy to have it with you.
