On this page
What the Opportunity score is
Ads Manager shows a list of recommendations and an Opportunity score from 0 to 100. You’ll find it in Account overview and above the Campaigns table. It goes up as you apply the recommendations (Meta’s help article “About opportunity score”).
Read that carefully. The score tracks how many of Meta’s suggestions you’ve adopted. Meta itself says it doesn’t measure performance. A score of 100 tells you that you’ve taken Meta’s advice. It doesn’t tell you whether your ads make money.
Why it deserves a second look
Meta makes its money when you spend yours. The recommendations, the score and most of the default settings are built to get more ads running and more budget spent. That doesn’t make them wrong. Some of the advice is good. It just isn’t measured against your margin, your bookings or the jobs you win.
The typical suggestions:
- Raise your budget.
- Broaden your audience, or widen your radius.
- Add placements.
- Switch on more Advantage+ features.
Any of them can be right for you. All of them mean more spend, or more of your decisions handed to Meta. Some feed a quieter problem: Meta taking credit for sales you’d have made anyway. Advantage+ not working? covers that one.
The same advice, second-hand
AI chatbots and most “how to run Facebook ads” content repeat Meta’s guidance, because that’s what they learned from. So following an AI’s setup steps usually means following Meta’s recommendations, without ever seeing the score. If your campaigns were built by asking a chatbot, treat them as built on Meta’s defaults.
Two settings that apply recommendations for you
Auto-apply. Recommendations can be set to apply themselves, in Account overview. Changes can then land in your account without you choosing them. Turn it off.
Duplicating “with recommendations”. If you duplicate a campaign with recommendations, Meta applies them to the copy: structure, audience, creative enhancements, placements, and bid or schedule settings (Meta’s help article “About duplicating an ad campaign with recommendations”). Duplicate without recommendations, then check music, placements and enhancements before you publish.
The one-question test
Before you accept anything, ask: will this lower my cost per sale, booking or good lead? If you can’t say yes, skip it.
| Recommendation | Ask yourself |
|---|---|
| Raise your budget | Has cost per result been under target for 7 days? |
| Broaden the audience or radius | Will I actually ship, serve or drive there? |
| Add placements | Am I optimising for a sale, booking or lead, not clicks? |
| Turn on Advantage+ creative | Will it change my images or words? |
A few notes on each:
- Budget: our plans only scale after 7 days under target, by 20% at a time. Online stores also need MER (store sales ÷ ad spend) at or above target.
- Audience and radius: for a tradie or a café, “widen your audience” is the one to watch. Inside any area, Meta finds the cheapest people first, often on the outer edge.
- Placements: when you optimise for a real result, Advantage+ placements is fine, because Meta chases that result everywhere. When you optimise for clicks, cheap placements can soak up the budget with accidental taps.
- Creative: enhancements can adjust brightness, crop, add music and move your text. Preview before you accept.
How to test one properly
- Accept one at a time. Two changes at once and you can’t tell which one did what.
- Give it 7 days. Changes can restart learning, so judge after a full week.
- Judge on your own numbers, not the score: store sales and MER for an online store, covers in your booking records for a venue, good leads and site visits for a trade.
- Keep it or roll it back. If cost per result rose, go back to what held.
The result below shows why this matters. A client’s online store acted on recommendations that raised Meta spend from A$7,560 to A$11,340. Purchases rose, but cost per purchase climbed and MER fell from 5.22x to 3.87x, while store revenue barely moved. After a reset, MER returned to 5.23x. That’s what moved alongside each change, not proof of what caused it.
So should you ignore the score?
Ignore it as a measure of success. Read the recommendations as a list of ideas, test the ones that pass the question, and keep score with your own numbers. If you sell online, know your break-even ROAS first: it’s the line every change has to stay above.
The free 3-minute quiz asks whether you follow Meta’s recommendations or a chatbot’s setup, and marks it in your diagnosis if it’s likely holding you back.