Meta Opportunity Score: Use Recommendations Without Optimizing to the Score
Updated: September 2026
Meta Opportunity Score is a 0–100 indicator based on the recommendations an advertiser has applied. Use it as a queue of hypotheses, not as a business KPI. A higher score can reflect closer alignment with Meta’s preferred setup while profit, incrementality, lead quality, and brand risk still require separate evidence.
What the score actually tells you
Meta describes Opportunity Score as a way to see how optimized campaigns are and prioritize recommended actions. Recommendations may involve Advantage+ features, creative, audience setup, placements, budget, or account configuration. Applying higher-impact recommendations can raise the score.
That makes the score operationally useful: it can surface settings a team might otherwise miss. It does not mean every recommendation is appropriate for every business. Platform recommendations are generated from broad patterns and account signals; your constraints may include margin, geography, inventory, sales capacity, regulated claims, or customer exclusions that the score does not fully represent.
Classify recommendations before acting
| Class | Examples | Default action |
|---|---|---|
| Low-risk hygiene | Broken destination, missing asset format, tracking warning | Verify and correct promptly |
| Reversible delivery change | Placement expansion, budget redistribution, bid setting | Test with guardrails |
| Signal or attribution change | Conversion event, value input, audience source | Require measurement owner review |
| Creative or brand change | Automated variation, generated asset, copy adjustment | Require brand and claims approval |
| Irreversible or high-impact change | Large budget increase or broad policy-sensitive expansion | Escalate and stage rollout |
Use an evidence gate for every recommendation
1. State the platform’s implied hypothesis
Translate the recommendation into plain language. “Expand placements” becomes: “Additional inventory will produce qualified conversions at an acceptable marginal cost.” If the hypothesis cannot be stated, the team cannot judge the outcome.
2. Identify the business constraint
Check the recommendation against exclusions, customer-acquisition policy, inventory, margin, sales coverage, and creative permissions. A suggestion can be statistically promising and still be operationally wrong.
3. Define the success measure and guardrail
Use qualified revenue, contribution margin, approved leads, or another business outcome. Add guardrails for frequency, brand safety, existing-customer share, refund rate, or sales rejection. Do not use the Opportunity Score itself as the test outcome.
4. Choose the smallest valid test
Prefer a controlled experiment or staged rollout over changing the entire account. Apply one recommendation family at a time when possible, record the date, and preserve the prior configuration for rollback.
When should you ignore a recommendation?
Ignore or defer it when the requested action conflicts with a documented business rule, relies on weak or duplicated conversion signals, cannot be evaluated with a suitable outcome, or would invalidate another live test. Record the reason. A lower score with intentional governance can be healthier than a perfect score built on automatic acceptance.
This approach extends our framework for evaluating Meta AI-powered recommendations. It also supports the system behind sustainable Meta AI scaling, where creative supply and CRM quality matter more than a dashboard badge.
Frequently asked questions
Is 100 the ideal Meta Opportunity Score?
Not necessarily. It means the account has applied the available scored recommendations, not that it has reached maximum profit or incremental growth.
Do recommendations apply automatically?
Some account settings can enable automated application, while others require action. Review permissions and change history so the team knows what is manual and what can change without a new approval.
How often should the score be reviewed?
Review it on a regular optimization cadence and after major campaign changes, but prioritize business alerts, tracking issues, and experiment integrity over chasing a daily score movement.