Which Meta opportunity score recommendations should a B2B advertiser actually apply?

TofuAnswer by the Tofu team · checked

A hand ticking off items on a handwritten checklist in a notebook.

Apply the fixes: errors, ad sets Meta says to combine, and the Conversions API. Switch to the qualified leads goal if your CRM sends lead stages back. Use Advantage+ audience and placements once your account controls are set, then check lead quality. Hold off on budget increases and landing page view goals until your CRM shows the leads qualify, and preview AI creative before you accept it. The score only counts how many recommendations you apply: Meta says it "does not reflect your actual or future performance," and every one you dismiss keeps it below 100.

Ask Tofu next

Pick a question and Tofu Ads Agent answers it in chat.

As of October 2026.

Quick answer by situation

If you…UseWhy
See errors or warningsFix them firstMeta adds no points until every error is resolved
Get a recommendation to combine ad sets or reduce auction overlapApply itFewer ad sets leave the learning phase sooner
Get a Conversions API, CRM or qualified leads recommendationApply it if you meet Meta's fit guidelinesMeta can then look for leads that become sales
Get an Advantage+ audience or placements recommendationApply it after setting account controlsCheaper leads aren't always qualified, so check your CRM
Are told to raise a budgetCheck your CRM firstMeta judges on cost per result, usually a form fill
Are told to switch to landing page views or another goalOnly for an event you'd pay forMeta finds more of whatever you optimize for
Get AI creative, music or image expansion recommendationsPreview, then apply what stays on brandMeta's AI changes what people see
See your score dropRead the new recommendations and keep campaigns runningMeta says the score isn't a performance measure

Does Meta's opportunity score matter?

Only as a to-do list. The 0 to 100 score "is based on how many Meta Ads Manager recommendations you apply," and each recommendation's points depend on your business type, objective, budget and how it did in Meta's experiments. In Meta's example, fixing audience fragmentation is worth 35 points and Advantage+ placements 5. Meta says the score "does not reflect your actual or future performance" and "you should not turn off campaigns with a low opportunity score." Dismissing a recommendation doesn't change the score, so if you can't apply some, it stays below 100.

The score shows in the Account Overview and Campaigns tabs of Ads Manager. The campaign score you see while drafting a campaign is separate and doesn't change it.

If you see errors or warnings

Fix these first. They cover issues "such as incorrect formatting, missing information or policy violations." A campaign gets no points until all its errors are resolved, and if the ad account itself has an error, no score shows.

If Meta says to combine ad sets or reduce auction overlap

Usually apply it. B2B accounts often split one small audience into many ad sets, one per job function, industry or offer. When several of your ads enter the same auction, Meta lets only the one with the highest total value compete, and Meta says this can keep ad sets from getting the results they need to leave the learning phase (usually about 50 in the week after the last significant edit). To see results by segment, use breakdowns in reporting instead, as Meta suggests. The combined ad set has new targeting, a significant edit, so it learns again.

If Meta suggests the Conversions API, CRM integration or the qualified leads goal

Apply these when you fit. Meta's developer docs list recommendations to add the Conversions API if you only use the Pixel, connect your CRM if you run Instant Form lead ads, send more complete match data when event match quality is low, and choose "Maximize number of conversion leads." That's the conversion leads goal in the developer docs and Qualified leads in Ads Manager.

That goal works only with Instant Forms. Meta's fit guidelines: at least 200 leads a month, CRM data uploaded at least daily, and a lead stage that happens within 28 days and converts 1% to 40% of leads. With Zapier or another partner integration, Meta estimates 3 to 4 weeks to full value, including 2 to 4 weeks of learning. Its developer docs say applying the recommendation duplicates your ad sets and ads with the new goal, so the copies start learning from scratch. With website forms, the Conversions API recommendation still applies.

If Meta suggests Advantage+ audience or Advantage+ placements

Apply them after you set limits, then watch lead quality. Advantage+ audience lets Meta's AI find the audience instead of you; Meta reports 9.7% lower cost per result for the Traffic, Engagement and Leads objectives and doesn't recommend it for retargeting. Advantage+ placements run ads across Facebook, Messenger, Instagram, WhatsApp, Audience Network and Threads, and cut cost per action by 11.7% in a Meta experiment. Meta is gradually removing placement choice in ad sets for Leads, App promotion and Sales campaigns anyway.

Set the limits in Advertising settings first. These account controls apply to every campaign and include locations, a minimum age, audiences to exclude for brand protection, and My business doesn't show ads to its own employees. To keep ads off Audience Network, turn on My business can only advertise on specific placements and uncheck the Audience Network options. Existing campaigns can take up to 48 hours to update. Then compare lead quality in your CRM for a few weeks: cheaper leads from outside your market aren't a win.

If Meta suggests raising a budget

Check your CRM first. Meta's developer docs say "Scale your high-performing campaign" picks ad sets with stable delivery and "a lower cost per result" than others with the same optimization goal that "you or your peers have run." In most B2B accounts a result is a lead, not a qualified one. If cost per qualified lead looks good, raise the budget in steps: Meta says $100 to $101 isn't likely to restart learning, while $100 to $1,000 may.

If Meta suggests a different performance goal

Apply it only if the new event is one you'd pay for. Maximize number of conversions suits you when the conversion is a real lead, such as a demo request. Maximize landing page views finds people likely to visit your website, so you get visitors, not form fillers.

If Meta suggests AI creative, music or new creative

Preview each one in every placement. Meta's list includes Advantage+ creative enhancements that deliver variations of your ad, automatic music, image expansion with generated images, AI backgrounds and extra text options. Generated expansion adds content that wasn't in your original, which matters for product screens and charts. Extra text options are lower risk when you write them yourself.

Act on creative fatigue, which small B2B audiences hit often. Meta shows Creative limited when an ad's cost per result is above your past ads but under twice as much, and Creative fatigue at twice or more. Its fix is a new ad with a materially different image or video, and it says keeping the original running may maximize results.

If your opportunity score dropped

The score moves when you apply recommendations, change your account, campaigns, ad sets or ads, or when Meta finds new recommendations, so a new campaign can lower it without results getting worse. A dismissal covers only that campaign, so the same recommendation can return for another. To see what moved it, the Marketing API's opportunity score history endpoint returns the daily score for up to 45 days with a changelog: budgets, recommendations, and campaigns switched on or off.

What to avoid

  • Applying everything to reach 100. Budget, audience, placement and goal recommendations change who you reach and what you pay.
  • Judging Meta by the score or by cost per lead alone. Use qualified leads and pipeline from your CRM.
  • Making several big changes at once. Changes to targeting, creative, optimization event or bid strategy restart learning, and you won't know which one helped.

Where Tofu fits

Tofu Ads Agent (our product) plans, builds and optimizes Meta campaigns alongside LinkedIn and Google, and connects to HubSpot or Salesforce to read pipeline and closed-won and closed-lost deals. Its weekly bid, budget and creative recommendations stay proposals until a person approves them, and campaigns it builds are staged switched off. Tofu doesn't read or apply Meta's opportunity score recommendations, and it doesn't set up Meta's Conversions API or CRM integration.

Sources (17)

Related answers

More on