Facebook Ads Not Converting? A Diagnostic Decision Tree for 2026

When Facebook ads stop converting, the instinct is to change the creative or move the budget around. Most of the time that is guessing, and it is guessing at the wrong layer. Ads that get impressions and clicks but no sales are almost always breaking at a specific point in the chain, and there is an order to check them in that saves you from fixing things that were never broken. This is that order: a decision tree you can run top to bottom before you touch a single setting.

Start here: is anything actually broken

Before diagnosing, rule out the two false alarms that send people chasing problems that do not exist.

The first is judging too early. A new ad set spends its first days in the learning phase, where delivery is unstable on purpose. Reacting after 20 euros of spend and a single day is how advertisers kill ad sets that were about to work. Give it enough spend to produce a real sample, past the learning phase, before you call it a failure.

The second is attribution lag. Meta restates recent conversions for up to about 72 hours, and much of your true conversion volume can land after the click through the attribution window. Judge performance on data that is at least 3 days old. A campaign that looks dead today can look fine once the window closes.

If you have cleared both of those and the ads genuinely are not converting, work down the tree.

Step 1: is your conversion tracking actually working

This is first for a reason. If conversions are happening but not being recorded, everything downstream is a lie. The optimizer is flying blind, your ROAS looks like zero, and you will spend days fixing creative that was never the problem.

Fire a real test conversion and confirm it appears in Events Manager under Test Events. Check that the pixel and the Conversions API are both sending the event and are deduplicated with a shared event ID, so you are neither missing conversions nor double-counting them. Confirm the ad set optimizes for the correct event on the correct dataset. A shocking number of accounts optimize for a pixel that fires on the wrong page, or for an event that stopped firing after a site redeploy.

The tell for a pure tracking problem is spend with genuine engagement and zero or near-zero recorded conversions across the whole account, not just one ad set. If that is what you see, stop here and fix tracking before anything else. The full checklist is in Facebook pixel not tracking purchases. Only once you trust the numbers should you continue down the tree.

Step 2: did something change right before it broke

If conversions fell off a cliff rather than slowly fading, something changed, and the fastest fix is finding what. Ninety percent of overnight drops trace to a specific edit or event.

Open the account change history and look at the days around the drop. A budget change beyond about 20 percent, a targeting edit, a new creative, or a bid change all reset the learning phase and can tank delivery while the ad set relearns. An accidental audience swap or a paused winning ad does it too. Outside the account, a broken checkout, an out-of-stock hero product, or a tracking change on the site will all show up as ads that suddenly stop converting.

If the break was sudden, the step-by-step version of this is in Facebook ads ROAS dropped overnight. If performance instead decayed gradually over weeks, skip ahead to the fatigue and audience steps, because slow decline is a different animal from a cliff.

Step 3: are you reaching the same people too often

Once tracking is trusted and no single change explains it, look at frequency. When the same people see your ads over and over, response falls and costs rise, and conversions quietly dry up even though impressions look fine.

Check cumulative frequency over the period, not daily frequency, which is a common reporting trap. On cold prospecting audiences, a cumulative frequency climbing past 3 in a 7-day window is a caution line. Pair it with a rising cost per thousand impressions and a falling link click-through rate, and you have audience fatigue rather than a bad week. The full method, including the reporting trap that makes frequency look far worse than it is, is in Facebook ad frequency and audience fatigue.

The fix is to give the audience something new or the ads a new audience: refresh the creative angle, broaden or swap the audience, and exclude recent purchasers so delivery stops re-hitting people who already acted.

Step 4: is the money sitting on the wrong ad sets

If individual ad sets convert but the account as a whole underperforms, the problem is allocation. Budget gets trapped on a mediocre ad set while a winner starves, and the blended result looks like nothing converts.

Compare cost per result across your ad sets. If one ad set takes a large share of spend while paying far more per conversion than your best one, that is recoverable budget. Under Advantage campaign budget this can still happen, because the system will keep funding an ad set whose short-term signals look acceptable, so pausing or consolidating the weak one is still your job. Move budget toward the efficient ad sets in steps of about 20 percent so you do not reset learning.

Step 5: is it the ads or the page

By this point tracking is sound, nothing broke suddenly, frequency is under control, and budget is allocated sensibly. Now separate the two halves of the funnel, because this is where most remaining cases live.

Look at the click-through rate and the landing page conversion rate as two separate numbers. If the click-through rate is healthy but conversions are low, the ads are doing their job and the break is after the click: the page, the offer, or the checkout. If the click-through rate is also weak, the problem is upstream in the creative or the audience, and you loop back to steps 3 and the creative.

High click-through rate with low conversion is the single clearest signature of a funnel problem, and it has its own guide: high CTR but low conversions. The usual culprits are a page that loads too slowly on mobile, a message mismatch between the ad and the page, a price or shipping surprise at checkout, or forced account creation.

Step 6: does the offer actually convert

The last step is the hardest to hear. If tracking is clean, delivery is stable, the page is fast and matched to the ad, and it still does not convert, the problem may be the offer itself. Ads amplify demand, they do not manufacture it. A price the market rejects, a product without a clear reason to buy now, or a weak guarantee will not convert no matter how good the media buying is.

This is not a reason to give up, it is a reason to test the offer rather than the ads: a stronger guarantee, a bundle, a limited-time incentive, a clearer value proposition. If a promotion suddenly makes the same ads convert, the media was never the problem.

The order matters more than any single fix

The reason to run these in sequence is that each step assumes the ones above it are clean. Fixing creative while your pixel is broken is wasted effort. Chasing the landing page while frequency is at 6 is treating a symptom. Work top to bottom: tracking, recent changes, frequency, budget allocation, ads versus page, then the offer.

Running this by hand across a full account is slow, which is the whole reason we built FixAds. It reads the account through the Marketing API, checks all six layers at once, and returns the findings in the order you should act on them, each with the number behind it and the exact fix. Your first finding is free, and it will usually tell you which step of this tree you are stuck on before you have finished reading it.