Key takeaway
If your Facebook ads are not working, do not start by changing everything. Start by checking where the funnel broke first: delivery, CTR, CPC, landing-page conversion rate, or profitability.
Start with the business target, not the dashboard
Before looking at CTR or ROAS, define the real target. That usually means target CPA, break-even ROAS, or acceptable payback period. If you do not know what success looks like economically, every metric becomes easy to misread.
This matters even more for beginners. A campaign can have weak-looking CTR and still be profitable. Another campaign can have cheap clicks and still lose money because the traffic does not convert.
Check the funnel in order
- Is spend or delivery unstable?
- Is CTR weak, which suggests a creative or hook problem?
- Is CPC rising because the ad is losing relevance or the auction is getting harder?
- Is the landing page converting well after the click?
- Are CPA or ROAS actually acceptable for the business?
- Did recent edits reset learning and make results noisy?
Common beginner mistake
Most advertisers react to the last metric that got worse. If ROAS dropped, they blame targeting. If CPC rose, they blame Meta. Usually the real issue started one step earlier, often in creative or post-click conversion.
What different failure patterns usually mean
Low CTR
This usually points to weak creative, weak hooks, or weak audience-message fit. If people are not clicking, the ad is not earning attention from the audience it is reaching. Check what good CTR looks like.
Good CTR but no conversions
This usually points to the landing page, the offer, or low-intent clicks caused by over-promising creative. If people click but do not buy or submit, the problem is often after the click.
High CPC and falling CTR
This often points to fatigue, weaker relevance, or audience saturation. The auction gets more expensive when the ad is no longer competitive.
Good top-of-funnel signals but weak profit
This points to margin problems, low order value, wrong optimization choices, or attribution confusion. A campaign can look fine in Ads Manager and still be a weak business result.
Do not judge one bad day
One bad day means almost nothing by itself. Read the last 7 days before making structural decisions. Good accounts at meaningful spend regularly have random down days and recover with no edits at all.
What to do next
- Replace creative if the problem starts at CTR.
- Fix the landing page or offer if the problem starts after the click.
- Simplify structure if multiple ad sets are stuck in learning or spending inconsistently.
- Only pause when the campaign is clearly below economic threshold and there is no believable recovery path.
The working rule
Facebook ads usually are not "not working" everywhere at once. Something broke first. Find that first broken link and the next action becomes much clearer.
The cause almost nobody checks first: what your pixel has learned
Meta optimises towards whatever you told it to find. If an account spent weeks on traffic or engagement campaigns "to build up data", the pixel is now full of people who click on everything and buy nothing, and the algorithm is faithfully finding more of them. No amount of new targeting fixes that, because targeting is not the broken part.
An agency that has audited hundreds of accounts sorts them into three states before recommending anything:
- Polluted. The pixel has plenty of events, but they are the wrong ones. Fix: optimise for purchases only, accept the volume drop, and let the old signal age out.
- Thin. Good events, but under roughly 50 conversions a month. Fix: chase volume, not efficiency — do not add ROAS goals or cost caps to an algorithm that is already guessing.
- Healthy. 50+ real purchases a month. Only here do tighter targets, cost caps and advanced retargeting behave the way the case studies promise.
Open Events Manager and look at what your pixel actually recorded in the last 28 days. Five thousand landing page views against twelve purchases tells you which state you are in, and which of the fixes below are even available to you.
What operators report
The single most common self-diagnosis error is copying a strategy from someone in a different state. As one practitioner put it: "Most of the strategies you see shared online assume you’re in Stage 3. When people try to apply them in Stage 1 or 2, they wonder why nothing works."
The counter-argument, recorded: a polluted pixel does not simply flush itself clean, and 50 conversions a month is directional rather than a switch — accounts at 40 can optimise fine, accounts at 60 can still struggle, depending on order-value spread and audience size.
Check the cheap things before the expensive ones
New creative is the most expensive answer to a performance dip and, in the accounts practitioners describe, rarely the right one. Run these in order, because each takes minutes and costs nothing:
- Break results down by placement. Look for a placement absorbing budget without producing conversions. Reels and Audience Network come up most often; one account recovered almost immediately after removing Reels. (Recorded dissent: excluding placements restricts delivery and can hurt — test it rather than assuming.)
- Stop reading the blended number. Split by ad set, ad, country, age. A retargeting campaign sitting at 3x blended was 7–9x in Canada and dragged down by the US. That is a budget allocation problem wearing a creative problem's clothes.
- Check the conversion event at the ad set level. A campaign objective of Sales with add-to-cart selected on the ad set is a common silent error.
- Audit your exclusions. A stale customer list can quietly remove a third of your viable reach, and Meta does not warn you when an exclusion audience breaks.
- Relaunch before you reshoot. Duplicate the campaign with the same ads, budget and targeting, and turn off the original. Campaigns accumulate delivery history from every edit; a clean duplicate resets that without changing anything about the creative.
Only when all five come back clean is new creative the right spend. Our guide on telling real fatigue from false positives covers what to look for.
Is it you, or is it Meta?
Sometimes it genuinely is the platform. The signature of a platform-side event rather than an account-side one: it starts abruptly rather than degrading, it hits several unrelated accounts at once, spend often continues while conversions do not, and traffic quality drops without any change in your targeting. If your own account changed nothing and the numbers fell off a cliff on a specific afternoon, look outward before rebuilding anything. Our guide on ads that suddenly stopped working goes through how to tell the two apart.
AskAds can diagnose this automatically
Instead of manually checking every metric, AskAds can walk your account through the same diagnostic order and tell you what probably broke first. Try it free →