Why Your Facebook Ads Aren't Working (and How to Fix Them)

There is a particular kind of quiet that comes about three weeks into a campaign that is not landing. Money is going out daily. The numbers in the dashboard mean nothing to you. Nobody has rung.

The instinct at that point is to change the picture, or widen the audience, or turn the budget up. Sometimes one of those is right. Far more often it is the equivalent of hitting the telly, and it resets whatever the campaign had learned in the process.

So before touching anything, spend twenty minutes finding out what is actually broken.

Facebook ads not working is three problems, not one

Every campaign is a chain with three links in it. Your ad gets shown to somebody. That somebody taps it. Having tapped it, they do the thing you wanted.

A campaign fails when exactly one of those links breaks, and which one it is decides everything you do next.

Working out which needs no software beyond what you already have. Open Ads Manager, set the date range to the last seven days, and read three columns: how many times it was shown, how many taps that produced, and how many results came out of the far end. Meta's own page for the tool was checked on 2 September 2026.

Notice how differently those four end. Turning the budget up rescues a delivery problem and makes a wasted-clicks problem cost you more, faster. That is why guessing is expensive.

Problem one: the ads are barely being shown

The giveaway is spend. If you set ten pounds a day and the campaign has managed four, nobody is ignoring your advert, because hardly anybody has seen it. This is the most misdiagnosed of the three and the quickest to fix, because the causes are mechanical rather than creative.

Work down this list in order, because the first two catch most of it.

1. Check whether it is still in review, or was rejected. Meta screens adverts against its published advertising standards before they are allowed to run, and a rejection sometimes lands with no email you actually notice. Health, weight loss, cosmetic and finance wording trips this constantly, and beauty clinics get caught more than anyone.
2. Look at whether the campaign, the ad set and the advert are all switched on. Three separate switches, and turning the top one on does not turn on the ones underneath. It happens to everybody once.
3. Check the payment. A declined card pauses everything, quietly, and the notification goes to whichever email set the account up years ago.
4. Look at how small the audience has become. A three mile radius, narrowed by age, then narrowed again by interest, can leave a pool too thin to spend even a modest daily budget in.
5. Check the schedule and the end date. Campaigns that finished a fortnight ago are a remarkably common cause of nothing happening.
6. See whether you are bidding against yourself. Two or three ad sets aimed at more or less the same people compete in the same auction, and you pay for the privilege.

Problem two: plenty of impressions, nobody tapping

Now it is genuinely being seen and people are scrolling past it. This is the one everybody assumes they have, and it is the one where the wrong instinct does most damage, because the instinct is to change the photograph.

The photograph is rarely the problem. What is being ignored is nearly always the offer, and there is a hard version of that sentence worth sitting with: if the thing you are offering is only interesting to somebody who already knows and likes you, it will die in front of strangers. Almost everybody seeing this advert has never heard of you.

So the questions in order are: is there an actual reason to stop, is it obvious in the first line, and would somebody who has never heard your business name care about it? "We offer a range of services" fails all three. "Boiler died? We can be there today" passes all three, and the difference is not photography.

Two specific causes of this are already written up properly elsewhere, so we will point rather than repeat. If you used the Boost button, the thing you asked Meta to go and get largely decides who it shows the advert to, and what the Boost button quietly picks for you explains why boosting for engagement fills your account with people who engage and never buy. And if it worked for six weeks and then faded, that is a different animal again: the reasons a campaign gets dearer over time covers creative fatigue, small audiences and the cost of editing an ad set that had already settled.

One thing worth doing before you write anything new. Go and look at your own feed for two minutes and notice what made you stop. It will not have been a nicely lit product photograph. It will have been something that looked like a person, or a price, or a problem you recognised.

Problem three: they tap, and then nothing

The advert did its job. You paid for the tap. Whatever happened next is where the money went.

Three things break here, and in our experience it is nearly always the first.

The promise does not match the destination. Somebody taps an advert about a specific offer and arrives on a homepage that mentions eleven services and not that one. They have about four seconds of patience and they were never that committed. Send paid traffic to the page about the thing you advertised, every time.

The page asks for too much too early. A stranger who has known you for nine seconds will not fill in eight fields. Ask for the least you need to be able to ring them back.

And nobody answers. Enquiries from paid advertising go cold faster than any other kind, because the person is comparing and you are one of three tabs. An enquiry answered on Thursday for a Monday tap has already been served by somebody else.

For what a page has to do to convert traffic once it arrives, whether that traffic is paid or not, the full diagnostic for a website that is not producing enquiries works through it properly and there is no sense repeating it here.

A worked example of what a paid click should land on. KAY Clinics is a prescriber-led aesthetics clinic in Wickersley, Rotherham, and their site is in our portfolio. Somebody arriving cold finds the GPhC registration, 2230779, printed in full so they can check it on the register themselves, the whole treatment list priced openly with no "from" figures hiding the real number, and a five-step booking flow that runs at any hour without anybody having to send a message and wait. Aesthetics is a market where anyone can hang a sign up, so every one of those answers a doubt a stranger arrived with. That is what a destination is for.

The fourth possibility: it is working and you cannot see it

This one is worth checking before you switch anything off, because we have watched people kill campaigns that were quietly earning.

Ads Manager can only report what it is able to observe. If nothing was ever installed on your website to tell it what happened after the tap, the results column shows a row of zeroes whatever the truth is. And if your customers ring you, which for most local trades and clinics is exactly what they do, that call is invisible to the platform by design.

The Meta pixel is the piece that closes the loop, and Meta's own description of it is straightforward enough: "Measure the actions and outcomes that matter most to your business, helping you optimise campaigns for real impact." Without it the platform is also flying blind while it decides who to show your advert to next, so this is not purely a reporting problem.

Three ways to find out today, none of which need a developer.

• Ask every single person who contacts you this fortnight where they found you. Tedious, unarguable, and it takes ten seconds a call.
• Compare enquiry volume with the fortnight before the campaign started. Crude, but a real rise nobody can attribute is still a real rise.
• Check whether your enquiry form records where the visitor came from. Plenty do and nobody ever looks.

Four things that are probably not your problem

Worth clearing these out of the way, because they absorb a lot of worry and rarely deserve it.

• The algorithm being against you. It has no view about your business. It is a pricing mechanism with no memory of you personally.
• Not having enough adverts. Splitting a modest budget across five variations is the most reliable way to ensure none of them gathers enough response to improve.
• Your budget being small. Small budgets work. Small budgets spread thin, or aimed at the wrong outcome, do not, and those are different faults.
• Needing to post more organically first. Useful, but unrelated. A quiet page does not stop a paid advert from being delivered.

The twenty-minute check, in order

Do it in this sequence. Each step only matters if the one before it passed.

1. Is it approved and switched on at all three levels, with a working card on file?
2. Did it spend roughly what you told it to over the last seven days? If not, stop here and fix delivery.
3. Are people tapping? Compare this week against your own last week, not against somebody's published average, which will have been produced by advertisers with nothing in common with you.
4. Where does the tap land, and does that page mention the exact thing the advert promised?
5. Is anything at all installed to record what happens after the tap?
6. How fast does an enquiry get answered, honestly, measured on the slowest one this month rather than the quickest?

Whichever step you stopped at is your problem. Fix that one thing, leave it a fortnight without fiddling, and look again.

When to hand it to somebody

You do not need an agency to run that check, which is why it is printed above in full rather than dangled as a reason to ring us.

Where it is worth paying somebody is when you have run it, fixed the obvious thing, and the campaign is still flat. Our paid advertising management is £300 a month, flat. Meta takes the ad budget off your own card and it never reaches our account. The fee holds steady at any level of spend, so we gain nothing at all by talking your budget upwards.

We would also rather tell you the campaign is fine and the phone is not being answered quickly enough. That happens more often than anybody selling advertising likes to admit, and it is free to fix.

Frequently asked questions

Diagnose first, then change one thing

Almost everybody whose Facebook ads are not working changes three things at once, out of impatience, and then cannot tell which of them helped. A fortnight later they are back where they started with less money and no more information than before.

Find the broken link. Fix it on its own. Leave it alone long enough to say whether it worked.

If you have been through the six steps and the numbers still are not telling you which of the three you have got, send us a screenshot of the ad set covering the last month, with the page you are sending people to. We will tell you which link is broken. If it turns out to be a setting you can change yourself in five minutes, we will say that rather than quote you for it.