Why Your Facebook Ads Stopped Working: Ad Fatigue, and Why One Good Ad Isn't Enough Anymore
The ad that worked brilliantly in August is probably costing you money now.
Nothing about it changed. Same photo, same words, same budget. But the enquiries have slowed, each one costs more, and you're starting to wonder whether Facebook ads just don't work for your kind of business.
They probably still do. What's changed is how Meta decides who sees your ad, and how quickly a single ad wears out as a result.
Meta now picks the audience by looking at your ads
A few years ago, the skill in Facebook ads was targeting. You told Meta exactly who to show your ad to: women aged 30 to 55, within 20km, interested in gardening. Then you ran one or two ads at that group.
That's mostly gone. Meta rebuilt the system that picks which ads go to which people, and it now leans heavily on the ad itself to work out who it's for. It reads your picture, your video, your words, and finds people who are likely to respond to that. In plain terms, your ads now do most of the job that targeting used to do.
So if you only run one ad, Meta only has one way to find customers. A photo of a finished deck with "Get a free quote" will find the people who respond to that, and once it's shown to most of them, it runs out of road.
That's also why ads tire faster than they used to. Industry reports this year put the useful life of a typical Meta ad at around 2 to 4 weeks, down from 6 to 8 weeks before the change. Meta is good at finding the right people quickly, which means it also gets through them quickly.
"Different" means a different reason to buy
This is where most small businesses go wrong when they hear "run more ads". They make five versions of the same ad: same photo with a blue background, a green background, a slightly different headline. Meta treats those as basically one ad. It groups ads that look alike and only really shows one of them.
What counts is a different idea. Different reasons someone might pick you, shown in different ways.
Take a dog groomer in Palmerston North. One ad isn't enough, but these four would each reach a different kind of owner:
A before-and-after video of a matted spoodle coming out fluffy. Speaks to owners who are a bit embarrassed by the state of their dog.
A short phone video of the groomer talking about how she handles nervous dogs. Speaks to owners of anxious or rescue dogs.
A photo of a happy customer's review on a plain background. Speaks to people who want proof before they book.
A simple text ad about the offer: "First groom includes a free nail trim and ear clean. Book online in 2 minutes." Speaks to people ready to go now.
Each has a different hook (the first thing that grabs attention) and a different format. Meta can now find four groups of people instead of one.
You don't need a film crew. Phone video often beats polished work on Facebook and Instagram because it looks like something a real person posted. A tradie walking round a finished job talking for 20 seconds is a perfectly good ad.
How many ads a small business actually needs
You'll read advice telling you to run 20 to 30 ads at once. That's aimed at businesses spending thousands a day. On a small budget, too many ads means none of them gets enough money for Meta to learn anything.
For a small account, say under $100 a day, this setup works well:
One campaign, one or two ad sets. An ad set is the group inside a campaign that holds your budget and your ads. Fewer of them means your budget isn't split too thin.
Six to ten genuinely different ads in that ad set to start with.
Three to five new ones every three to four weeks, replacing the ones that have faded.
Bigger accounts have shown the same thing on a larger scale. In one 2026 test by the agency Logical Position, a single ad set holding 25 different ads got 17% more conversions at 16% lower cost than the same spend split across five ad sets with five ads each. Fewer buckets, more variety inside them.
If you're still setting up the account itself (which campaign goal to choose, how much to spend, how tracking works), we've laid all of that out in our guide to Google and Meta ads. This post is just about keeping the ads fresh once it's running.
How to tell an ad is worn out
Ads rarely die overnight. They fade, and there are three numbers that show it, all in Ads Manager:
Frequency is the average number of times each person has seen your ad. Around 2.5 is when to start watching closely. Past about 3.5, the ad is usually costing you more than a fresh one would.
Click-through rate is the share of people who see the ad and click it. Compare it with how that same ad did in its first week, not with some industry average.
Cost per result is what you pay for each enquiry, booking or sale.
When frequency is climbing, clicks are dropping and your cost per result is going up at the same time, the ad is tired. One on its own can be a blip.
The trick is to have the replacement ready before the old ad crashes, not after. If an ad has been running three weeks and the clicks are starting to slip, put the new ones in now.
A 30-minute routine to keep ads fresh
This is what we'd suggest for a small business doing its own ads:
Once a month, spend half an hour making new ads. Look at which of your current ads did best and ask why. Was it the before-and-after? The face talking to camera? Make two more in that style, plus one or two in a style you haven't tried.
Keep a running list of ad ideas on your phone. Customer questions are gold here. If three people this month asked "do you do weekends?", that's an ad.
Film as you work. Ten seconds of a job in progress, a finished result, a quick "here's what we found under the floor". Save them in one album. That album becomes your ad supply.
Turn off the losers, don't fiddle with them. If an ad hasn't produced anything after a fair amount of spend (roughly what you'd pay for two or three enquiries), switch it off and let a new one take its place.
Leave the winners alone. Resist the urge to edit an ad that's working. Changing it can reset what Meta has learned about it. Add new ads beside it instead.
None of this is technical. It just takes regular effort, and that's the bit that falls over when you're flat out running the business.
