there's a specific moment i hear about on almost every first call. "we had an ad that was working — properly working — and then it just stopped. we didn't touch anything."
i believe you. you didn't touch anything. that's actually the problem, and it's worth understanding why, because the thing that killed your ad has changed a lot in the last year and most business owners are still running the old playbook.
your ad didn't break. it wore out.
every ad has a shelf life. once meta has shown it to most of the right people a few times, there's nobody fresh left to show it to. frequency climbs, clicks fall, and the cost of every sale creeps up until the ad that was carrying your account is quietly draining it.
that's always been true. what's new is the speed. meta's current algorithm (the andromeda era, if you like the jargon — you don't need it) decides who sees your ad largely based on the creative itself, and it burns through an ad's audience in two to three weeks. a couple of years ago a winner might have run for two or three months. now the industry data is pretty consistent: concepts fatigue in weeks, and engagement on a fading ad drops sharply week after week.
so if your results "suddenly dropped off" — the ad probably didn't break on a tuesday. it faded on schedule. the schedule just got shorter than anyone told you.
the expensive part isn't the fading. it's the not-noticing.
a dying ad doesn't send you a message. it keeps spending exactly as before. i recently watched an account where the winning creative died on roughly day 14 — frequency spiked, click-through halved — and then spent another four weeks and about £1,100 running to an audience that had already said no.
nobody did anything wrong, in the sense that nobody did anything at all. the account was set up months ago by someone competent, and then everyone went back to running the actual business. which is reasonable. it's just expensive now, because the platform moved from "set up well, check monthly" to "feed constantly, check weekly."
what watching for it actually looks like
if you run your own ads, the whole early-warning system is two numbers, once a week.
frequency: how many times the average person has seen your ad. past 3-4 and climbing, with results worsening, means the ad is tired.
click-through rate: falling week over week while frequency climbs is the confirmation. fatigue shows up in these two before it shows up in your cost per sale — which means you can act while the ad is still profitable, instead of two weeks after it stopped being.
and the response isn't "more budget" — more spend on a tired ad just pays extra to annoy the same people. the response is the next creative. new angle, new opening line, new footage, same offer. not fifty ads. just the next one, ready before you need it.
the uncomfortable conclusion
meta's own numbers now put creative quality at over half of what determines campaign performance — more than targeting, budgets and placements combined. which means running ads well in 2026 isn't really a setup task anymore. it's a weekly pipeline: make concepts, watch the leading indicators, kill winners while they're still winning, repeat.
some owners genuinely enjoy doing that themselves, and the two-numbers-weekly habit above will carry you a long way. but if your ads keep "suddenly stopping" and each time it costs you a few quiet weeks of wasted spend before anyone notices — the maths on having someone whose actual job is watching starts to look different.
either way: go look at your frequency column this week. if your best ad is past 4 and climbing, it's not your best ad anymore.
0 comments