- creatives
- Meta Ads
- optimisation
Why your ads cost more and more after a few weeks
Ad fatigue explained through the auction mechanism, the signals that announce it, and how to spin off a creative that works instead of letting it die.
Published on · Joh · 8 min read
Your campaign was running well. Same budget, same targeting, nothing changed, and for three weeks the cost per result has been climbing for no visible reason.
There is a reason, and it sits in the auction mechanism.
What is actually happening
Meta does not sell ad space to the highest bidder. It sells it to the best total score, which combines three things: what you are willing to pay, the estimated probability that people will react, and the perceived quality of the ad.
The first term is yours to control. The other two degrade over time.
The longer your ad runs, the more often the same people see it. Interactions fall, hides rise, the estimated action rate drops. Your total score falls, and Meta charges you more for the same slot.
This is not a punishment. It is an auction noticing that your ad interests people less.
Direct consequence: a good creative is cheaper to distribute than a bad one. The money you put into production, you get part of it back on the price of the space.
The signals, in the order they appear
Fatigue does not show up on cost per sale first. It shows up upstream, which is what lets you get ahead of it.
Frequency rises. The same people see you more often. It is the first sign, visible before any result degrades.
Hook rate falls. Your three-second views against impressions go down. People recognise the video and scroll past.
Click-through rate drops below 1%. By this stage the decline is established.
Cost per thousand impressions climbs. The auction has priced in the drop in interest.
Cost per result explodes. You discover the problem at the moment it costs you the most.
If you only watch that last line, you are three or four weeks behind what is happening.
A narrow audience tires faster
A business working within a 40-kilometre radius does not have the same reservoir as a brand selling across the country.
The smaller the area, the faster the rotation has to be. On a local audience I plan for monthly renewal. On a national audience with a large budget it can hold longer.
It is also why "we have a great video, it will last us a year" does not hold for local trades. It will last as long as it takes your area to see it.
Spin off rather than start again
When a creative works, it contains an angle that works. Throwing the whole thing away to start from a fresh idea means discarding the most expensive information you have paid for.
What to change, in order of effectiveness:
The hook. The first three seconds. Same video, five different openings, five different performances. It is the most profitable lever.
The person on screen. Owner, employee, customer. A different face revives attention on the same message.
The format. A demonstration becomes a testimonial, a before-and-after becomes a list.
The setting. Same script, another site, another light.
The soundtrack and the pace of the edit. The least powerful, but the cheapest to produce.
A validated angle easily yields eight variants. Thirty minutes of editing per variant, and you have enough to last two months without setting foot on another shoot.
The rhythm that avoids the wall
What I put in place, and which mostly takes discipline:
- Every week, read the upstream signals: frequency, hook rate, click-through rate. Not just cost per result.
- Always one creative ahead. When an ad tires, the replacement is already edited. Otherwise you let it run by default.
- One change at a time, with 72 hours minimum before judging. Changing something daily sends the campaign back into learning permanently.
- Judge by the week, not by the day. A bad day means nothing.
- Do not cut everything at once. Replacing gradually keeps the learning history alive.
What doing nothing actually costs
An ad that tires does not stop dead. It carries on running, carries on spending, and returns less and less. The spend does not fall on its own.
It is the argument I give clients who hesitate to produce new videos: the budget you do not put into creative, you lose in auction surcharge on the ad you keep too long.
Put another way: you pay either way. The only question is whether you pay for something you own.
Further reading: Meta Ads for an e-commerce brand and Freelance media buyer: what it is, what it changes.
Costs climbing with no explanation? . We look at your upstream signals and the renewal rhythm you need.