← All posts
Creative testing14 November 20267 min read

Creative fatigue is a math problem. Here's the math.

Your ad didn't break. You used it up. Every ad only has so many people who will ever care about it, and once Meta has shown it to most of them, results drop. That's creative fatigue. The fix isn't a better ad. It's more ads, on a schedule.

What creative fatigue actually is

Creative fatigue is when the same ad has been shown to the same people so many times that they stop noticing it. The ones who were going to buy already did. Everyone else has learned to scroll past. So Meta has to work harder, and pay more, to find anyone left who'll respond. Your cost per result climbs.

Meta even tells you when it happens. In the Delivery column of Ads Manager, an ad gets marked Creative fatigue when its cost per result is at least twice as high as it has been, and Creative limited when it's higher but less than double (Jon Loomer has a good breakdown).

Read that again: twice. By the time Meta flags it, you've already been overpaying for a while.

Cost per result: one ad versus a steady stream of new ads Illustration. The line for one ad running alone starts low and climbs steeply after week 3. The line for an account that adds new ads every week stays roughly flat. 2× = Meta flags "Creative fatigue" Week 1Week 3Week 5Week 7Costperresult One ad, running alone →New ads every week
Illustration, not real account data. The shape is what matters: one ad climbs, a pipeline of new ads stays flat.

The mistake almost everyone makes

They try to fix the tired ad.

New background. New button colour. Same headline, same photo, same idea. To your customer, that's the same ad they've already ignored. To Meta, it's close enough to look the same. You just paid to make one ad twice.

Here's the thing. Fatigue isn't a design problem. It's a supply problem. You don't need a fixed ad. You need the next one ready before this one dies.

The math: how many ads you actually need

Two numbers decide it. How long a winner lasts at your spend, and how often a new ad beats it (your hit rate). Then:

Ads to test per month = (4 ÷ weeks a winner lasts) × (1 ÷ hit rate)

Let's run it with simple, made-up numbers. Swap in your own.

Now look at what happens when you scale. Double the budget, and Meta reaches your audience twice as fast. Winners tire faster. Your "3 weeks" becomes something closer to 2, and your ad count goes up with it.

More spend needs more ads. Brands that scale budget without scaling creative don't hit a ceiling. They hit fatigue.

Not all "new" is new: three levels of change

When you make the next ad, how different it is matters more than how good it looks.

Level 1

New angle

A different reason to buy. Comfort, then gifting, then durability. This reaches people the last ad never moved, and it's the biggest reset you've got.

Level 2

New hook

Same reason to buy, new first line or first visual. Cheaper to make, and often enough to wake an angle back up. (Ten hooks to start with.)

Level 3

New edit

Same ad, new crop, format or colour. Useful to stretch a proven winner into 9:16 or a new placement. Useless as a fix for fatigue on its own.

A simple split to start with: most of your tests on new angles and hooks, the rest on edits of whatever is winning.

Spot it before Meta does

One warning: don't kill a new ad on day two. Give it enough spend to tell you something. Fatigue is about ads that worked and stopped, not ads that never got a chance.

The weekly system

  1. Monday: read the last 7 days. Turn off ads whose cost per result has blown past your target.
  2. Tuesday: pick the next angles and hooks to test, based on what's winning and what isn't.
  3. Wednesday: make the batch.
  4. Thursday: launch it into the account, next to the current winners.

It's boring. That's the point. Brands that do this every week rarely have a "my ads stopped working" month. (For the longer version of why ads die, read why your Meta ads stop working after a few weeks.)

The short version

Want to feel it instead of reading it? ROAS Lab is a free 90-second test: six weeks of Meta ads, with fatigue built in.

Questions

What is creative fatigue?

Creative fatigue is when the same ad has been shown to the same people so often that they stop noticing it. Meta then has to pay more to find the people who haven't tuned out, so your cost per result climbs. Meta flags it in the Delivery column of Ads Manager as Creative fatigue when cost per result is at least twice as high as it has been, and Creative limited when it is higher but less than double.

How long does a Meta ad last before it fatigues?

There's no fixed number. It depends on how much you spend and how big your audience is. A small budget can run one winner for months; a big budget can burn through it in a week or two. Watch cost per result on the ad, not the calendar.

How many new ads should a DTC brand test each month?

Use this: ads to test per month = (4 ÷ weeks a winner lasts) × (1 ÷ your hit rate). If a winner lasts about 3 weeks and 1 in 5 new ads beats it, that's about 7 ads a month. Round up to 10, because some tests never get enough spend to judge.

Does changing the colour or background fix creative fatigue?

Usually not. To your customer and to Meta, it's the same ad with a new coat of paint. Small edits can stretch a winner for a little while, but fatigue is fixed by new angles and new hooks.