Creative

Spotting TikTok creative fatigue before CPA doubles

By the time cost per result moves, the creative has usually been dying for several days. Three earlier signals tell you first — if you look at them per creative rather than per account.

Freshlytics team · Updated August 2026 · 7 min read

Fatigue is a sequence, not an event

On TikTok, a creative rarely dies all at once. It decays in a fairly consistent order, and each step is measurable:

  1. Hold rate falls. The share of people still watching at three and six seconds drops first. The hook is no longer novel to the audience being reached.
  2. CTR follows. Fewer people reach the part of the video that motivates a click. CTR usually lags hold rate by a day or two.
  3. Frequency climbs. Delivery has to reach the same people more often to spend the same budget, because the auction is finding fewer fresh, responsive users.
  4. CPA moves last. By this point you have already spent real money learning something the first three signals told you earlier.

Most reporting setups only alert on the fourth step. That is the expensive one.

Measure each creative against itself

A creative with a 22% three-second hold rate might be excellent or terrible — it depends entirely on the format and the offer. Comparing it to an account average tells you almost nothing. Comparing it to its own first week tells you a great deal.

The working approach:

  • Take the creative's first 3–5 days of meaningful delivery as its baseline
  • Track the rolling 3-day value of hold rate, CTR and frequency against that baseline
  • Only score creatives that passed a minimum spend or impression threshold, so noise cannot trigger a verdict
  • Deduplicate: the same asset running in six ad groups is one creative, and its fatigue is a property of the asset, not of any one ad group

Rule of thumb we use: a creative whose 3-day CTR has fallen more than 25–30% below its own baseline, with frequency up meaningfully over the same window, is in decline — regardless of whether CPA has moved yet.

Refresh the hook or replace the asset?

The two failure modes need different fixes, and the metrics separate them cleanly.

PatternMost likely causeWhat usually works
Hold rate down, CVR of clickers stableThe opening has stopped stopping the scrollNew hook — first 1–2 seconds, new opening frame, different first line of copy. The body of the video can survive.
Hold rate stable, CTR downThe middle no longer motivates the click Rework the offer framing or the call to action; keep the hook that still works.
Both down, frequency up sharplyGenuine audience saturation for this asset Replace the creative and consider widening targeting or shifting budget to a different angle.
Everything stable, CPA upNot creative fatigue Look at pacing, competition, landing page, tracking or seasonality before you blame the video.

That last row matters more than it looks. A lot of creatives get killed for a problem that lived somewhere else entirely — and you lose a proven asset for nothing.

What fatigue looks like in a dashboard

app.freshlytics.site/creatives/fatigue
Creative fatigue radar
28 active creatives · hook retention & CTR decay, 14-day window
Rotate 4 ads
CreativeSpendCTR trend3s holdVerdict
hook_gym_v4.mp4live 19 days$9,340 21% −38%Replace now
ugc_review_02.mp4live 11 days$6,720 29% −11%Refresh hook
demo_split_a.mp4live 4 days$3,110 41% +16%Scale
static_offer_b.jpglive 26 days$2,480 18% −44%Replace now

Freshlytics creative fatigue view — illustrative sample data.

How fast should you rotate?

There is no universal number, and anyone quoting one is selling something. What is reliable is the shape: the higher your spend against a fixed audience size, the faster frequency accumulates and the shorter the useful life of a creative. A broad-targeted ad group at modest spend can run the same asset for weeks; a narrow retargeting pool at high spend can exhaust one in days.

Practical version: stop rotating on a calendar, and start rotating on the curve. Build the habit of always having the next hook ready before the current one crosses its threshold, so a replacement is a scheduled move rather than an emergency.

Test the replacement honestly

When you swap a creative, keep track of what it replaced. The question is never "did the new one work?" but "did it beat the one it replaced, at comparable spend and in a comparable window?" Without that comparison, a rotation habit can quietly degrade your account while everybody feels productive.

In Freshlytics: every creative is scored against its own baseline, deduplicated across ad groups and accounts, gated behind a minimum spend, and shown with the asset it replaced. You get a verdict — scale, refresh the hook, or replace — instead of a wall of metrics. Request beta access.

← All guides