Is your TikTok creative tired, or has something else changed?

The ad that carried last month’s acquisition is getting more expensive. It has been running for a while, so creative fatigue sounds plausible. The next brief practically writes itself: new hook, new creator, new batch.

Before commissioning that batch, establish what has become less effective. An asset can look tired because it is reaching different people, buying more expensive impressions, or sending users into a weaker app experience. Replacing it might temporarily change delivery without fixing the underlying problem.

Start this review after you have aligned reporting definitions and cohort age. The question here is narrower: does the evidence point to this creative losing effectiveness in a comparable context, and which part deserves a refresh?

Make one row for each meaningful version of the asset. A different opening, caption, language, or offer is a different version for this purpose. Keep the creative identifier alongside a human-readable name so similar uploads do not get merged accidentally.

Then record the setting in which it ran: app, operating system, country, language, campaign objective, optimization event, placement where available, and acquisition dates. Note material changes to targeting, bids, destination, and app version.

You do not need every possible breakdown. Begin with the contexts that changed enough to explain the result. If most recent spend moved into a new market, an account-wide average is a poor test of whether the original asset wore out.

Consider this fictional comparison. In both periods, Asset A received 100,000 impressions in each market, and a different Asset B received 100,000 impressions in Market X:

• Market X: Asset A’s clicks fell from 2,000 to 1,200, so click-through rate fell from 2.0% to 1.2%.
• Market Y: Asset A’s clicks stayed at 1,000, so click-through rate stayed at 1.0%.
• Market X: Asset B’s clicks stayed at 1,800, so click-through rate stayed at 1.8% across the same periods.

This is a useful lead: the deterioration appears specific to Asset A in Market X. It does not prove repeated exposure caused it. Check whether delivery within that market changed, and whether enough downstream outcomes remain to judge business performance.

Keep a second possibility open. If Assets A and B both weakened at the same time, inspect what they share before replacing either. A common auction change, destination problem, or product release is a better starting hypothesis than two simultaneous creative failures.

Delivery: Did impressions become more expensive? Did reach contract? Did delivery shift toward a narrower audience? Where frequency is available, rising repeat exposure can support a fatigue hypothesis, but an average does not reveal how every person responded.

Response: Are people watching, clicking, or leaving at different points? Use the same click definition and denominator in both periods. A falling click-through rate is evidence of weaker response in the delivered audience, not a complete explanation of why.

Downstream quality: Among comparable acquired users, did activation and payment change? A lower click-through rate with stable cost per valuable user may be tolerable. A stable click-through rate with collapsing activation deserves a different investigation.

TikTok’s creative guidance recommends refreshing creatives when delivery results show a consistently declining trend and maintaining a supply of creative assets. It does not establish a universal expiry date for every video. Treat that guidance as a prompt to investigate and prepare alternatives. TikTok’s creative best practices

Watch the counts behind the rates. A large percentage swing based on a handful of purchases cannot carry the same weight as a sustained pattern with substantial exposure. Write “insufficient evidence” when that is what you have; it is more useful than a confident fatigue label attached to noise.

Suppose early attention weakens while matched-cohort activation among acquired users remains stable. That suggests testing a different opening while preserving the product demonstration and promise.

If people watch and click but fail to activate, changing the opening seconds may attract more of the same mismatch. Review whether the ad accurately represents what someone can do after installing, including any payment or setup requirement.

TikTok’s Video Insights supports video comparisons and key-frame analysis. Its key-frame metrics are estimates, and view-through conversions are excluded from that analysis. Use the view to locate a question worth testing rather than force it to reconcile exactly with summary reporting. TikTok’s Video Insights guide

Qualitative evidence can sharpen the question. Repeated comments asking whether a demonstrated feature is included may reveal unclear proof or qualification. They do not tell you how common that confusion is across all viewers. Keep the observation and its limitation together.

Choose the smallest meaningful change that addresses your explanation:

• Opening: Make the same relevant problem recognizable sooner. Preserve the offer and demonstration so the test is interpretable.
• Proof: Show the actual task and result more clearly. Use a genuine product demonstration rather than a stronger unsupported claim.
• Promise: Correct the expectation if the creative attracts people the product cannot satisfy. A narrower promise may reduce clicks while improving useful acquisition.
• Next step: Explain what happens after the tap, especially if the first session requires setup or a trial decision.

Some situations justify a new concept rather than a small edit. If the entire premise targets the wrong need, polishing the hook is unlikely to answer the important question. State that explicitly in the brief.

Where feasible, keep a usable control and compare it with the proposed refresh under compatible conditions. Simply adding an ad and comparing its eventual results with last month’s winner is exploratory evidence. TikTok’s split-testing tool separates audiences into mutually exclusive groups and supports creative variables; check whether the available setup fits your question and budget. TikTok’s split-testing overview

Complete six lines before production begins:

1. Suspected problem: Asset A’s opening is losing response in Market X.
2. Supporting evidence: Response weakened there while Asset B held steady.
3. Competing explanation: Asset A may have reached a different audience mix.
4. Next change: Replace the opening; preserve proof, offer, destination, and first-use experience.
5. Success and guardrail: Improve the chosen acquisition outcome without weakening matched-cohort activation quality.
6. Review condition: Evaluate after the agreed exposure and outcome window, within the spending ceiling.

Also write what would weaken your theory. If the refreshed opening lifts clicks but not valuable app actions, the next brief should not celebrate a solved fatigue problem. It should investigate the remaining handoff.

You are trying to learn which message needs to change, for whom, and why. The age of the file cannot answer those questions on its own.