Meta Native Creative Fatigue vs Third-Party Detection

Meta native creative fatigue vs third-party detection stopped being an academic comparison on 6 August 2026. Ads Manager has flagged fatigue natively for a while, which made the third-party layer look optional. Then Meta made the frequency breakdown opt-in, and a lot of fatigue tables quietly went blank without anyone being told.

Why this matters right now

Three Ads Insights breakdowns became per-account opt-in on 6 August 2026: impression_device, hourly stats, and frequency_value. According to Elevarus, the API "still returns HTTP 200. The Ads Manager UI still renders the report. But the rows are blank or zeroed." AdsUploader describes the same behaviour: requests from non-enabled accounts return a success response with an empty data array rather than an error.

The full mechanics of that change, and the other three definitional shifts it compounds with, are covered in the full breakdown of Meta's 2026 attribution changes. The short version for this post: frequency_value is the field most fatigue calls are built on, and a blank frequency table does not read as "no data." It reads as "no fatigue."

That is a false negative in the exact diagnosis you are trying to make. It has been running since 6 August in every account that never opted in.

What Meta's native flag actually is

Meta's native detection is not a score or a trend line. It is two delivery statuses, and they are defined by one variable: cost per result against the account's own history.

Per the Meta Business Help Centre:

• Creative limited — cost per result is higher than your past ads, but less than double.
• Creative fatigue — cost per result is at or above 2× your past ads.

That precision is worth respecting. It is a clean, unambiguous definition, it is free, and it is native to the surface where you make changes. It is also, structurally, a lagging indicator: both statuses fire only after cost per result has already moved. By the time the label appears, you have paid for the fatigue.

Native vs third-party vs pre-diagnosed

| Layer | Signal it uses | When it fires | What you still have to do | |---|---|---|---| | Meta native statuses | Cost per result vs account history | After CPA has already risen (2× for the fatigue label) | Find the flagged ads, decide which matter, decide what to replace them with | | Creative analytics platforms (Motion, Madgicx) | Per-creative trends, frequency curves, predictive scoring | Earlier — leading indicators before CPA confirms | Read the charts, build the interpretation, translate it into a decision | | Good Morning | The same underlying signals, pre-diagnosed | Weekly, on a fixed cadence | Execute the action list. Zero analysis required. |

The category split is not detection quality. Detection is close to commoditised — that argument is laid out in full in do you still need a creative fatigue tool in 2026, and the tool-by-tool version lives in the creative fatigue tool head-to-head and the 2026 detection tool roundup.

The split is what arrives at the end. Native gives you a label. Creative analytics gives you a chart. An actionable Meta Ads dashboard gives you an urgency-tiered action list, already diagnosed.

Four things the native flag cannot see

1. It cannot see fatigue that has not hit cost per result yet. Creative limited requires cost per result to already be above history. Frequency climbing from 2.4 to 3.6 while CTR softens and CPA holds is textbook early fatigue, and it produces no status at all.

2. It cannot rank across accounts. If you manage 12 client accounts, native gives you 12 separate Delivery columns and no view of which of the 40 flagged ads to touch first. Prioritisation is the work, and it is not done for you.

3. It compares against your history, not against the truth. Both statuses benchmark the ad against your own past ads. An account whose baseline CPA has been drifting up for a quarter has a moving benchmark. Ads stop getting flagged because the bar came down to meet them.

4. Since 6 August, it can be running on an empty frequency table. This is the new one. The native statuses do not depend on the frequency_value breakdown, but almost everything you would use to sanity-check them does — your own frequency reporting, your rules, your rotation logic. Elevarus notes that teams running frequency-cap enforcement rules or creative rotation logic tied to reach breakdowns are now operating on empty data without knowing the signal disappeared.

What this looks like in an account

Take an account spending $80K/month with eight active creatives. On 5 August the frequency breakdown shows three creatives above 3.5 on cold audiences. On 6 August the same report shows nothing above 1.0, because the rows came back empty.

Nothing in the interface changes. The report renders. The chart draws. The fatigue watchlist that had three names on it now has zero, and the obvious reading is that the last creative refresh worked.

Two weeks later the native Creative limited status starts appearing on those same three creatives, because cost per result has now moved enough to trip it. The fatigue was real the whole time. The detection layer was blank, and the native layer was waiting for the damage.

That is the specific failure mode worth checking for this week, and it is checkable in about a minute.

Fix the blind spot first

Act today. Open Ads Manager, go to the reporting view, select Breakdown, and enable the toggles in Additional Breakdowns. This is per-account and requires admin access — there is no bulk business-manager option. Agencies with long client lists can use the the insights/feature-settings endpoint endpoint instead, which AdsUploader flags as the practical route at scale.

This week. Do not trust any fatigue conclusion drawn from data between 6 August and the day you opt in. Sources disagree on what happens to that window: AdsUploader reports that asynchronous report jobs return your full history after enabling, while Elevarus states that historical periods before opt-in stay permanently empty. Assume the pessimistic version until your own account proves otherwise, and re-baseline from the opt-in date.

Monitor. Dataslayer lists a fourth affected breakdown — hourly stats aggregated by advertiser time zone, alongside the audience time zone version. If you run dayparting analysis, check that one too.

Common mistakes

• Treating the native flag as a monitoring system. It is an alert on a lagging metric. It tells you fatigue already cost you money. It does not tell you fatigue is coming.
• Reading an empty table as a clean account. Since 6 August this is the single most likely misread in Meta reporting. Blank is not zero.
• Refreshing everything the moment a label appears. Creative limited and Creative fatigue are different severities with different responses. Rotating a Creative limited ad that still clears your CPA ceiling throws away a working asset. The 2026 creative refresh cadence guide covers the trigger thresholds by funnel stage.
• Assuming frequency alone is the diagnosis. Frequency is an input. Frequency thresholds of 2.5 vs 3.5 mean different things on different audience types, and neither number is a decision on its own.
• Paying for a third-party tool to duplicate detection you already have. If the tool's output is another chart, you bought a second opinion, not a decision.

FAQ

Does Meta's native creative fatigue flag replace a third-party tool? For detection alone, largely yes. Meta's two statuses are precise and free. What they do not do is prioritise across ads or accounts, fire early, or tell you what to change. That is where a third-party layer still earns its cost.

What is the difference between Creative limited and Creative fatigue? Cost per result. Creative limited means cost per result is above your historical ads but under 2×. Creative fatigue means it is at or above 2×. Both compare the ad against your own account history, not an external benchmark.

Did Meta remove the frequency breakdown? No. It became opt-in per ad account on 6 August 2026. Accounts that have not enabled it get an HTTP 200 response with no rows, which renders as an empty report rather than an error.

How do I know if my account is affected? Pull a frequency breakdown for a date range you know had impressions. If the rows are blank or zeroed while the summary numbers look normal, you are not opted in.

Should I re-run fatigue analysis for August? Yes, after opting in. Any fatigue call made from frequency_value data between 6 August and your opt-in date was made on an empty table, and backfill behaviour is disputed between sources.

The version that does the thinking for you

Every layer above still hands you an input. Native hands you a label. Creative analytics hands you a chart. Both assume you will sit down, work out which ads matter, and decide what to do.

Good Morning does that part. The creative fatigue tool reads the account, applies the diagnosis, and returns an action list sorted into Act today / This week / Monitor. Action items, not analysis. The same logic runs across the rest of the account in the weekly Meta Ads reporting software, and if you are weighing this against a creative analytics platform, the Good Morning vs Motion comparison covers where each one fits.

Read-only access. $50/month. 14-day free trial.