Most Meta accounts that stall don't stall all at once. A winner slows, CPA drifts up a little each week, and someone launches three new ads that look a lot like the old ones. By the time anyone asks whether the creative is the problem, the account has been running on the same handful of ideas for months, and nobody can say which ones are still pulling their weight.
A Facebook ads creative audit answers that properly. It is different from a media audit: it leaves bids, budgets and audiences alone and looks only at what people actually see, how much of it there is, how varied it is and whether the account can learn from it. Below is the checklist in full, scored so you get a number rather than a feeling, followed by how to run it in about an hour.
If you only need to read one report, how to read a Meta creative report covers the cut, keep or scale call ad by ad. The audit sits one level up and asks whether the account is set up to find its next winner at all.
The creative audit checklist (copy it)
24 checks in six sections. Score each one 0, 1 or 2 using the "scores 2 when" column, then add them up. Work at ad level, over the last 30 days, for every ad that spent money in that window.
1. Spend and volume
| # | Check | Where to look | Scores 2 when |
|---|---|---|---|
| 1 | How many live ads actually get spend | Ads tab, sorted by amount spent | Most live ads have spent enough to judge; few sit at near zero |
| 2 | Spend concentration | Share of spend on the top 3 ads | Spend is concentrated on proven ads but new ads still get a real test budget |
| 3 | Ads too small to read | Impressions column | Few ads are stuck under 500 impressions after a week |
| 4 | Freshness | Created date of ads with spend | At least one ad launched in the last 14 days is getting meaningful spend |
2. Diversity of what's live
| # | Check | Where to look | Scores 2 when |
|---|---|---|---|
| 5 | Distinct concepts, not variants | Group ads by the idea they sell | Several genuinely different ideas are live, not one idea in ten edits |
| 6 | Personas | Who is on screen and who the ad speaks to | Two or more distinct buyer types are addressed |
| 7 | Formats | Talking head, demo, static, founder, street interview and so on | Three or more formats are live with spend |
| 8 | Awareness stages | What each ad assumes the viewer already knows | Ads exist for people who don't know the problem yet, as well as for people comparing options |
| 9 | Near duplicates | Thumbnails side by side | No cluster of ads that share the same first frame, creator and opening line |
3. Hooks and holds
| # | Check | Where to look | Scores 2 when |
|---|---|---|---|
| 10 | Hook rate tracked | Custom column: 3-second video plays ÷ impressions | The column exists, and you know which hooks are below your account's norm |
| 11 | Hold rate tracked | Custom column: ThruPlays ÷ 3-second video plays | The column exists, and weak bodies have been flagged separately from weak hooks |
| 12 | Silent first frame | Watch each top ad muted | The first second makes sense with the sound off |
| 13 | Hook variants on winners | Ads sharing a body with different openings | Every proven body has been tested with more than one hook |
4. Fatigue and relevance signals
| # | Check | Where to look | Scores 2 when |
|---|---|---|---|
| 14 | Delivery status | Delivery column, ad and ad set level | No spending ad shows "Creative fatigue", and any "Creative limited" ad has a replacement planned |
| 15 | Frequency on winners | Frequency, week on week | Frequency on the top ads is stable, or a rise is matched by new creative going in |
| 16 | Relevance diagnostics | Quality, engagement rate and conversion rate ranking columns | No high-spend ad sits in the bottom bands on all three |
| 17 | Age of the main winner | Created date of the top-spending ad | The top ad has had iterations launched off it, whatever its age |
5. Message and offer
| # | Check | Where to look | Scores 2 when |
|---|---|---|---|
| 18 | Claims you can back up | Every number and promise in the top 10 ads | Each claim has a source or proof you could show a reviewer |
| 19 | Ad to page match | Click through from each top ad | The offer, price and promise on the page match the ad |
| 20 | Angles beyond the product | What each ad leads with | Ads lead with problems, outcomes and objections, not only features |
6. Naming, tracking and learning
| # | Check | Where to look | Scores 2 when |
|---|---|---|---|
| 21 | Readable ad names | Ad name column | Every name follows one fixed pattern you can split into fields |
| 22 | Results by concept | Your report or sheet | You can total spend and CPA by concept, creator and format, not only by ad |
| 23 | Test log | Wherever the team writes things down | Each test has a written question and a written result |
| 24 | Where new ads go in | Campaign structure | New creative has one clear home where it gets a fair test |
Want it in a doc? Here is the same list as plain text to paste into a sheet, one row per check.
FACEBOOK ADS CREATIVE AUDIT (score each 0 / 1 / 2, last 30 days, ad level) SPEND AND VOLUME 1 Most live ads have spent enough to judge 2 Spend concentrated on winners, new ads still tested 3 Few ads stuck under 500 impressions after a week 4 An ad launched in the last 14 days has real spend DIVERSITY 5 Several distinct concepts live, not one idea re-edited 6 Two or more buyer personas addressed 7 Three or more formats live with spend 8 Ads for unaware and comparing audiences both exist 9 No cluster of near-duplicate ads HOOKS AND HOLDS 10 Hook rate column tracked (3s plays / impressions) 11 Hold rate column tracked (ThruPlays / 3s plays) 12 First second works muted 13 Proven bodies tested with more than one hook FATIGUE AND RELEVANCE 14 No spending ad on "Creative fatigue" 15 Frequency on winners stable or offset by new creative 16 No high-spend ad bottom-ranked on all three diagnostics 17 Top ad has iterations launched off it MESSAGE AND OFFER 18 Every claim in the top 10 ads can be backed up 19 Landing page matches the ad's offer and promise 20 Ads lead with problems, outcomes and objections NAMING AND LEARNING 21 One fixed, parseable naming pattern 22 Spend and CPA available by concept, creator, format 23 Each test has a written question and result 24 New creative has one clear testing home TOTAL ___ / 48
How to score the audit
Add up the 24 scores for a total out of 48. The bands below are our own rule of thumb, not a Meta benchmark, but they map well to what you would do next.
| Total | What it usually means | Next move |
|---|---|---|
| 38 to 48 | The creative engine works. Problems are more likely to be offer, page or budget | Keep the testing rhythm, look outside creative for the drag |
| 24 to 37 | Holes in one or two sections, usually diversity or hooks | Fix the lowest section first, then re-score in 30 days |
| 0 to 23 | The account is running on a few tired ideas and can't tell which ones still work | Fix naming and tracking (section 6) first, then rebuild the pipeline of new concepts |
Look at the section totals as well as the grand total. A score of 30 made up of a perfect section 6 and an empty section 2 is a very different account from a flat 30, and the lowest section is almost always the right place to start.
How do you run a creative audit on a Facebook ad account?
- Set the view. Ads Manager, Ads tab, last 30 days, filtered to ads with spend. Add the three ranking columns, frequency, 3-second video plays and ThruPlays.
- Export it. Pull the table into a sheet so you can sort, group and add the hook and hold rate formulas.
- Watch the top ads. Play every ad that took meaningful spend, muted first and then with sound. Sections 2, 3 and 5 need eyes, not columns.
- Score in one sitting. Go through the 24 checks in order and write one line of evidence next to every 0 or 1.
- Pick one section to fix. The lowest one. Re-run the audit after 30 days to see if the score moved.
The checks people get wrong
A few of these need more than a column to score fairly, and they're the ones where people tend to be kind to themselves.
Fatigue is a status Meta gives you, not a hunch
Meta's help centre spells out how its creative fatigue statuses work. "Creative limited" shows when cost per result is higher than ads you ran in the past but less than twice as much; "Creative fatigue" shows when it is twice as much or more. Meta also counts all recent exposures of the image or video, including from your other campaigns on the same Page. So reusing one winning video across five campaigns does not give you five fresh ads. For check 14, take the status at face value. It is Meta telling you, in its own numbers, that this ad costs more than it used to. The fix itself is covered in how to spot and fix creative fatigue.
Variants are not diversity
Check 5 is the easiest one to score too generously. Ten edits of one testimonial still count as one idea. Meta's own creative differentiation guidance says that when creative is too similar, its system may not recognise it as different and could deliver it to the same audiences, which ends in a limited or fatigue status. That matters more now that retrieval runs through Andromeda, which Meta's engineers describe as narrowing tens of millions of ad candidates down to a few thousand relevant ones. Our guide to creative diversity by persona, angle and format shows how to count concepts properly.
Relevance rankings are a clue, not a verdict
Quality, engagement rate and conversion rate ranking compare your ad with others competing for the same audience. Meta's ad relevance diagnostics page notes two limits worth remembering: they don't appear for ads with fewer than 500 impressions, and they aren't inputs into the ad auction. That's why check 16 only marks an ad down when it is bottom-ranked on all three while taking real spend. One weak ranking on its own tells you very little.
One ad per ad set makes fatigue more likely
Check 24 is about structure, but it shows up as a creative problem. Jon Loomer's breakdown of creative fatigue causes lists a single ad per ad set and narrowed placements among the setups most likely to fatigue. If every new ad goes into its own tiny ad set, you'll score badly on sections 1 and 4 however good the ads are. Our two-campaign testing structure is one way to give new creative a single, fair home.
Naming is the check that makes the rest possible
Checks 21 and 22 look like admin, but without them you cannot score section 2 at all, because you can't group ads by concept or creator. We name our own deliverables with the same fields in the same fixed order every time (script date, studio, brand, creative name, creator, ratio and hook), so any report can be split by any field. You can copy the pattern from our ad creative naming convention template.
What a creative audit can't tell you
It won't tell you whether the offer is right, whether the price is the problem, or whether your tracking is losing conversions. A high score with poor results points you away from creative, and that is useful too. If ads aren't getting delivery at all, start with whether it's the creative or the account setup. If a single winner fell off a cliff, the five-cause performance drop diagnostic is faster.
Where Spark fits
Full disclosure: Spark UGC is our studio, and the gaps in sections 2 and 3 are exactly the work we do: strategy, scripts, creators, editing into hook and ratio variants, and the next round planned from the account data. We have made more than 5,000 ads, and across Spark client accounts blended ROAS is 8.7x. If your score is low in sections 2 and 3, how it works shows how we'd fill those gaps. If it's low only in section 6, you don't need us; you need an afternoon with the naming template.
Frequently asked questions
What is a Facebook ads creative audit?
A Facebook ads creative audit is a review of the ads people actually see in a Meta account: how many get spend, how varied they are, how well their hooks and bodies hold attention, whether Meta flags them as fatigued, and whether results can be grouped by concept. It leaves bids, budgets and audiences to a separate media audit.
How often should you audit your Meta ad creative?
Once a month is a sensible rhythm for most accounts, scored over the last 30 days so each audit covers fresh data. Run one straight away if a top ad shows "Creative fatigue" in the Delivery column or CPA climbs for two weeks running.
What does "Creative limited" mean in Meta Ads Manager?
Meta shows "Creative limited" when an audience has seen the same ad many times and its cost per result is higher than ads you ran in the past, but less than twice as much. At twice as much or more, the status changes to "Creative fatigue".
Why don't I see quality ranking for some ads?
Meta doesn't show ad relevance diagnostics (quality, engagement rate and conversion rate ranking) for ads with fewer than 500 impressions. The rankings also aren't inputs into the ad auction, so treat them as clues for diagnosis rather than scores to chase.
Run it this week
Copy the checklist, block an hour, and score the account before you brief anything new. Your lowest section tells you what the next round of creative should fix. The resources hub has guides for every section, pricing shows how Spark plans work, and you can start a project if you'd like us to fill the gaps.