What a Meta account auditactually finds.

Most "free ad account audits" are a sales call wearing a clipboard: someone skims your Ads Manager for ten minutes, declares your account structure outdated, and pivots to the retainer pitch. A real audit is different, and you can tell them apart by one property — a real audit shows its workings. It names each finding, shows the number behind it, and tells you what it can't know from the outside. This note is the full map of what a genuine Meta account teardown looks for, so you can run one yourself, or hold anyone who offers you one to the standard.

Same rules as everything we publish: figures are either Meta's documented definitions (cited below), our own measured data labelled as ours, or arithmetic marked illustrative. No invented benchmarks — anyone quoting you a universal "good CPA" without seeing your margin structure is guessing.

After enough teardowns, the findings stop being surprises and start being categories. Nearly everything a Meta account audit uncovers falls into one of six waste classes:

1. Fatigued spend still running

The single most common finding, and the most mechanical. Meta's own definition: an ad is creative limited when its cost per result is worse than your account's own historical baseline, and fatigued when cost per result reaches roughly twice that baseline — with exposures of the same creative counted across your other campaigns too.^[1] An audit reads every live ad against that line and totals the spend still flowing to ads past it. That total is not an opinion; it's a number, in your currency, per week.

2. Frequency creep

Frequency — impressions divided by reach^[2] — read per ad, per audience, over lifetime. Climbing frequency against a flat audience means the account is paying to show the same people the same idea again. On its own it's a warning light; combined with a drifting cost per result it's a diagnosis. The audit question is never "is frequency high?" but "is frequency rising while response falls?"

3. Creative concentration

Sort spend by creative and look at the top of the table. In many accounts one or two ads carry the overwhelming majority of results. That's not success — it's exposure. When the carrying creative crosses the fatigue line (and it will — see why your ads fatigue), the account has no successor tested and ready, and performance falls off a cliff instead of a slope. An audit measures the concentration and, more importantly, the depth of the bench behind the winner.

4. Self-competition and structure sprawl

Accounts accumulate structure the way desks accumulate paper: duplicated campaigns from old tests, overlapping audiences bidding into the same auctions, and results fragmented across so many ad sets that none exits learning. The audit reads the campaign map as a whole and asks the only question that matters: is every division here deliberate, and is each one getting enough results to be judged? "We're not sure why that campaign exists" is a finding — a common one.

5. Tracking and attribution gaps

Everything upstream is worthless if the measurement lies. The audit checks that the pixel and server events fire on the actions that matter, that the events driving optimisation match the business goal (an account optimising for add-to-cart when the problem is purchases is aiming at the wrong target), and that attribution settings are consistent enough that week-to-week comparisons mean something. Misfiring events are the quietest waste class — the account isn't just spending badly, it's learning badly.

6. The post-click leak

Media buying stops at the click; the money doesn't. An audit follows the winning ad to its landing page and asks: does the page keep the ad's promise, in its words, above the fold, at phone speed? A mismatch here taxes every pound of media upstream of it, which is why a paid-social audit that never opens the landing page is half an audit.

A list of findings without weights is a worry generator. A usable audit scores each class the same way every time — what's the finding, what's the number behind it, how much spend does it touch, and what's the fix's effort-to-impact ratio. Then it sequences: kill fatigued spend first (instant, free), fix tracking second (everything downstream depends on it), simplify structure third, and only then talk about new creative — because new creative poured into a broken measurement layer is money spent learning nothing.

An honest audit also names what it can't see, and you should treat that section as a credibility signal rather than a weakness. From outside the business, no auditor knows your true contribution margin, your repeat-purchase economics, or which products you actually want to scale — which means nobody outside the business can tell you what your target CPA should be, only whether the account is buying results efficiently against the target you set. An audit that hands you conclusions about your margins it was never given the data to reach is manufacturing certainty. The good ones hand you the mechanical findings with the numbers attached, plus a short list of questions only you can answer — and let the two halves meet in the plan.

Waste tells you what to stop. The creative-gap read tells you what to start, and it's the half most audits skip because it's the half that requires craft. Map every live and recent ad by angle (the argument being made), format (video, static, UGC-style, carousel) and hook. Most accounts discover they aren't running ten concepts — they're running two concepts wearing ten outfits. The empty cells in that map are the cheapest tests the account hasn't run, and a good audit hands them to you as a ranked test queue, not a vague "diversify your creative".

This is the discipline we built on our own money before pointing it at anyone else's. Running it on our own brand (TITAN — our own accounts, stated as ours) produced 247 winning creatives for one brand, with click-through moving from 3% to 11% across the winning set and site traffic up +400%. On a client account, the same teardown-then-replace sequence took cost per lead from €22.14 to €4.56 — 4.9× cheaper (−79%) — real client, name withheld under NDA, Meta Ads reporting, account billed in EUR, and past performance, not a promise. Behind the method sits €10M+ of the founders' own Meta spend, self-reported, which is where each waste class on this page was first paid for the hard way.

Everything above is checkable from inside your own Ads Manager, and we've published the self-serve version as a nine-check sequence: audit your own ad account. The wider system it belongs to — what AI changes about creative production this year — is in the free, ungated AI Marketing Hacks 2026 guide.

And if you'd rather have senior operators read the account: the free Crank Audit is this exact teardown, run for you — strategy, gaps, the creative-gap map, free ads, a competitor read and a 30-day plan, delivered in 24 hours. Workings shown, every figure labelled, no call required.