Paid performance · Free read-only scan
Request the scan, grant read-only access, and we send back a 28-day breakdown of every campaign, the creatives dragging your CTR down, and the tracking gaps that make the numbers lie. Nothing in your account is changed.
The free-audit market for Meta is crowded, and some of it is good. Adzooma's free audit goes as deep as pixel and conversions-API gaps and creative fatigue at ad-set level, so we are not going to pretend the category is shallow.
Useful, and blind to the thing that has quietly broken underneath half of today's ad accounts. The pixel on its own misses a large share of real events, so every CTR, CPA and ROAS figure is only as trustworthy as the signal beneath it. Optimising on a broken signal is optimising on noise.
So this scan asks the three questions an operator actually asks, in the order that matters. Is my tracking reliable enough to believe any of this? Which campaigns are burning budget with nothing to show for it? And which creatives are pulling my CTR below my own account average, and why? The tracking audit comes first, because the other two answers depend on it.
Event match quality
Missing on low-scoring events: email, phone, external_id.
Creative fatigue, last 7 days
An illustrative example, not your data. Your scan shows your own event match quality scores and your own creatives across the last 28 days.
We read your account through Meta's official read-only reporting permission. That is all we ask for and all we take. We never request write access, we never change a campaign, a budget, a creative or a setting, and we never spend a penny of your budget.
Before any optimisation call, the scan checks the signal: missing UTMs, missing events, spend with no matching sessions, and per-event match quality. If your data is not reliable enough to act on, we say so and we hold the calls rather than dress up a guess.
Every metric is computed once, from the official Meta Marketing API. CTR and CPC use billable link clicks rather than every tap on the ad, and conversion rate is measured against landing-page sessions so inflated click counts cannot flatter a bad campaign.
A creative is flagged when its CTR falls below 80% of your impression-weighted account average, so the bar is your account and not a generic industry figure. Where we compare one creative against another we run a statistical test, and where the volume is too thin to call it, we label it that way.
Your wasted spend, your weakest creatives, and an honest verdict on whether your data can be trusted.
You get a 28-day campaign table with your own account benchmark row, an itemised wasted-spend panel with each line explained, every below-average creative flagged with the trigger that fired and its likely cause, and a per-event signal-quality report ending in a plain verdict: safe to optimise, or fix tracking first.
Each tracked event is scored out of 10 with the missing fields named, so you can see exactly what to fix. A score of 7 or higher is the target. At 4 to 6 you are running on half a signal, and we will tell you so before we tell you what to cut.
Underperformance is measured against 80% of your impression-weighted account average, not an industry benchmark you have never met. Six fatigue triggers with defined thresholds mean a flag tells you which pattern fired, not just that something looks tired.
Significance is not the same as sufficiency. A difference that is real but tiny, or large but under-powered, is labelled as such rather than sold to you as a verdict. Today the creative evaluator scores CTR, and we do not imply otherwise.
The scan checks whether your Meta data can be trusted before it says anything about your campaigns. This is the complete list of what it does today, with nothing on it that is roadmap or managed-service only.
The serious tracking tools have no creative angle. The creative and fatigue tools assume your tracking is fine. Getting both in a single audit usually means paying for it. This scan puts them together: per-event signal quality sits alongside creative-fatigue analysis, read in the same pass, on the same 28 days.
It is read-only and advisory by construction. We do not need write access to produce it, so we do not ask for it, and there is nothing in the process that could change or spend anything in your account. It also shows its working: the thresholds, the rubric and the statistical test are all stated, and when the data is not good enough to call we say so instead of guessing.
And we say what it does not do. You request the scan and we run it for you. It is campaign level, not ad-set or ad level. It reports what your account already contains rather than promising an outcome, and we have no volume claims to make because we would rather show you the method than a number we cannot source.
Most graders check your spend and your creative. This one checks whether the data underneath them can be trusted, and says so before it recommends anything.
| MxD Meta Ads Scan | WordStream (LOCALiQ) Facebook Ads Grader | Adzooma Meta Ads Performance Audit | |
|---|---|---|---|
| Read-only, nothing changed in your account | Yes, read-only only | Yes | Yes for the audit itself |
| Benchmark used | Your own impression-weighted account average | Competitor benchmarks | Not published |
| Creative fatigue flags with the trigger that fired | Yes, six named triggers | No | Yes, ad freshness flags |
| Per-event match quality scored on your tracking | Yes, scored per event | No | Pixel and CAPI presence only |
| Withholds optimisation calls when the signal is unreliable | Yes | No | Not published |
| Statistical test before calling one creative a winner | Yes | Not published | Not published |
| Depth of the read | Campaign level, last 28 days | Account and audience level | Campaign, ad set and ad level |
Swipe the table to see every column.
Competitor capabilities and pricing verified against vendor pages on 22 July 2026. Where a vendor does not publish a behaviour we say so rather than mark it down. Pricing changes often; reverify before relying on it.
Read-only reporting access through Meta's official permission, and nothing else. We never ask for write access, so there is no mechanism by which we could change a campaign, a budget, a creative or a setting, or spend any of your money. You can revoke the access the moment the scan is delivered.
The scan is free, with no card and no obligation. The honest catch is that we are an agency: a scan that finds real problems is how you meet the way we work. If you would rather we ran Meta for you from there, that managed service is what we sell. The scan stands on its own either way.
Because the pixel alone misses a large share of real events, and every CTR, CPA and ROAS figure is only as good as the signal underneath it. We target an event match quality of 7 or higher out of 10. At 4 to 6 you are running on half a signal, and optimising on that is optimising on noise.
The last 28 days, at campaign level, with an account benchmark row for context. Ad-set and ad-level analysis is not part of this scan. It reads live from the Meta Marketing API at the point we run it.
Honestly, on coverage Adzooma's free audit goes deeper in places: it reads down to ad-set and ad level, where this scan reports at campaign level. What this adds is discipline about the data underneath. We score signal quality per event rather than flagging that tracking exists, we benchmark against your own account history rather than a generic average, and we withhold optimisation calls when the data is not reliable enough to support them. If you want ad-level depth, run both.
Nothing you did not ask for. You keep the scan. If you want the same engine run continuously, our managed Meta service does that, and every change it proposes passes a fail-closed guardrail and a human approval step before anything is applied. That is a conversation, not an auto-upgrade.
The free scan tells you where the waste and the broken signal are, once. Our managed paid performance service is the continuous version: it recommends a change, gates it through a fail-closed guardrail with a human approval step, applies it, then judges the result with the same statistics the scan uses, and remembers what did not work so it is not tried again.
