When you take over a Meta Ads account — whether it’s a new client, an internal handover, or you’ve just been given access to something that hasn’t been touched properly in months — the first hour matters more than most people realise. A systematic meta ads audit tells you not just what’s broken, but whether it’s worth fixing or whether you’re rebuilding from scratch.
At Choco Media, we’ve inherited enough Meta accounts to know that the chaos is almost always predictable. The same patterns show up: duplicate audiences, tracking that hasn’t fired properly in months, campaigns that were built for a brief that changed six months ago. The good news is that the audit itself is learnable. This guide walks through the exact process we run, the pixel checks that matter most, and the structural signals that tell you whether you’re optimising or starting over.
This is written for anyone stepping into a Meta account they didn’t build — paid media managers, agency teams, and founders who’ve just hired their first ads person and want to understand what they’re inheriting.
Start with access and hygiene before you look at performance
Before you open a single campaign, get the account structure right. Access issues create invisible constraints that make everything downstream harder to diagnose.
- Check who has admin access and at what level (Business Manager, Ad Account, Pixel). Remove anyone who shouldn’t be there.
- Confirm the pixel is installed via Business Manager, not a personal account. A pixel owned by an individual means you lose it if they leave.
- Verify the ad account currency and time zone match your client’s billing and reporting expectations. Mismatches cause subtle reporting headaches for months.
- Check payment method status. If there are failed charges in the billing history, Meta will throttle or pause delivery — sometimes silently.
This takes fifteen minutes but saves you from diagnosing performance problems that are actually account administration problems.
Pixel and event tracking: the audit step most people skip
Tracking issues are the most common and most damaging thing we find in inherited accounts. If the pixel isn’t firing correctly, every optimisation decision downstream is built on bad data.
Pixel health checks
Open Events Manager and check the activity for each event in the last 30 days. You’re looking for:
- Event match quality score. Anything below 6/10 means you’re losing signal. Low scores usually mean missing hashed parameters — email, phone, first/last name — in the event payload.
- Deduplication status. If you see the same events firing via browser pixel and server-side (Conversions API), confirm deduplication is configured with an event_id. Without it, Meta double-counts and the algorithm optimises against inflated numbers.
- Event volume trends. A Purchase event that drops 80% in a single week and never recovers is almost always a tracking break, not a sales drop. Check the date against any site changes.
In client work we’ve found that roughly half of inherited accounts have a tracking issue significant enough to affect campaign performance — but the previous manager treated it as a campaign problem and kept trying to fix it with budget changes.
Conversions API
If the account doesn’t have Conversions API (CAPI) set up, that’s the highest-leverage addition you can make in week one. Browser-side pixel data has been degrading since iOS 14. CAPI sends events from the server directly to Meta, improving signal quality and reducing the gap between what your analytics shows and what Meta reports. If they’re already using CAPI, check the overlap rate in Events Manager — you want 20-30% overlap to indicate healthy deduplication.
Campaign structure: signals that tell you rebuild vs. optimise
Campaign structure problems are less urgent than tracking problems, but they’re usually the reason the account hasn’t been scaling.
Red flags that suggest a rebuild
- Audience fragmentation. More than 8-10 active ad sets targeting slightly different custom audiences or interest combinations is usually a sign the account was built to “test everything” without a framework. The result is audience overlap, internal auction competition, and no statistical significance on any single test.
- Budget allocation inversions. Spend concentrated in campaigns where no conversions have been recorded, while converting campaigns run at low budgets. This happens when campaigns are optimised for the wrong event (Reach, ThruPlay, Link Click) instead of the outcome you care about.
- Expired creative. Creative frequency above 3.0 is a signal that audiences have saturated. Look for ad sets running the same creative for more than 60 days without a refresh — this is one of the most common reasons accounts plateau.
- Advantage+ Shopping Campaigns misused. ASC works well for e-commerce brands with a healthy pixel. For service businesses or accounts with under ~50 conversions per week, manual campaign structure typically outperforms it. If you see an ASC running on a thin-signal account, it’s worth testing against a manual structure.
Signs the structure is worth keeping
- Clear funnel logic: prospecting campaigns targeting cold audiences, retargeting campaigns targeting site visitors and engagers.
- Consistent naming convention across campaigns, ad sets, and ads — even if imperfect, it suggests intentional management.
- Active creative testing: multiple ads per ad set, new ads added in the last 30 days.
Audience quality and overlap analysis
Open the Audience Overlap tool in Ads Manager (under the Audiences section) and check your active custom and lookalike audiences against each other. Significant overlap between audiences in different ad sets means you’re bidding against yourself in the same auction — this inflates CPMs and fragments learning.
For custom audiences, check the data sources:
- Customer list audiences: when were they last uploaded? A customer list that’s 18 months old has significant decay — people change emails, unsubscribe, or buy elsewhere.
- Website custom audiences: check the window lengths. A 180-day window is significantly broader intent than a 30-day window. If you’re retargeting everyone who visited in 6 months with the same message, you’re almost certainly over-reaching.
- Lookalike audiences: check the seed size and geography. A lookalike built from 50 conversions in a small market is a weak foundation. We typically look for 1,000+ seed events to build a reliable lookalike.
Creative audit: what’s running, what’s learning, what’s dead
This is where most of the actionable insight comes from. Filter your ads by “Active” and sort by spend over the last 30 days.
What to look for
- Spend concentration: Is 80%+ of spend going to one or two ads? That’s not necessarily a problem — Meta’s delivery system should push budget toward winners — but it does mean you’re one creative fatigue away from a performance drop. The question is whether there’s a creative pipeline feeding new tests.
- Zero-spend active ads: Ads marked active but with zero spend in the last 14 days are usually losing the internal auction. They’re cluttering the account and not contributing learning. Archive them.
- Hook rate and hold rate: If the account has video ads, check three-second video views divided by impressions (hook rate) and ThruPlay divided by three-second views (hold rate). Hook rates below 25-30% suggest the opening frame isn’t stopping scroll. Hold rates below 20% suggest the content isn’t earning the watch after the hook.
- Text overlay compliance: Meta removed the 20% text rule formally, but heavy text overlays still underperform in delivery. If you see ads with heavy graphics and text, it’s worth testing against cleaner creative.
Bidding strategy and budget distribution
Check the bidding strategy at campaign level. Most accounts should be running Lowest Cost (automatic bidding) unless there’s a specific reason to use Cost Cap or Bid Cap. We typically see Cost Cap used incorrectly: set too aggressively, it constrains delivery and prevents the algorithm from finding volume. Unless the account has strong conversion volume and a clear cost-per-acquisition target, Lowest Cost is the right default.
For paid media management, budget distribution should broadly reflect funnel logic: more spend in prospecting (which is harder and more expensive) than retargeting, which should convert at lower cost but is limited by audience size.
A common problem: retargeting has a higher budget than prospecting. This means you’re spending more re-engaging a small warm audience than finding new people — the pipeline dries up within weeks.
Attribution settings and reporting windows
Check the attribution window at the campaign level. Meta’s default is 7-day click, 1-day view. This is reasonable for most accounts, but you need to verify it matches how the client measures performance externally.
If they’re using 28-day click attribution in the account but measuring results in a 7-day window in their reporting spreadsheet, they’ll see a persistent gap between what Meta claims and what they can actually verify. This creates trust problems. Get attribution settings aligned with how the client tracks their business before you run any performance reviews.
Also check the account-level attribution model in Settings. If it’s been changed to last-touch or a non-default model, understand why before you change it back — sometimes there’s a reason.
The output: rebuild or optimise?
After a full audit you should be able to answer this question clearly. In our experience:
- Optimise if: tracking is clean or fixable with minor changes, campaign structure has sound logic even if execution is loose, and there are active creatives with meaningful delivery data. Work from what’s there.
- Rebuild if: the pixel has been broken for more than 60 days (the learning data is compromised), campaigns are optimising for wrong events, or the structure is so fragmented that consolidation would take longer than a clean rebuild.
Document your audit findings in a structured format before presenting to the client. We use a simple three-column structure: area, finding, recommended action. It makes the scope of work concrete and prevents scope creep when priorities shift later in the engagement.
For a deeper look at how we approach the work beyond audit, see our conversion rate optimisation process — many of the same diagnostic principles apply across the funnel.
A quick note on what the audit doesn’t solve
A good audit surfaces structural and tracking problems. It doesn’t solve the underlying creative or strategic problems — what offer you’re putting in front of which audience, whether the landing page converts, whether the product-market fit justifies paid spend at this stage. Those are separate questions.
The audit buys you confidence that when you run the next campaign, you’re measuring the right things and the infrastructure can support what you’re about to build.
If you’ve just inherited an account and want a second pair of eyes, get in touch. We run account reviews as part of our paid media work and can usually tell you within 60 minutes whether you’re in optimise or rebuild territory.
Documenting what you find and setting expectations
One thing we’ve learned from doing these audits repeatedly: the document you produce matters as much as the findings themselves. Inherited accounts come with inherited politics — the previous manager might still be in the client’s network, or the client may feel defensive about what they’ve been paying for.
The audit isn’t a blame exercise. Frame it as a baseline: here’s where things stand, here’s what we’ll prioritise, here’s the expected timeline before you see meaningful data. Specifically:
- Tracking fixes typically show impact within 2-4 weeks as Meta re-trains its algorithm with cleaner signal.
- Campaign structure consolidation enters a new learning phase — budget 1-2 weeks of potentially softer performance while the algorithm re-learns.
- Creative refresh impact is usually visible within 7-14 days if the new creative is genuinely different (not just a new thumbnail on the same script).
Setting this timeline upfront prevents the client from interpreting a learning phase as “the new agency made things worse.”
The full account audit process we run in 90 minutes covers some of these same areas in a live-account walkthrough format — worth reading alongside this guide if you want the timing and sequencing we actually use.