The Complete Meta Ads Audit: How to Find What's Actually Broken

Most Meta ads audits are checklists. Verify your pixel. Check your campaign objective. Review your audience overlap. Confirm your attribution window. You work down the list, tick every box, and at the end you have a tidy document and no idea which of those things is responsible for your poor performance. A checklist tells you what to look at. It doesn't tell you where the money is leaking, and those are different questions.

A real audit starts somewhere else. Not with a list of things to check, but with a single diagnostic question: where in the funnel is money actually leaking? Everything else follows from that. After managing north of $100M in Meta spend, I can tell you the accounts that turn around aren't the ones that checked the most boxes — they're the ones where someone found the one broken layer before touching anything, and fixed that. This guide is how you do that: not the Ads Manager tour every other audit gives you, but the five places money actually leaks, in the order you should interrogate them.

Before You Start: The One Question That Determines Everything

Is the problem above the click or below the click?

Above the click is everything that happens before someone taps your ad: the impression, the reach, the hook, the click-through rate. If your problem lives here, it's a traffic-and-creative problem — the ad isn't reaching the right people, or it isn't earning the click from the people it reaches.

Below the click is everything after the tap: the landing page, the offer, the checkout, the conversion. If your problem lives here, the ad did its job and something downstream squandered it.

You have to answer this before you do anything else, because it determines where you spend the next 90% of your audit time — and because the two halves are diagnosed completely differently. The fastest way to answer it: look at your link click-through rate and your landing-page conversion rate side by side. Weak CTR on cold traffic means the problem is above the click. Healthy CTR with weak conversion means it's below. Get this wrong and you'll spend two weeks optimizing creative when the leak was a broken checkout, or rebuilding a landing page when nobody was clicking in the first place.

Why Everyone Audits the Wrong Layer

Creative and CTR are seductive because they're visible, they're fast to change, and they feel like progress. You can ship a new hook this afternoon. You cannot ship a new offer this afternoon — that involves pricing, positioning, and usually a conversation with someone who outranks you.

So advertisers optimize the thing they're allowed to touch. They A/B test thumbnails while the actual problem sits two layers down, untouched, for quarters. The platform encourages this. Meta's interface is built to make you act on the metrics it can move with more of your money — impressions, reach, clicks. It is not built to tell you that your landing page contradicts your ad, because that's not Meta's problem to solve. It's yours.

The result is an entire industry of advertisers and agencies who are world-class at the top layer and structurally blind to the four beneath it. That blindness is expensive: across nearly 450 sales-effect studies, creative drove 49% of sales lift while targeting drove just 11% — and most of that creative-and-offer leverage lives in the layers below CTR. They'll tell you "the creative is fatigued" when the creative is fine and the offer is the fatigue. They genuinely believe it, because the only instrument on their dashboard is pointed at creative.

A proper audit looks at five dimensions, in order, because order is how the money moves. Those five are the hook, the ad copy, the offer, message alignment, and the landing page — and they're exactly the five FunnelBoost scores in every audit.

Layer 1: The Hook

The hook is the first 1–3 seconds of video or the top third of a static. Its only job is to interrupt a scroll and buy you the next three seconds. Nothing else. It does not need to explain your product. It does not need your logo. It needs to stop the thumb.

‍What good looks like: The first frame is already the interesting part — a recognizable problem, a sharp claim, a specific number, a face mid-sentence. What bad looks like: a logo animation, a slow establishing shot, or a person saying "Hey guys" for two seconds. Watch your video ads with the sound off, because most of your audience does. Then pull your 3-second video play rate and your hook rate — 3-second plays divided by impressions. On a healthy account, the best hooks clear the field by a wide margin, and the gap between your best and worst hook is usually larger than the gap between your best and worst full creative. That's how much the opening matters.

A concrete example: a DTC supplement brand I looked at was running a polished 30-second brand film as their top-of-funnel ad. Beautiful production. Hook rate was dismal. We cut the same footage so it opened on a customer mid-sentence saying "I stopped taking it for a week and the bloating came straight back" — same asset, same offer, same page. Hook rate roughly tripled. We changed three seconds.

‍The trap: A weak hook depresses every downstream metric. If nobody watches past second three, your CTR looks broken, your conversion rate looks broken, your CPA looks broken. You'll be tempted to blame all four layers below. Fix the hook first, then re-read the rest, because the rest was never measured on a fair sample.

Layer 2: The Ad Copy

Once the hook has earned attention, the copy carries it to the click. This is the body text, the on-screen captions, the script after the opening line. Its job is to take interest and convert it into enough intent to leave the platform.

‍What good looks like: copy that names a specific problem the reader recognizes in themselves, then earns the click by implying the resolution lives one tap away. What bad looks like: "We're the leading provider of..." — a sentence written for an internal stakeholder, not a buyer. Read your primary text and ask whether the first line works as a standalone line, given that Meta truncates after roughly 125 characters before "...see more." If your hook in the copy is hiding behind a fold nobody opens, it isn't a hook. Then check whether the copy makes a single clear promise or hedges across four benefits. Four benefits is zero benefits — the reader can't hold them, so they hold none. One idea per ad, built around the single most compelling reason this person should act.

‍The trap: Copy problems disguise themselves as targeting problems. When intent is low at the click, the traffic that lands on your page is unqualified — they clicked on a vague tease, not a clear promise. Your conversion rate craters, and the obvious-but-wrong conclusion is "the audience is bad." You go change targeting. The audience was fine. The copy invited the wrong click.

Layer 3: The Offer

The offer is what the prospect actually gets and what they give up to get it: price, terms, risk reversal, bonus, urgency, the shape of the deal. This is the layer almost nobody on the marketing side is allowed to touch, which is exactly why it's where so much budget dies.

‍What to look for: Be honest about whether your offer is genuinely compelling to a stranger who has never heard of you, or merely acceptable to someone who already trusts you. "$200/month, annual contract, book a demo" is an offer for a warm buyer. To cold traffic it's a wall. Look at the friction between the click and the value: free trials that demand a card, demos that require a sales call, lead magnets that ask for ten form fields. Every unit of friction has to be paid for by a stronger hook and stronger copy upstream — and at some point the upstream can't carry the weight.

‍The trap, and this is the big one: A bad offer looks exactly like a creative problem. Picture a great hook, sharp copy, a high CTR — and a CPA that won't come down. Your dashboard says the creative is winning. Clicks are cheap. So you make more of that creative, you scale the budget, and the CPA gets worse, because you're now buying more clicks into the same broken offer. I've seen six-figure monthly budgets pour into "winning" creative that was winning at everything except the part that mattered. The hook was doing its job perfectly. It was delivering qualified, interested people to an offer they looked at and walked away from. No thumbnail test on earth fixes that.

When the upper layers look strong and the money still leaks, the offer is the first place to dig.

Layer 4: Message Alignment

This is the layer that gets skipped in essentially every audit I've ever inherited, and it's the one that does the most quiet damage. Message alignment is the continuity between what the ad said and what the landing page says — the promise, the language, the visual, the offer, all of it matching across the click.

The click is the most fragile moment in the entire funnel. The prospect has just decided to leave the place they were comfortable to go somewhere unknown, on the strength of a promise. If the page they land on doesn't immediately confirm that promise — same headline, same offer, same imagery, same tone — they don't reason through the discrepancy. They feel a flicker of "this isn't what I clicked" and they're gone before it becomes a conscious thought.

‍What to look for: Click your own ads. Actually click them, on a phone, like a customer. This is a message-match check — as Nielsen Norman Group frames it, a link is a promise, and the page has to confirm within a second that the visitor is in the right place — so assume you have the problem until you've proven you don't. Does the landing page headline echo the ad's promise, ideally in the same words? If the ad sold "cut your reporting time in half" and the page leads with "Welcome to the all-in-one platform," you've broken continuity. Does the offer match exactly? An ad promising "30% off" that lands on a page showing full price destroys trust instantly — and "instantly" is not an exaggeration; it happens above the fold, in the first second. Does the creative carry through? If the ad starred a specific person or product shot, that face or that product should be on the page.

‍Why it's the worst leak: Misalignment burns money you've already spent. You paid full freight for the impression, the hook did its work, the copy earned the click, the click cost you real dollars — and then you lose the person at the doorstep over a mismatch that costs nothing to fix. It rarely shows up as a single broken metric. It shows up as a landing page conversion rate that's "just kind of low" across everything, so it never triggers an alarm. It's not dramatic enough to investigate, and that's precisely why it survives audit after audit while the budget keeps draining through it.

Layer 5: The Landing Page

The page is where intent becomes action, or doesn't. Assuming the message aligns, the page still has its own job: load fast, prove the claim, remove friction, and make the next step obvious.

‍What to look for: Mobile load time first, because most of your traffic is mobile and every second of delay sheds conversions — Google found 53% of mobile visitors abandon a page that takes over three seconds to load. Then the hierarchy: is the primary action visible without scrolling, or is your "Buy" button buried under three sections of brand storytelling? Is there proof near the claim — a testimonial, a number, a logo, a demo — within sight of where the promise is made? Count the form fields and cut every one you don't strictly need — form and checkout friction is one of the most documented causes of abandonment. Find the leaks between page load and conversion: a quick scroll-depth and funnel-step read usually shows you exactly which section is where people quit.

‍The trap: A page problem and a message-alignment problem produce the same symptom — low conversion on traffic that arrived fine. The difference is diagnostic discipline. Misalignment loses people in the first second, above the fold, before they engage. A page problem loses them deeper, after they've started to engage and hit friction. If they bounce instantly, look up at Layer 4. If they scroll, read, start a form and abandon, the problem is here on Layer 5. Treating one as the other sends you optimizing the wrong thing for a month.

How the Layers Mask Each Other

The reason single-metric optimization fails is that the layers are a chain, and a chain reports its weakest link as a failure of the whole.

A weak hook makes strong copy, a strong offer, and a great page all look broken, because nothing downstream gets a fair sample. A weak offer makes a strong hook and strong copy look like the wrong kind of winners — cheap clicks, expensive customers. Misalignment makes a great page look mediocre. Each layer can sabotage the readings of every other layer, which is why you can't audit any one of them in isolation, and why reading the dashboard top-down, metric by metric, leads you to the wrong fix nearly every time. You audit in order — hook, copy, offer, alignment, page — because each layer's metrics are only trustworthy once the layer above it is doing its job.

How to Read Your Findings and Fix Them in Order

Finding the leaks is half the job. Fixing them in the right order is the other half, and most people get it wrong by fixing whatever's easiest instead of whatever's costliest.

‍Biggest leak first, by money lost, not by how easy it is to fix. A misaligned landing page wasting 40% of qualified clicks beats a hook tweak that lifts CTR a few points. Size each leak in dollars — spend flowing in, multiplied by the percentage lost at that step — and rank by that. The most expensive leak is the priority even when it's the most annoying to fix.

‍Upstream before downstream when leaks are close in size. Fixing the hook first means every downstream layer finally gets clean data, so your later diagnoses get sharper. Fix an upstream leak before you fully trust your read on the layers below it.

‍One variable at a time, with enough volume to read it. The whole point of an audit is to stop guessing. Don't undo that by changing the offer, the page, and three creatives in the same week and then having no idea which move worked. Change one thing, let it reach significance, read the result, then move.

Most accounts have one or two leaks doing the overwhelming majority of the damage. Find those, fix those, and ignore the long tail of micro-optimizations the dashboard keeps waving at you. The platform will always offer you another button to push. Discipline is knowing which one actually moves money.

The Five Layers in Practice: A Worked Example

Here's the method running on a real-shaped account, because the sequence is easier to trust when you watch it find the leak.

A coffee subscription brand spending in the mid five figures a month came in convinced their creative was tired — they'd been cycling new ads for weeks with no improvement in add-to-cart rate. Walking the five layers in order told a different story. Layer 1, the hook: mostly fine. One ad in the set — "$7 for a terrible coffee. There's a better way." — was genuinely strong, specific and thumb-stopping. Layer 2, the copy: weaker. The primary text led with "The reigning champions of coffee, delivered every month," a brand assertion any competitor could copy word for word, while the actual ownable differentiator — competition-winning, top-1% roasters — was buried in secondary creatives and absent from the headline. Layer 3, the offer: a real 25%-off first-month deal, but hidden in the ad description where most people never read it, with no urgency attached.

Then Layer 4, message alignment — and there was the leak. The ads promised award-winning coffee at 25% off. The landing page opened with "Are you more Gold or Platinum?" — forcing a cold visitor to choose between two subscription tiers they'd never heard of, before any desire had been built, with the 25% promise nowhere above the fold. A visitor who clicked on champion-level coffee at a discount landed on a product-architecture quiz. That's a bait-and-switch in experience terms, even though nothing was technically broken, and it was quietly suppressing add-to-cart on every dollar of paid spend.

The overall Funnel Score came in at 58 — hook 63, copy 55, offer 52, alignment 48, landing page 54. Notice the shape of that: no catastrophic single number, just a set of scores in the 40s–60s, which is exactly what "functioning but leaking" looks like. The prioritized fix wasn't more creative — it was resequencing the landing page to confirm the ad's promise first and build desire before asking the visitor to choose a tier. The creative the team was about to throw out was never the binding constraint. The seam between the ad and the page was.

Where Andromeda Fits Into a 2026 Audit

One thing has genuinely changed since the old audit playbooks were written, and it raises the stakes on the five layers rather than replacing them. Meta's delivery system, Andromeda, now reads the actual content of your creative — the visuals, the hooks, the language — and uses it as a primary signal for who to show your ad to. A companion system, GEM, learns from your conversion patterns and shapes what gets prioritized next. Meta detailed Andromeda publicly in late 2024 and published GEM's technical foundations in November 2025, and the system rolled out globally through 2025.

The practical implication is simple: your creative is now a targeting input, not just a persuasion tool. What your ad says and shows helps decide who sees it. Weak, generic creative doesn't just fail to persuade — it mismatches you to the wrong audience before persuasion even starts. And if your ad and landing page tell different stories, the mismatch doesn't only cost the conversion; it can train the system to deprioritize that creative going forward. That makes the first four layers — hook, copy, offer, and alignment — more consequential than they were in 2022, not less.

A note on honesty, because it matters in this category: Meta has not published Andromeda's weighting formulas or scoring mechanics, and anyone claiming to know them exactly is guessing. What's described here is grounded in Meta's own disclosures and consistent practitioner observation. The audit doesn't reverse-engineer the algorithm — it evaluates the creative and funnel signals the algorithm is known to read.

When to Do It Yourself vs. Pay for an Audit

You can run this audit yourself. Everything above is the actual method, not a teaser. So the honest question is when doing it yourself makes sense and when it doesn't.

Do it yourself if you have the time, the methodical discipline to change one variable at a time, and — the hard part — the objectivity to judge your own account clearly. That last one is where most self-audits fail. You wrote the hook. You commissioned the creative. You approved the landing page headline. You are the single worst-positioned person to notice that your ad and your page make two different promises, precisely because you know they're the same campaign. The blind spot is structural, not a matter of skill.

Pay for an audit when the cost of misdiagnosis dwarfs the cost of the audit — which, on any account spending real money, it does. Three months spent optimizing the wrong layer on a $5,000/month account is $15,000 of spend aimed at the wrong problem. An external, trained eye that has pattern-matched across hundreds of accounts finds the binding constraint faster and without the emotional attachment to the assets you built. FunnelBoost is $99. The case for paying isn't that you couldn't do it — it's that a wrong guess costs far more than a right diagnosis.

What a Good Audit Report Looks Like

Whether you do it yourself or buy it, the output should look the same: a diagnosis you can act on today, not a 40-page PDF you'll skim next month.

A good report is prioritized, not exhaustive. It leads with the two or three things costing you the most, in order, not a comprehensive dump of everything that's less than perfect. It gives you a score you can track — FunnelBoost produces a Funnel Score from 0–100 plus a section score for each of the five layers (hook, ad copy, offer, message alignment, landing page), so you can see at a glance which layer is dragging and by how much. Most functioning accounts land somewhere in the 40s to 60s; the score exists to show headroom, not to shame. And it gives you specific changes, not vague suggestions — "rewrite the hero headline to lead with the competition-win credential and surface the 25%-off promise above the fold," not "improve your landing page." Every finding is tied to the specific funnel you submitted, not a generic best-practices checklist that could apply to anyone.

One scope note worth being precise about, because it's how the audit stays honest: FunnelBoost evaluates what it can see from your submitted ad creative, ad copy, and landing page — the five layers above. It doesn't audit your ad account's tracking setup, pixel health, or audience configuration, because those aren't visible from the outside, and a report that claimed to assess them from a creative and a URL would be guessing. What it does, it does specifically.

FunnelBoost delivers this exact audit for $99, with results in under an hour and 100% ROI on your audit cost or your money back — start your audit at funnelboost.io.

‍

Q: How long does a Meta ads audit take?

A: A thorough self-audit across all five layers takes a few focused hours if you're disciplined about walking the funnel as a stranger would. A FunnelBoost audit is delivered in under an hour — you submit your creative, copy, and landing page URL, and get back a structured report with a Funnel Score and prioritized fixes.

Q: How much does a Meta ads audit cost?

A: Agency audits typically run $1,200–$5,000+, and many are generic checklists rather than a specific diagnosis of your funnel. FunnelBoost is $99, with results in under an hour and a 100% ROI guarantee on the audit cost. The economic case is simple: on any account spending real money, the cost of misdiagnosing the problem for a few months dwarfs the audit fee.

Q: What's the difference between a Meta ads audit and a Meta ads manager?

A: An audit is a one-time diagnosis: it tells you where your funnel is leaking and what to fix first. Ongoing management is execution: someone running, testing, and optimizing your campaigns week to week. An audit is what you do before you decide whether you even need management — and often it reveals that one or two specific fixes, not ongoing management, are what the account actually needs.

Q: How often should you audit your Meta ads?

A: Quarterly is a sensible baseline for a stable account, plus any time performance shifts meaningfully — a CPA climb, a ROAS drop, or after a significant change to your creative, offer, or landing page. Because creative is now a delivery signal under Andromeda, it's also worth re-auditing whenever you refresh your creative library or relaunch a page.

Q: What does a Meta ads Funnel Score mean?

A: The Funnel Score is a 0–100 rating of your funnel's overall health, backed by a separate score for each of the five layers — hook, ad copy, offer, message alignment, and landing page. It's designed to show headroom: most functioning accounts score in the 40s to 60s, and the section scores tell you exactly which layer to fix first rather than leaving you to guess.

Related reading: What Is Meta Andromeda? · Meta Ad Creative That Converts · Message Alignment: The Silent Killer · High CTR, Low Conversions · Why Your Meta Ads Stopped Working · FunnelBoost home