10 Step 72 Hour Audit to Fix Facebook Ads That Stopped Converting

Most accounts that suddenly stop converting have broken or misconfigured tracking, not a creative or audience problem. Check your Meta pixel and Conversions API (CAPI) event match quality first, before you touch budgets or ads. Once tracking is confirmed clean, work through optimisation settings, audience structure and the landing page in that order.

TL;DR:

• Most conversion drops are caused by broken or mismatched tracking events, making it crucial to audit the Meta pixel and API before adjusting budgets or creatives.
• Ensuring your campaign objectives align with your business goals and that the optimization event matches actual revenue actions improves ad performance and measurement accuracy.
• Audience fatigue, creative mismatch, or landing page friction often underlie performance issues, but fixing tracking problems takes priority for reliable results.
• Accurate reconciliation of Meta’s reported conversions with CRM revenue and lead data is essential for making informed scaling or cut decisions.
• Compliance with privacy laws, including transparent consent mechanisms, affects tracking validity and should be integrated into the setup.

Table of Contents

• Top reasons your ads stop converting
• Conversion tracking problems: pixel, CAPI and event mismatch
• Campaign objective and optimisation alignment
• Audience and targeting problems
• Creative, offer and message mismatch
• Landing page and funnel leaks
• Budget, bidding, learning phase and auction dynamics
• Measurement, attribution and reconciling Meta conversions with CRM revenue
• Compliance and privacy: what the rules mean for tracking and targeting
• A one-page diagnostic: 10-step audit you can run in 72 hours
• When to cut spend versus when to double down
• How we can help: the Senior Ad Managers audit
• Sources
• FAQ

Top reasons your ads stop converting

When conversions dry up, the cause almost always sits in one of nine places. Work through them in this order, because fixing tracking before anything else stops you from “optimising” a campaign based on broken data.

1. Tracking is broken or incomplete. Pixel and CAPI events are missing, duplicated or mismatched. Fix: audit Events Manager (developer help often needed).
2. Campaign objective doesn’t match the goal. You’re optimising for traffic or engagement, not leads or purchases. Fix: switch to conversions or value optimisation (DIY).
3. The ad set never left the learning phase. Too few conversions mean Meta is still guessing. Fix: consolidate ad sets, raise budget slightly (DIY).
4. Audience is too broad or too narrow. Either Meta can’t find buyers or you’ve exhausted the pool. Fix: rebuild with overlap checks (DIY).
5. Creative fatigue or offer mismatch. Frequency is high, CTR is falling. Fix: rotate creative every 1-2 weeks (DIY).
6. Landing page or form has friction. Slow load, broken fields, confusing checkout. Fix: run a Core Web Vitals check (developer help often needed).
7. Budget or bids are out of step with CPA. Daily spend can’t produce enough data to optimise. Fix: resize budget to target CPA (DIY).
8. Measurement doesn’t match CRM revenue. Platform conversions look fine, sales don’t. Fix: reconcile weekly (DIY, agency helps at scale).
9. Compliance gaps are limiting signal. Missing consent mechanisms affect what Meta can track. Fix: review consent banners against OAIC guidance (legal or privacy advice recommended).

Most businesses jump straight to blaming the audience or the creative. In practice, the first three items on this list account for the bulk of accounts we see that have “just stopped working.”

Conversion tracking problems: pixel, CAPI and event mismatch

If Meta can’t see your conversions accurately, it cannot optimise towards them. That’s the whole mechanism: the algorithm learns from the signals you send it, and a noisy or incomplete signal produces noisy, incomplete delivery. This is why tracking sits at the top of the triage list, not the bottom.

The common failure patterns are predictable:

• Duplicated events, where both the browser pixel and server-side CAPI fire the same purchase without a shared deduplication ID, inflating reported conversions.
• Missing events, typically on a redesigned checkout or a new landing page where the tracking code was never copied across.
• Incorrect event names, like a “Purchase” event firing on an order confirmation page that also loads for cancelled orders.
• Deduplication failures, where the event ID passed from the pixel doesn’t match the one passed from the server, so Meta counts both.

To diagnose this, open Events Manager and check your event match quality score for Purchase or Lead. Run a private browser session and complete a real test purchase or lead submission, then confirm it appears once, with matched parameters, within a few minutes. Cross-check against your server logs if you have access, since a mismatch between what the browser reports and what the server confirms is a strong sign of a dedupe problem.

The OAIC’s tracking pixel guidance also matters here beyond compliance. A poorly configured consent banner that blocks the pixel for a large share of visitors will quietly starve your event volume, which looks identical to a tracking bug from inside Ads Manager.

The practical fix is implementing server-side CAPI with a consistent deduplication ID shared with the browser pixel, then validating key events in Events Manager before resuming spend. If you can’t confirm a clean signal within a day of testing, that’s your cue to pause the campaign rather than let it optimise against bad data.

Pro Tip: Run one test transaction a day for three days straight before you touch budgets. One clean match could be luck; three in a row is a pattern.

If your internal team doesn’t have access to Google Tag Manager or your website’s backend, this is a developer-level fix, not something to leave to the person running the ads. Consider using WarmDoc for sales document tracking and coaching to help validate lead quality through secondary signal sources. Our website tracking work exists largely because this exact gap shows up in most accounts we audit.

Campaign objective and optimisation alignment

Tracking can be perfect and your ads can still underperform if Meta is optimising for the wrong thing. An ad set to “Traffic” will dutifully deliver cheap clicks from people who were never going to buy, because that’s what it was told to find.

1. Check the campaign objective against the business goal. Traffic and engagement objectives optimise for cheap interactions, not sales; conversions and value objectives optimise towards your pixel events.
2. Confirm the optimisation event in the ad set. Open Ads Manager and check whether the ad set is optimising for “Purchase,” a custom conversion, or a lower-value proxy event like “Add to Cart.”
3. Move to value optimisation once you have enough purchase history. This tells Meta to chase higher-value customers, not just any conversion, which matters if your average order values vary widely.
4. Set up custom conversions tied to actual revenue rather than generic events, so the platform’s idea of “success” matches your CRM’s.
5. Test changes in isolation. Duplicate the ad set, change one optimisation setting, and compare real revenue outcomes after a full week rather than switching everything at once.

A catalogue sales objective behaves differently again, since it optimises for product-level performance against your feed rather than a single conversion event. If you’re running eCommerce and still on a generic conversions objective with no catalogue, you’re leaving a tool on the table that’s built for exactly this.

The test here isn’t “did CPA look better,” it’s whether revenue reported in your CRM moved in the same direction as the platform’s reported conversions. If they diverge, the objective is tracking the wrong kind of success.

Audience and targeting problems

Poor audience health shows up in the numbers before it shows up as a feeling. A falling click-through rate paired with a rising frequency usually means you’ve exhausted an audience. A high cost per click with very low conversion volume often means the audience is too broad for Meta to find the right people within it.

• Check frequency first. Anything climbing past 3 to 4 within a short reporting window, alongside a falling CTR, points to audience fatigue rather than creative fatigue alone.
• Run the Audience Overlap tool across your active ad sets to confirm you’re not bidding against yourself for the same people.
• Separate prospecting from retargeting into distinct campaigns, since mixing cold and warm audiences in one ad set muddies the signal Meta uses to learn.
• Seed lookalikes from your highest-value customers, pulled from your CRM, rather than from everyone who’s ever clicked an ad.
• Exclude existing converters and competitor brand searchers from acquisition campaigns, so budget isn’t spent re-selling to people who already bought or were never going to.

First-party data from your CRM is consistently a stronger signal than Meta’s own behavioural targeting once you have a few hundred genuine customers to work from. A lookalike built on real purchasers tends to outperform one built on page visitors or add-to-cart events, because it’s learning from people who actually paid, not people who just looked.

Creative, offer and message mismatch

Creative fatigue is real, but it’s often blamed for problems that are actually tracking or objective issues in disguise. The genuine symptom is a steady CTR decline alongside rising frequency on the same unchanged assets, typically after two to three weeks of consistent delivery to the same audience.

1. Rotate creative every one to two weeks for accounts running meaningful daily spend, sooner if frequency is climbing fast.
2. Test one variable at a time, such as the headline, the thumbnail image, or the call to action, so you know what actually moved the result.
3. Match the ad’s promise to the landing page. An ad offering “20% off your first order” that lands on a generic homepage creates friction the moment someone arrives.
4. Check the call to action makes sense for the objective. “Learn More” on a lead-generation ad is a weaker signal than “Sign Up” or “Get Quote,” which tell Meta what kind of action to optimise towards.
5. Test short-form video alongside static images if you haven’t already, since the format mix can shift performance independently of the message itself.

The quickest wins to test in the next week or two are usually the cheapest to produce: a new primary text angle, a genuine customer-style photo instead of a stock image, and a headline that states the offer plainly instead of hinting at it.

Pro Tip: If you change creative and the CPA doesn’t move within about a week of stable delivery, the problem probably isn’t the creative.

Landing page and funnel leaks

An ad can do everything right and still fail to convert if the page it sends people to loses them. This is the part of the funnel most Meta advertisers check last, which is backwards, since it’s often the cheapest thing to fix.

• Pull your analytics funnel report to see exactly where visitors drop off, whether that’s on page load, mid-form, or at payment.
• Check Core Web Vitals and mobile rendering, since a page that loads slowly or breaks on mobile will lose a meaningful share of Meta traffic, which skews heavily towards mobile devices.
• Look for JavaScript errors and broken UTM handling, which can silently prevent your tracking from firing even when the pixel code itself is correct.
• Reduce form fields to the minimum needed, enable autofill, and add a visible trust signal like a review count or a security badge near the submit button.
• Simplify the checkout flow, removing unnecessary account-creation steps or extra pages that add friction without adding value.

Once you’ve made changes, validate them properly rather than assuming they worked. A short A/B test, or even a few days of live comparison against the previous version, tells you whether the fix moved the needle or whether you’ve just changed the page for its own sake. Our website optimisation work is built around this exact reconciliation step, because a page that “feels better” and a page that converts better aren’t always the same thing.

Budget, bidding, learning phase and auction dynamics

Meta’s delivery system needs stable inputs to learn, and frequent edits reset that process more often than advertisers realise. Changing budget, audience or creative inside an ad set restarts the learning phase, which typically needs around 50 conversions within a seven-day window to exit cleanly.

1. Set a realistic budget relative to your target CPA. If your target cost per acquisition is $80 and you’re spending $20 a day, you won’t generate enough conversions to exit learning in a reasonable window.
2. Avoid editing live ad sets more than once every three to four days unless something is clearly broken, since constant small tweaks keep resetting the clock.
3. Hold off changing bid strategy for the first three to six months of clean, reconciled data. Lowest cost bidding is the right default while you’re still gathering volume; cost cap or bid cap only make sense once you have enough reconciled CPA history to set a sensible ceiling.
4. Watch CPM and CTR together as auction signals. A rising CPM with a steady CTR usually means increased competition in the auction, not a problem with your account.

Seasonal shifts compound this. Retail-heavy periods see CPMs climb across the board as more advertisers compete for the same inventory, which can look like your account “broke” when it’s actually the market getting more expensive. Separating a genuine account problem from a market-wide auction shift is part of what a proper diagnostic should rule out before you start pulling levers.

Measurement, attribution and reconciling Meta conversions with CRM revenue

Platform-reported conversions and CRM revenue tell different stories more often than most advertisers expect, and the gap is where wasted spend hides. Meta will happily report a conversion for an event that never closed as a sale, or miss one that did close through an assisted path it couldn’t attribute cleanly.

The reconciliation sequence we run looks like this:

• Export Meta’s reported conversions for the period, including the attribution window and event names used.
• Match each one against your CRM by order ID, email or phone number, not just by rough timing.
• Remove duplicates, which commonly appear when both pixel and CAPI fire the same event without a shared dedupe ID.
• Calculate revenue per matched conversion, so every ad set has a reconciled ROAS, not just a platform-reported one.

Once that’s done, prioritise three numbers: closed revenue from matched conversions, the matched lead rate (how many platform-reported leads actually show up in your CRM as real contacts), and cost per real acquisition, calculated against the matched number rather than the platform count.

You’re scaling against a number that doesn’t reflect real demand.*

Items that typically surface immediately in this kind of audit include duplicate purchase events inflating reported ROAS, leads attributed to the wrong campaign due to a broken UTM parameter, and conversions counted twice because a lead filled out the same form on two different sessions. A $99 AUD Google Ads audit from Senior Ad Managers follows the same logic for Meta accounts: check tracking, reconcile the numbers to your CRM, and hand back a short list of what’s broken and what to fix first, before anyone talks about scaling spend.

Compliance and privacy: what the rules mean for tracking and targeting

Privacy obligations aren’t a side issue for Meta advertisers. They directly affect how much reliable data you can collect in the first place, and getting them wrong carries real enforcement risk.

• APP 7 applies to tracking pixels used for targeted advertising, and the OAIC’s guidance is clear that businesses need a simple, accessible opt-out for individuals who don’t want to be targeted this way.
• Sensitive data needs explicit consent. OAIC determinations involving health providers found that using tracking pixels to collect sensitive health information for targeting without consent breached privacy obligations, which matters directly if you’re in health, finance or any category touching sensitive personal information.
• Commercial electronic messages need consent and a working unsubscribe. ACMA’s guidance on spam requires accurate sender identification and a functional unsubscribe mechanism, which applies to the lead-nurture sequences many Meta lead-gen campaigns feed into.
• Build consent into the UI, not just the policy. A clear consent banner that actually blocks the pixel until a visitor opts in protects you legally and still allows tracking for those who consent.
• Avoid collecting anything sensitive through ad tracking unless you have a specific, documented consent process for it.

None of this needs to gut your tracking setup. It needs a consent mechanism that’s honest about what it’s doing, and a process for honouring opt-outs when they come in. If your business handles health, financial or other sensitive categories of data, or you’re unsure where a campaign sits against APP 7, that’s a conversation for a privacy professional, not a guess based on what a competitor seems to be doing.

A one-page diagnostic: 10-step audit you can run in 72 hours

This is the order we’d work through an account that’s stopped converting, fastest checks first.

1. Confirm the pixel fires on every key page using Meta’s Pixel Helper browser extension; pass means the event appears live.
2. Check event match quality in Events Manager; a score above 6 out of 10 is generally workable.
3. Run a real test purchase or lead in a private browser session and confirm it appears within minutes.
4. Cross-check CAPI and browser events for duplicates using the event ID field; pass means one event, not two.
5. Review the ad set’s optimisation event in Ads Manager, confirming it matches your actual business goal.
6. Check frequency and CTR trend over the last 14 days for signs of audience or creative fatigue.
7. Run the Audience Overlap tool across active ad sets to rule out internal competition.
8. Export the last 30 days of platform conversions and match them to your CRM, calculating a matched rate.
9. Check Core Web Vitals and mobile load speed on your primary landing page.
10. Review consent banner behaviour against current pixel firing to confirm you’re not blocking or over-collecting data.

If steps 1 to 4 fail, pause spend on the affected campaign until tracking is confirmed clean rather than letting it keep optimising on bad data. If steps 5 to 7 flag issues, that’s usually fixable within the account directly. If step 8 shows a matched rate well under what you’d expect, or steps 9 and 10 turn up technical or legal gaps, that’s the point to bring in a developer or an agency rather than keep guessing.

When to cut spend versus when to double down

Once tracking is reconciled, the decision about whether to cut or scale a campaign comes down to one number: cost per real acquisition, calculated against your CRM, not the platform’s reported cost per result. If reconciled CPA sits comfortably under what a customer is worth to you, and that’s held for at least a week of stable delivery, that’s the signal to add budget. If reconciled CPA is above that line even though the platform shows a “good” result, the honest move is to cut, not to keep feeding a campaign that looks fine from inside Ads Manager and loses money everywhere else.

This is the single biggest shift in thinking most business owners need to make. Platform-only metrics reward activity, not revenue, and a campaign can look efficient on cost per lead while quietly producing leads that never close. If you don’t have an in-house developer to implement server-side tracking or reconcile exports against your CRM, that’s not a reason to keep making decisions on platform numbers alone. It’s a reason to get the reconciliation done properly once, so every decision after that is based on something real.

— Jesse Lees

How we can help: the Senior Ad Managers audit

If you’ve worked through the diagnostic above and found gaps you can’t fix in-house, that’s exactly what our $99 AUD Google Ads audit was built for. It checks your tracking setup end to end, reconciles a sample of platform conversions against your CRM, and comes back with the top three fixes we’d make first, in plain language, before any conversation about an ongoing engagement.

Experienced specialists run the audit, not a junior learning on your account. We’ll tell you upfront if we don’t think we can help. There’s no 12-month lock-in attached to the audit itself, and if tracking turns out to be the core issue, we can quote separately for the website tracking work or ongoing Meta Ads management depending on what the account needs.

Start with the $99 AUD audit and get a clear answer on what’s actually broken before you spend another dollar guessing.

This article is general information, not a substitute for advice from a qualified lawyer. Consult a qualified legal professional about your own circumstances before acting on anything here.

Sources

• Tracking pixels and privacy obligations — OAIC
• Privacy Commissioner finds privacy breaches in third-party tracking pixel investigation — OAIC
• Dealing with spam — ACMA

FAQ

Why are my ads getting clicks but not converting?

Clicks without conversions usually point to a landing page or tracking problem rather than an ad problem, since the ad has already done its job of generating interest. Check that your pixel is firing correctly on the confirmation page first, then look at page load speed, mobile rendering and form friction before assuming the audience or creative is wrong.

Do Facebook ads still work in 2026?

Meta Ads remain a viable channel for businesses with clean tracking and a clear objective set correctly in Ads Manager. The accounts that struggle are typically the ones running on broken tracking or mismatched optimisation events, not the platform itself.

Why are my Facebook ads suddenly not working?

A sudden drop in performance is most often caused by a tracking break, such as a website update that removed or altered pixel code, or a shift in the auction from increased seasonal competition. Check Events Manager for a drop in matched events first, then compare your CPM trend against the same period last year before assuming a deeper problem.

Is there something wrong with Facebook ads today?

There’s no single platform-wide issue behind most conversion drops; the cause is almost always specific to the individual account’s tracking, objective setup or audience structure. Running through a structured diagnostic, starting with Events Manager, usually identifies the exact point of failure within a day or two.

Recommended

• Social Media Advertising — Meta, Instagram, TikTok
• Website & Conversion Rate Optimisation