How to Fix Meta Ads That Suddenly Stopped Working: A Diagnostic Framework
When Meta ads stop working, the instinct is immediate: change something. Refresh the creative. Adjust the targeting. Restructure the campaign. Do something, anything, to stop the bleeding.
That instinct is usually what makes the problem worse.
After 26 years in marketing and working with thousands of businesses on their Meta ads, the most expensive mistake I see isn’t bad creative or wrong targeting….
It’s fixing the wrong problem. Changing the ads when the issue is the funnel. Restructuring campaigns when the issue is budget. Refreshing creative when nothing was actually broken in the first place.
Before you touch anything in your account, you need to know what you’re actually dealing with. That’s what my Follow the Trail Framework is for.
It’s the diagnostic process I run every time a client comes to me with a declining account. It follows the evidence from the top of the funnel to the bottom, in a specific order, until we find where things actually broke down.
Sometimes the problem is the ads. Often it isn’t. And knowing the difference before you start making changes is what separates a fix that works from one that wastes another month of ad spend.
Why Most People Diagnose This Wrong
When someone tells me their Meta ads stopped working, the first thing I do is not look at their ads.
The first thing I do is ask what they mean by “stopped working” and look at the actual numbers.
This sounds obvious. It isn’t. Most people come to a diagnostic conversation with a conclusion already formed. The ads are the problem. The creative is stale. The targeting is off. And because they’ve already decided what’s wrong, they start looking for evidence that confirms it rather than following the evidence to wherever it actually leads.
I’ve learned, after working with thousands of advertisers, never to trust anyone’s self-diagnosis when they bring me a problem. Not because they’re wrong to have an opinion, but because the thing that looks like the problem at first glance is rarely the actual problem. And if you solve for the wrong thing, you’ve spent time, money, and energy and the real issue is still sitting there waiting.
The Follow the Trail Framework exists to prevent that. It forces you to look at the full picture before forming any conclusion, and to rule out every other possible cause before you land on the ads as the source of the problem.
The Follow the Trail Framework
Step 1: Read the Trail
Understand the real situation and the actual numbers before forming any opinion.
The first step isn’t looking at your ads. It’s looking at your data. All of it. The full funnel, the trend lines, the timeline of when performance changed, and the context around that change.
Specifically, I want to know:
- What are the actual numbers, not the interpreted version, the raw data?
- How long has this campaign been running?
- How well was it really performing before, and for how long?
- When exactly did performance change?
- Did anything change around that time, inside or outside the account?
That last question matters more than most people realise. Performance doesn’t usually decline in a vacuum. Something changed. It might be inside the account (a budget adjustment, a creative swap, a structural change) or outside it (a seasonal shift, an economic change, a competitor entering the market, a change in the offer or the landing page). Finding the inflection point on the trend line and asking what happened around that time is often the fastest route to the real answer.
I also look at the full funnel stats at this stage, not just the ad metrics. Landing page conversion rate, cost per lead, lead to call rate, call to close rate. The numbers tell a story and I want to read the whole story before I start forming conclusions about which chapter has the problem.
Step 2: Find the Break
Determine whether there’s actually a problem, and if so, where.
This is the step most people skip entirely, and it’s the one that saves the most money.
Before I start looking for what’s wrong, I ask a more fundamental question: is anything actually wrong?
You’d be surprised how often the answer is no.
A common situation I encounter is a business owner who’s frustrated with their lead volume and convinced their ads have stopped working. We look at the numbers together and the ads are performing well. The cost per lead is good. The ROAS is solid. The creative is converting.
The problem is they’re spending $1,500 a month and expecting results that require $5,000 a month. Nothing is broken. The math just doesn’t support the expectation. When I lay out the numbers and show them what their current ROAS means at a higher spend level, the reaction is almost always relief. They don’t have a broken ad account. They have an opportunity to scale, and scaling will get them exactly the results they want.
That conversation is very different from “your creative is fatigued and here’s how to fix it.” And it requires looking at the real numbers first, rather than assuming there’s a problem because someone feels like there’s a problem.
The other question I ask at this stage is: where in the funnel is the breakdown? Because ads stopping working and a funnel stopping working are not the same thing, and they have completely different solutions.
If your ads are generating clicks and leads at a good cost but those leads aren’t converting to calls or sales, the ads are doing their job. The breakdown is further down the funnel. Maybe the webinar isn’t converting. Maybe the follow-up sequence is weak. Maybe the sales call process needs work. Nine times out of ten, refreshing your creative in that situation does nothing, because the creative is rarely the problem in this scenario.
Following the trail means finding exactly where the numbers drop off before you decide what to fix.
Step 3: Check the Ground
Test the full technical setup end-to-end before assuming it’s a strategy problem.
Once I’ve established that there is a real problem and I have a sense of where in the funnel it lives, I check the technical layer before I touch anything strategic.
This means going through the entire customer journey myself, end-to-end, the way a real prospect would experience it. Click the ad. Land on the page. Fill out the form. Go through the checkout. Complete the signup. Every step, in full, on both desktop and mobile.
You would be amazed at how often something is broken that nobody knew about.
I had a client who was frustrated because their upsell offer wasn’t converting. They’d rewritten the sales copy twice. They were convinced the offer was wrong or the price was too high. We went through the purchase process together and the upsell page wasn’t loading. It had been broken for weeks. Nobody had tested it end-to-end with live pages, so nobody knew. Every person who tried to buy the upsell hit a dead end and left. The offer was fine. The copy was fine. The page was broken.
That’s an extreme example, but the principle applies at every level. Tracking pixels stop firing. Thank you pages break. Email sequences stop triggering. Links in ads go to the wrong place. Checkout processes develop errors that nobody catches because nobody goes through them regularly.
I also check:
- Is the pixel firing correctly and tracking the right events?
- Are the links in the ads going where they should?
- Is the landing page loading quickly on mobile? (Mobile will account for the majority of your traffic.)
- Are the lead notification emails arriving so leads are being followed up?
- Is there actually a follow-up sequence in place or are leads going into the void?
That last one is more common than it should be. Leads come in, nothing happens, and the business owner wonders why their ads aren’t producing clients. The ads produced leads. The absence of a nurture sequence meant those leads went cold. Marketing 101, but it happens all the time.
Rule out every technical and process issue before you move to the next step. Because if something is broken in the infrastructure and you spend a month refreshing creative, you’ve wasted a month.
Step 4: Look at the Ads
Only once everything else is ruled out, examine the creative itself.
If you’ve followed the trail through the first three steps and everything checks out, the ads are where you look. And at this point, you’re looking with a specific set of questions rather than a general sense that something needs to change.
How long have these ads been running?
Creative fatigue is real and it arrives faster than it used to since Andromeda changed how Meta reads and distributes ads. An ad that ran profitably for six months two years ago might fatigue in six weeks today. If your ads have been running for a long time without meaningful creative changes, fatigue is the most likely culprit.
Are they refreshing too frequently?
The opposite problem is just as common. A business owner sees performance dip, panics, pulls the ad after three days, and starts over. The algorithm never gets enough data to optimise properly. Performance stays flat not because the creative is wrong but because it’s never given enough time to find its audience. If you’re constantly restarting, you’re not giving Meta’s system what it needs to work.
Is there genuine creative diversity or just variations of the same thing?
This is where most accounts fall short, especially since Andromeda. I look at the creative pool and ask: are these actually different ads or are they the same ad in slightly different outfits?
A real test means different personas, different visual worlds, different emotional entry points, different proof formats. If everything in the account shares the same basic DNA, the algorithm reads it as one undifferentiated signal and the account plateaus regardless of how much you spend.
Are they only using one format?
Video only. Static only. No carousels. No UGC-style content. No talking-head pieces. Format diversity matters because different people respond to different formats, and Meta’s algorithm needs genuinely different signals to find genuinely different subsets of your ideal audience. If you’ve never tested a format that feels uncomfortable or unfamiliar, that’s usually where the opportunity is.
Is the hook actually working?
The hook is the first two seconds of a video or the first line of copy. If the hook isn’t stopping the right person mid-scroll, nothing else in the ad gets seen. I look at hook rate, the percentage of people who watch past the first few seconds, as a specific metric rather than just overall performance. A low hook rate tells you the problem is at the very beginning of the video ad, before the message even lands.
Are they speaking to one persona or several?
Most underperforming accounts are running one generic ad to everyone. One hook, one angle, one version of the ideal client. Since Andromeda, that approach produces weak signals and weak results. The accounts that perform well have creative built around distinct personas, each with their own hook, their own copy angle, and their own proof format. If your account has one persona running on repeat, that’s where to start.
Is the ad attracting the right person in the first place?
There’s one more scenario worth examining, and it’s easy to miss because it looks like a funnel problem when it’s actually a creative problem.
Earlier I said that if your ads are generating leads at a good cost but those leads aren’t converting to sales, the breakdown is usually further down the funnel. That’s true most of the time. But occasionally the ad itself is the culprit, just not in the way you’d expect.
If your leads are cheap but they never buy, ask this before you touch your webinar or sales page: is the ad speaking to the same person your funnel is built for?
It’s more common than you’d think. An ad uses certain language, calls out certain pain points, speaks to certain desires, and attracts a specific type of person. That person resonates with the ad, clicks through, and enters the funnel. But then the webinar or sales page is speaking to someone slightly different. The language shifts. The problem being solved changes. The person who clicked because the ad felt like it was written for them suddenly isn’t so sure. So they leave.
The ad got the click. The funnel lost the sale. And the temptation is to fix the funnel.
But sometimes the funnel is doing a better job of speaking to the right person than the ad is. The webinar qualifies well. The sales page converts the people who make it through. The problem is the ad is calling in the wrong crowd in the first place.
So before you rewrite your webinar or overhaul your sales page, ask which one is doing a better job of speaking to your actual buyer. If the funnel is strong but leads aren’t converting, look at who the ad is attracting and whether that person is genuinely who you’re trying to serve. Sometimes the fix isn’t the webinar. It’s an ad that was put together quickly without enough thought about exactly who it’s calling out and whether that person is the right fit for what comes next.
The ad is the first conversation you have with a potential client. If that conversation attracts the wrong person, no funnel in the world converts them into the right one.
The Most Common Finding
After running this diagnostic process across hundreds of accounts, the most common finding at Step 4 is some version of the same thing: there’s no real testing happening.
Not because people aren’t trying. But because what looks like testing isn’t actually testing. It’s the same hook on repeat with minor variations. The same template with different text. The same persona with slightly different language. The same visual world across every single ad.
Since Andromeda changed how Meta reads creative signals, that approach produces diminishing returns faster than ever. The algorithm needs genuinely different creative to find genuinely different people. When everything looks the same, it treats everything as the same, and the account plateaus.
The fix isn’t just more ads. It’s more genuinely different thinking before the ads are made.
When to Call in Help
Some of this diagnostic process you can run yourself. The technical checks in Step 3 are straightforward once you know to do them. The funnel math in Step 2 is something any business owner can work through with their own data.
But Steps 1 and 4 are harder to do objectively when it’s your own account. It’s difficult to read your own numbers without confirmation bias. It’s difficult to look at your own creative and see it the way a cold audience sees it. And it’s difficult to know what genuinely different creative looks like when you’ve been staring at the same account for months.
If you’ve run through this framework and you’re still not sure where the problem is, or you can see the problem but aren’t sure how to fix it, that’s exactly what a Strategy Intensive is for. We get on Zoom, go through your account together using this exact framework, and find the real problem, not the one that looks like the problem. It’s a fraction of what another month of misdirected ad spend costs.
If you’re not sure whether a Strategy Intensive is the right next step, start with a free 20-minute Clarity Call. We’ll look at your situation together and figure out what kind of help actually makes sense for where you are.
And if you want to hand the whole thing off, including the ongoing diagnostic work, the creative strategy, and the media buying, done-for-you ad management might be the conversation to have instead.
Read Next
This article is part of a complete Meta ads creative strategy library built specifically for coaches, consultants, and service businesses:
- Meta Ads Creative Strategy in 2026: The Complete Guide for Coaches, Consultants, and Service Businesses
- What Is Meta Ads Creative Strategy? And Why It’s Different From Media Buying
- How Meta’s Andromeda Algorithm Changed Creative Strategy Forever
Julie Lowe is a Meta ads creative strategist and founder of Socially Aligned. She has 26 years of marketing experience and specialises in helping coaches, consultants, and service businesses build Meta ad creative systems that generate quality leads without requiring massive budgets or endless content production.