InsightsNovember 14, 2026

Paid Ads Stopped Working? What to Check First

An ordered checklist for when Facebook or Google ads stop performing: tracking, offer, landing page, audience, creative fatigue and auction, before you touch the budget.

Short answer: when paid ads stop working, check in this order: tracking, offer, landing page and checkout, audience structure, creative fatigue, and only then the auction and your budget. Most drops come from the first three. Changing budget or creative before you have ruled those out usually makes the problem more expensive.

The usual reaction to a bad week is to touch the ads: new creative, new audience, a bigger or smaller budget. I understand why, because the ad account is where you see the pain. But it is often the least likely place for the cause. I have watched teams swap creative for months on a campaign whose real problem was a broken checkout.

On a recent project, a DTC apparel brand had 14 months of flat return on ad spend, even after raising spend by 22%. Nobody found a magic creative. We fixed the product feed, split cold, warm and retargeting audiences and cut the checkout from five steps to two, all before touching a single creative. Blended ROAS went from 1.4 to 3.2 in 90 days and CAC fell 38%. Retargeting conversion rate rose 61%. The order of the checks is the lesson, and it is the order below.

Is your tracking telling the truth?

Start here, because every other check depends on it. If conversions are undercounted or double counted, you may be fixing a problem that does not exist, or ignoring one that does.

  • Compare the platform’s reported conversions with orders or leads in your own system for the same days.
  • Check that the pixel or tag fires once per conversion, on mobile and desktop.
  • Look for recent changes: a site update, a new cookie banner, a checkout app. Many sudden drops line up with one of these.

Attribution will never be perfect, and platforms often disagree with each other. What you want is a stable, honest baseline. Our note on marketing attribution covers how to think about the gaps.

Has the offer stopped being attractive?

An offer that worked in spring can be tired by autumn: a competitor discounted, shipping got slower, a price went up. Ask whether you would click your own ad today. If the same offer has run for months, test a different first step, such as a guarantee, a bundle or a lower-commitment entry point, before blaming the platform.

Is the landing page or checkout leaking?

This is where most hidden losses sit. Look at the funnel step by step: ad click, landing page, add to cart or form start, checkout, purchase or submit. Find the step with the steepest fall. In the apparel project, the five-step checkout was costing sales that no ad change could recover.

  • Load the page on a mid-range phone on mobile data. Slow or broken pages waste the click you already paid for.
  • Count the steps and form fields between interest and purchase. Remove everything that is not required.
  • Check that the headline matches what the ad promised.

Is the audience structure muddled?

When cold prospects, past visitors and existing customers sit in one campaign, the platform tends to spend on the people most likely to convert anyway, which flatters the numbers and starves new demand. Separate cold, warm and retargeting so you can see each one’s real cost. Then the average stops hiding a problem in one of them.

Is it creative fatigue?

Only now. Signs of fatigue: rising frequency, falling click rate on a steady audience, and costs climbing slowly over weeks instead of dropping suddenly. A sudden drop usually points to tracking or the site, not fatigue. I go deeper on this in our piece on ad fatigue solutions.

Could it be the auction?

Costs go up around busy seasons, when competitors raise budgets, and when a platform changes how it delivers. You cannot control the auction, so the question is whether your unit economics survive the higher cost. If they do not, the answer is in your margins and conversion rate, not in the ad settings.

What is the order of checks, in one table?

StepWhat to checkTypical sign
1. TrackingPlatform vs your own ordersNumbers disagree, drop starts on one date
2. OfferPrice, shipping, competitor movesClicks fine, few buyers
3. Page and checkoutFunnel step drop-off, mobile speedTraffic steady, conversion rate down
4. AudienceCold, warm, retargeting splitBlended numbers look fine, growth flat
5. CreativeFrequency, click rate trendSlow, steady decline
6. AuctionCost trends, season, competitorsCosts rise across all campaigns at once

I would not change the budget until step three is clean. Cutting spend on a leaking funnel saves a little money, and raising it loses more. The case where all of this played out is on the DTC apparel case page.

What should you do while you diagnose?

Do not freeze the account, and do not start changing five things at once. Lower spend moderately on the weakest campaigns so you stop the loss, and keep the best performer running so you still collect data. Write down the date performance changed and everything that happened around it: site updates, new products, price changes, a new agency or freelancer touching the account, a platform notice.

That list is often the diagnosis. In my experience, a sudden change on a specific date points to something that changed on your side. A slow slide over several weeks points to fatigue, audience saturation or rising competition. Knowing which pattern you have narrows the checks from six to two or three.

When is it time to change the creative?

When the earlier steps are clean and the pattern fits fatigue. Then change the idea, not only the colors. New hooks, a different promise, a different proof point. Swapping a headline on an ad that has run for two months usually buys a few days. Testing a new angle, such as a customer result in place of a product shot, gives the platform something new to learn from.

One rule that has saved clients money: test creative on the audience that already converts, and only then widen it. If a new ad cannot win with warm people, it will not win with cold ones.

FAQ

Should I pause ads when performance drops?

Not automatically. Pause only if you are losing money on every sale. Otherwise lower spend slightly while you diagnose, so you keep learning data.

How long should I wait before changing a campaign?

Give a change enough conversions to read, often a week or more, depending on volume. Changing things every day makes it impossible to tell what worked.

Why did my ads work last month and not now?

The most common reasons are a tracking or site change, a tired offer, or seasonal cost rises. Check those before assuming the platform changed.

Want a second opinion on a campaign that stalled? Get in touch.