Google Ads sits on top of a layer most owners never look at.
Underneath your campaigns there is a set of tags on your website, an analytics property collecting behavior, and a tag manager that is supposed to keep it all straight. When that layer is healthy, nobody thinks about it. When it isn't, the symptoms show up somewhere else entirely: numbers that don't match, remarketing audiences that never fill, conversions that stop arriving after a website update nobody told you about.
We run every account through a tag check before we look at a single campaign. Here is what keeps turning up.
Three Tags Firing, Two of Them Pointless
Run a chiropractor's website through Tag Assistant and you find three separate Google tags loading. One of them is a Universal Analytics tag that stopped processing data in 2023 and has been sitting in the page ever since.
This is the single most common finding in this whole layer, and it happens through completely ordinary events. A developer added a tag in 2019. An agency added their own in 2021 because they couldn't get access to the first one. Someone installed a plugin that added a third. Nobody ever removed anything, because removing things feels risky and leaving them feels free.
It isn't free. Every tag is a script the browser has to fetch and execute before your page finishes loading, and page speed is part of how Google scores your landing page experience. You are paying for slower pages to run code that reports to an account nobody has opened in three years.
Worse, duplicate tags can double-count. Two properties both recording the same session produce two different numbers, and then somebody spends an afternoon trying to work out which one is right.
The check: install Tag Assistant, load your site, and look at what fires. If you see more than one Google tag, or anything labeled Universal Analytics, you have cleanup to do. Universal Analytics stopped processing data on 1 July 2023 for standard properties and July 2024 for 360, so any UA tag still on your site in 2026 is pure dead weight.
Installed Is Not the Same as Working
Google Tag Manager gets installed and then abandoned more often than any other tool in this stack.
We audited a psychiatry practice where GTM was loaded on every page and firing nothing. The container was there. The script was there. It contained no working tags. Somebody had installed the plumbing and never connected anything to it.
From outside, an installed GTM container and a working one look identical. The page loads, the script runs, no error appears anywhere. The only way to know is to open the container and look at what is inside, then use preview mode to confirm those tags actually fire on the pages you think they do.
This matters more than it sounds, because GTM is usually where your conversion tags live. An empty or broken container is not a tidiness problem. It is your conversion tracking, silently absent.
The same practice hides a second failure that is harder to spot: tags that fire on the wrong pages. A conversion tag set to fire on every page view rather than only on the thank-you page will report a conversion for every visitor who lands anywhere. The number goes up, everyone is pleased, and Smart Bidding is being trained that a bounce is a win.
Preview mode is the answer to both. It lets you walk your own site while GTM shows you exactly which tags fired on which page, in real time. Load the homepage and confirm the conversion tag stays quiet. Complete a form and confirm it fires once. That five-minute exercise catches the majority of what we find in this layer.
Your Ads Account and Your Analytics Will Disagree. That's Normal.
Sooner or later somebody notices that Google Ads reports 61 conversions and GA4 reports 40, and the natural conclusion is that one of them is broken.
Usually neither is. They count differently, and the differences are documented:
- Attribution. Google Ads credits the ad click. GA4 credits the last non-direct channel by default.
- Timing. Google Ads counts on the date of the click. GA4 counts on the date of the conversion.
- Window. Google Ads uses your conversion window, often 30 to 90 days. GA4 uses its own model, shorter by default.
- Cross-device. Google Ads counts it via signed-in Google data. GA4 counts it differently, and usually lower.
- Repeat actions. Google Ads can count every conversion. GA4 typically counts the key event once per session.
A month where Ads shows more conversions than GA4 is expected, not evidence of a fault. What matters is that the gap stays roughly stable. A gap that suddenly widens or collapses is worth investigating. A gap that simply exists is not.
Knowing this saves you from the worst response to the discrepancy, which is "fixing" one platform until the numbers match. They are not supposed to match.
The Metrics You Are Paying For and Not Importing
Here is the one that costs money quietly.
GA4 can push engagement metrics straight into your Google Ads columns: bounce rate, average session duration, pages per session, engagement rate. Most accounts we audit have GA4 running perfectly and have never turned this on.
Without it, you evaluate keywords on clicks and conversions alone. With it, you can see that a keyword sends plenty of traffic that lands and leaves in four seconds. That keyword is not merely unprofitable. It is actively damaging, because landing page experience feeds quality score, and a high bounce rate on paid traffic tells Google your ad is answering the query badly.
This matters most on broad match, where you have the least control over what you attract. Engagement metrics are how you spot the traffic that is technically relevant and practically worthless, well before the conversion data has enough volume to tell you.
The check: in Google Ads, open the columns editor and look for the Google Analytics section. If those columns are empty or unavailable, your properties are not linked, and it takes about five minutes to fix.
The Audiences You Never Built
The second thing GA4 gives Google Ads is audiences, and this is where most small accounts leave real money behind.
You can build audiences in GA4 from behavior on your site and import them into Google Ads. Not one blunt "people who visited," but genuinely useful segments: visitors who reached a specific product category, people who started a form and abandoned it, visitors who spent more than two minutes on a service page.
For a prefab building company we audited, that meant separate audiences for people who looked at sheds, at accessory dwelling units, and at saunas. Three different buyers with three different price points, and every one of them worth a different message and a different bid.
Almost nobody does this. The audience sits unbuilt, and remarketing either doesn't run or runs against everybody who ever loaded the homepage, which is close to useless.
Two constraints worth knowing before you build them. Audiences need to accumulate before they can be used, so build them early even if you won't use them for a month. And if you are in healthcare, this is one of the areas where the rules are genuinely restrictive, so check before you build anything based on health-related browsing.
It Breaks When the Website Changes, and Nobody Tells You
The recurring failure in this layer is not bad setup. It is good setup that quietly stopped working.
A site gets redesigned. A form plugin updates. A thank-you page URL changes. A developer moves a script during unrelated work. None of that generates a warning inside Google Ads, and the account keeps spending.
When conversions drop for a client, the first question we ask is not about the campaigns. It is: has anything changed on the website? A form changed, a page moved, a plugin updated. That question resolves more sudden performance drops than any amount of bid analysis.
The habit that prevents most of it is dull and takes five minutes:
- Submit your own form. Use a real email, from a phone that isn't in your office.
- Call your own number from a phone that has never called you before.
- Check tomorrow that both appear in the account.
- Do it monthly, and always after any website change.
That is the only test that proves the entire chain works end to end. Everything else confirms that a piece of it looks right.
What Lives Where
The confusion underneath most of this is that three tools overlap and people aren't sure which one owns what.
- Google Ads conversion tracking tells the bidding algorithm what success looks like. Get it wrong and Smart Bidding optimizes toward the wrong outcome.
- GA4 is for understanding behavior on your site and building audiences. Get it wrong and you cannot see engagement quality or run useful remarketing.
- Google Tag Manager deploys and maintains the tags cleanly. Get it wrong and tags go missing, duplicate, or never fire at all.
They are not alternatives, and this is where the reasoning usually goes wrong. "We have GA4, so we're tracking conversions" is the most common version, and it isn't true: GA4 knows what happened on your site, but unless those key events are imported as conversion actions, the bidding algorithm never sees them. The reverse mistake is just as common. An account with immaculate Google Ads conversion tracking and no GA4 link can tell you how many leads arrived and nothing whatsoever about whether the traffic was any good.
The Twenty Minute Version
If you do nothing else:
- Run Tag Assistant on your site. Remove anything Universal Analytics, and any duplicate Google tag.
- Open your GTM container and confirm it actually contains working tags, then use preview mode to watch them fire.
- Link GA4 to Google Ads and import the engagement columns.
- Build two or three genuine GA4 audiences and let them start filling.
- Submit your own form and call your own number, then check tomorrow.
None of that changes a campaign, a bid, or a keyword. It changes whether the information you use to make those decisions is real.
We do this all day long, and if you would rather not spend an afternoon in a tag manager, we can take a look. If you would rather do it yourself, start with Tag Assistant. It takes two minutes and it usually finds something.
