Last week I audited a Google Ads account that was reporting a 43% conversion rate. The business behind it had received exactly one phone call.
That's not a typo. 300 clicks, 129 reported conversions, one real phone call. On paper, one of the best-performing accounts I've ever opened. In reality, a few thousand dollars spent on a campaign nobody could actually evaluate.
The account belonged to a financial advisory firm, and I walked through it live with a partner. What we found is one of the most common and most damaging problems in Google Ads conversion tracking: the account wasn't underperforming, it was lying. Here's what was actually happening, why inflated conversion data is worse than having no tracking at all, and how to audit your own account in about 15 minutes.
Where the 129 conversions came from
The first thing I check in any account is the conversion actions list. Not the campaigns, not the ads. What is this account counting as a conversion, and which actions actually have volume.
This account had a click-to-call action with 124 recorded conversions. The firm confirmed they had received one phone call. Their best guess at the cause: the tag was firing when visitors hovered over the phone icon on the site. Not clicks. Not calls. Hovers.
The rest were contact form submissions, six in total. At that volume, some of those are usually the company testing its own form.
Back in the campaign view, segmenting the conversions column by conversion action confirmed it: 120 of the 129 conversions were the phantom click-to-call. The 43% conversion rate was fiction.
The label said secondary. The data said otherwise.
Here's the detail that makes this one sneaky. The click-to-call action was set as a secondary conversion at the account level, which is supposed to keep it out of the main conversions column and out of bidding decisions. It was showing up in the campaign's conversion data anyway, flowing through the campaign's account-default conversion goals.
I won't pretend the settings behavior there is intuitive, because it isn't. The lesson is simpler: don't trust the labels, verify the data. Segmenting by conversion action shows you exactly what's flowing into the column Google reports and optimizes against. That view doesn't lie. The labels sometimes do.
Why inflated conversions are worse than no tracking
If you have no conversion tracking, at least you know you're blind, and you act accordingly. Inflated tracking is worse because it manufactures confidence in numbers that mean nothing.
It also sets up what we call the Reverse Optimization Trap. Ad platforms optimize toward whatever conversion data you feed them. Feed Google hover events, and the moment you switch to conversion-based bidding, Google will go find you more people who hover over phone icons. It will get genuinely good at it. The algorithm isn't broken. It's doing exactly what it was told, against a signal that has nothing to do with revenue.
This account was saved from that fate by an accident: it was bidding on Maximize Clicks, so the junk data wasn't steering the bidding yet. The damage was to decision-making instead. The firm thought they had a performance problem, and they had no clean way to see it, because a measurement problem was sitting on top of it. You can't evaluate a campaign, an agency, or a budget when the conversion column is fiction.
How to audit your Google Ads conversion tracking in 15 minutes
This is the exact process I run on every account review. No tools required beyond the Google Ads interface and some honesty.
- Open Goals, then Conversions, and view all conversion actions as a list. Read every action. Note what's primary versus secondary and where the volume sits. Auto-added actions like store visits and get-directions show up here too.
- Segment your campaigns by conversion action. In the campaign view, use Segment, then Conversions, then conversion action. This breaks the conversions column into its parts and shows what's actually firing. This single view is what exposed the 120 phantom calls.
- Sanity-check the numbers against reality. Pull up your CRM, your phone logs, your calendar. If Google says 124 calls and your team remembers one, the audit is over. Reality wins.
- Interrogate how each action fires. A phone tag built on a hover trigger in Google Tag Manager, a thank-you page people can reach directly, a form tag that fires on click instead of submit. Every one of these inflates the count. For calls specifically, Google's built-in call tracking with number swapping beats a homemade tag on a phone icon every time.
- Apply the smell test. A 43% conversion rate on cold search traffic isn't a win, it's a warning. If your numbers look too good, they almost certainly are.
Bad tracking rarely travels alone
The same account had one campaign, one ad group, 242 keywords, and a single ad. The top-spending keywords had Quality Scores of 1 and 3 out of 10, largely because one ad can't possibly be relevant to 242 different searches at once.
And here's the kicker: that one ad showed an Excellent ad strength rating. Dynamic keyword insertion was pulling each search term into a headline, which is enough to satisfy the rating. Ad strength grades the ad in isolation. It says nothing about whether your account structure makes sense. Treating it as a health score is how a 1 out of 10 Quality Score hides behind a green Excellent label.
The fix isn't complicated. Group keywords into tight themes of roughly 5 to 15, write ads that speak directly to each theme, and make sure the landing page continues the conversation the search started. This account's landing page led with a clever brand slogan that had nothing to do with what anyone searched, and the phone number, for a firm that wanted phone calls, was nowhere near the top of the page. Structure and message match matter even more if you ever want to run broader targeting, which only works when the setup isn't broken.
To be fair, this account got real things right. Someone was actively reviewing search terms and excluding irrelevant queries, which I often don't see when I do these audits. And the location settings were correct, targeting people physically in the service area rather than the default that includes anyone merely interested in it. That default is not a small mistake. In another audit, we found an account that had quietly spent over $200,000 across four years on clicks from outside the US, a market the business couldn't even serve.
I graded the account a 5, maybe a 6 out of 10, and told them the honest version: a 6 out of 10 is still a failing grade. Effort inside the wrong structure doesn't compound.
What good conversion tracking actually looks like
Fixing the phantom conversions gets you to zero. It doesn't get you to good.
Good tracking captures the actions that matter, forms and calls, and then adds a layer of quality on top. Was the caller qualified? Did the form fill turn into a real opportunity? That's the data worth feeding back into Google, because that's when the algorithm starts hunting for your buyers instead of your browsers. Your CRM data is the most valuable asset in your ad account, and connecting it through offline conversion tracking is how lead quality becomes the thing Google optimizes for.
Then, and only then, does conversion-based bidding become an advantage instead of a trap. The order matters: clean tracking first, quality signal second, smart bidding last. Most accounts run it in reverse.
One more thing. This is the entire first pass of a paid audit, published for free, on purpose. If this post is enough for you to find and fix the problem yourself, you were never going to hire an agency to audit it anyway, and now you don't have to.
But if you'd rather have experienced eyes on it, get a free system review and we'll tell you what your conversion column is actually made of.
About the author
Kyle Rutledge
Owner
I’m Kyle, founder of Gradari, a paid ads lead generation agency that helps B2B and SaaS companies stop wasting budget on low-quality leads and start building systems that actually drive growth.