Conversion tracking health check: 15 minutes to know if Google Ads measures properly

Most accounts I audit share the same underlying problem: money is going out, but nobody knows for certain whether the tracking measures the right thing. If it measures wrong, Google optimises in the wrong direction — and you’re paying for an algorithm that buys the wrong clicks.

You don’t need a developer for the first check. In about 15 minutes you’ll know whether your account measures cleanly or has been flying blind for months. Here’s the checklist, step by step.

Step 1: which conversions are you counting at all?

In Google Ads, open the menu at the top left and go to Goals → Conversions (some accounts just call it “Conversions”). You’ll see a list of your conversion actions.

The first question is the important one: what’s actually being counted? Not every “conversion” is a real conversion.

• Real conversions are outcomes: a submitted enquiry form, a phone call, a purchase, a booked appointment.
• Not real conversions are pure activity: a page view, “60 seconds on the page”, scroll depth, a click on “Contact”. Count things like that and Google optimises for busyness instead of revenue — and sends your CPA off in the wrong direction.
• An empty list is just as bad. If nothing is listed, you’re measuring nothing. Then Google bids blind, and so do you.

Only then go through each action:

• Status. Does it say “Recording conversions”? Or “No recent conversions” / “Inactive”? The latter is an alarm bell.
• Last conversion. When was the last one recorded? If that’s days or weeks ago while enquiries keep coming in, something is broken.
• Source. Website, Google Analytics (GA4), Google tag, import? Worth knowing so you don’t end up double-counting later (see step 6).

Step 2: is the right thing counted as “primary”?

The most common expensive mistake. Every conversion action is either primary (counts for the bidding strategy) or secondary (observation only).

• Primary should only be real, valuable outcomes: submitted enquiry forms, qualified phone calls, purchases.
• Not primary: page views, clicks on “Contact”, scroll depth, newsletter sign-ups. Set those to primary and Google optimises for people who click, not people who buy.

Rule of thumb: if your “conversions” number looks suspiciously high — dozens a day on a small budget — you’re probably counting micro-actions instead of real enquiries.

Step 3: are you counting twice?

Two classic double-counting traps:

1. Two tags for the same enquiry — for example the Google Ads tag and a GA4 import, both reporting the same “form submitted” action.
2. Counting set to “every” instead of “one” — the conversion action settings offer “every” and “one”. For leads (one enquiry = one prospect) that almost always belongs on “one”. On “every”, a prospect who submits three times counts as three conversions.

Step 4: are values coming through? (for ROAS)

If you want to bid on value (target ROAS), conversion values have to be passed. If the value column reads “€0.00” or “no value” everywhere, Google can’t tell a €200 job from a €20,000 one. For local service businesses, a fixed average value per lead is often enough to start with.

Step 5: do the conversions actually land in the campaigns?

Back to Campaigns. Add the Conversions column and use Segment → Conversions → Conversion action.

• Are conversions spread across campaigns and search terms — or is it almost all 0? Zeroes across the board while you’re spending mean the tracking isn’t firing, or it fires on a page your ads never reach.
• Go one level deeper to search terms (not keywords): are conversions coming from real, purchase-ready searches — or from “… do it yourself”, “… materials”, “… salary”? That doubles as your first wasted-spend check.

Step 6: does the number match reality?

The most honest test needs no tool at all: compare one real day. Take yesterday. How many genuine enquiries came in — forms in your inbox, calls, messages? And how many conversions does Google report for that same day?

• Google shows considerably more → you’re counting micro-actions or double-counting (steps 2 and 3).
• Google shows considerably fewer → tracking isn’t firing reliably, consent is blocking it, or calls and forms aren’t being captured at all.
• It roughly matches → good sign. That’s how it should look.

Step 7: are calls being measured at all?

For local businesses the most valuable lead often arrives by phone — and that’s exactly the one that most often stays invisible. The classic: someone clicks your ad, reads your number and types it into their phone by hand. To Google that looks like a click with no result — the €6 click that brought you a €5,000 job never gets booked as a conversion.

A clickable tel: link on mobile can still be tracked, more or less. But numbers typed by hand, calls from a landline, or callbacks days later? You only see those with real call tracking — a dynamic phone number assigned to each visitor. I’ve written up why this is such a blind spot, and which tool I love for it, here: Call tracking: do you know which call came from Google?

Step 8: GDPR and Consent Mode

In the EU, tracking may only fire after consent. Without a properly configured Consent Mode v2, one of two things happens: you lose data (tracking blocked), or you track without consent, which is legally dicey. If your cookie banner and tag setup don’t talk to each other, you risk both.

My GDPR tracking check (in German) takes a URL and shows you in seconds whether Consent Mode v2 is set correctly and which tags fire before consent.

The red-flag checklist

Tracking is solid — now what?

Clean tracking is half the job. The other half is whether the numbers add up. Run your real figures through once: my Google Ads calculator (in German) works out CPA, ROAS and break-even in about 60 seconds.

If the check turns up gaps and you’d rather not tinker with it yourself, this is exactly what I do. Write to me at visnakovs@clickspire.de — we’ll go through the account together and I’ll tell you honestly what’s clean and what isn’t, without the technical waffle.