How to set up Google Ads conversion tracking (and verify it works)

Most underperforming Google Ads accounts we review are not suffering from bad keywords or weak copy. They are suffering from a measurement fault. The account is optimising towards a number that does not describe the business: form clicks that never submitted, three overlapping conversion actions counting the same lead, a thank-you page that stopped firing after a website update, or a conversion value that was never passed through at all.

This matters more than it used to. Google's bidding now makes the spending decisions, and it makes them using your conversion data. If you feed it a flattering, inflated signal, it will efficiently buy more of the traffic that produces that flattering, inflated signal. Broken tracking does not just mislead your reporting. It actively redirects your budget. Getting this right is the cheapest performance work available to a small business, because it costs an afternoon and changes every decision afterwards.

What should a small business actually count as a conversion?

Start here, not with the tag. A conversion action should represent something that moves money, or reliably precedes money. Most small businesses should run one or two primary conversion actions and mark everything else secondary, so the headline number stays meaningful and bidding has one clear target.

Action	Count it as	Why
Purchase or completed checkout	Primary, with value	Actual revenue; pass the order value so bidding can optimise profit, not order count
Quote, booking or contact form submitted	Primary	The real lead event for most service businesses
Phone call over a minimum duration	Primary	A 40-second call is a conversation; a 5-second call is a wrong number
Click on a phone number or email link	Secondary	Proves intent to contact, not contact
Newsletter signup, PDF download	Secondary	Useful signal, very different commercial weight
Pricing page view, time on site, add to cart	Secondary or off	Engagement, not commitment; as a primary target it buys browsers

If you sell at meaningfully different values, pass a conversion value rather than counting every conversion as equal. For ecommerce that is the order total. For a service business it can be a sensible average value per lead type, which is enough for bidding to prefer the better enquiries. Guessing roughly right beats treating a $200 job and a $20,000 job as identical events.

How do you install Google Ads conversion tracking?

There are three common ways to get the data in, and the right choice is mostly about how many moving parts you want to maintain.

Method	Best for	Trade-off
Google tag directly on the site	Simple sites, one or two conversion actions	Fewest dependencies, most reliable; code changes need developer or theme access
Google Tag Manager	Several tags, events and marketing platforms	Flexible and auditable, but a container nobody owns is a common failure point
Import conversions from GA4	One shared definition across reporting and ads	Consistent naming, longer chain; an analytics change silently changes your bidding signal

Whichever you choose, the mechanics are the same. The Google tag must load on every page, including the confirmation pages, because a conversion cannot be recorded on a page the tag never reaches. The conversion event must fire on success, not on attempt: the thank-you page load, or a confirmed-submission event from your form tool. And auto-tagging must stay on, because that is what appends the click identifier to your landing page URLs and lets Google match a later conversion back to the ad click that earned it. Lose auto-tagging and attribution collapses even though the tag itself looks healthy.

Two settings deserve a deliberate decision rather than a default. The conversion window defaults to 30 days from the click; a long consideration purchase such as a trade quote or a B2B service often deserves 60 or 90 days, or you will systematically undercount the work your ads did. And counting method: use "one" for lead forms so a customer who submits twice is one lead, and "every" for purchases so a repeat order is counted. Also enable enhanced conversions where you can, as it uses hashed first-party data from your own forms to recover conversions that browser restrictions would otherwise hide. If your site shows a cookie consent banner, make sure consent handling is configured intentionally, because a banner that blocks tags until acceptance will legitimately reduce what you measure.

How do you track phone calls from Google Ads?

For a great many Australian service businesses, the phone is the conversion, which makes call tracking the difference between a useful account and a blind one. You have three distinct mechanisms. Calls from call assets and call-only ads are counted by Google directly, since the call starts inside the ad. Website calls can be attributed by letting Google swap a forwarding number onto your site for ad visitors, which links the call back to the click. Clicks on a phone link are the weakest of the three and should stay secondary, because tapping a number is not the same as having a conversation.

Set a minimum call length that reflects a real enquiry, commonly somewhere between 30 and 60 seconds, so misdials and hang-ups do not train your bidding. If your staff genuinely qualify leads on the phone, consider recording the outcome in your CRM and uploading qualified leads back as an offline conversion, which is the single biggest upgrade available to a lead-generation account.

How do you verify conversion tracking is actually working?

Never trust a tag you have not tested. The verification sequence is short:

1. Submit a real test conversion yourself, end to end, on the live site, on a mobile device as well as desktop.
2. Check the tag fired using Google Tag Assistant or Tag Manager's preview mode, confirming the conversion event appears on the confirmation step and not on the button click.
3. Confirm the status in Google Ads reads as recording conversions rather than inactive, unverified, or no recent conversions.
4. Wait at least 24 hours before concluding anything is broken, then reconcile the count against your inbox, CRM or order list for the same actions.
5. Check for double counting by listing every conversion action and asking which single real-world event each one represents.

Then put it on a recurring review, monthly at minimum. Tracking does not break when you install it; it breaks three months later when the site is redesigned, the form plugin updates, or the thank-you page URL changes. That is also why conversion data belongs on the same dashboard as your actual sales, as covered in how to build a KPI dashboard for a small business, where a divergence between the two is visible within a week rather than a quarter.

Why don't my Google Ads conversions match my actual sales?

Because they are answering different questions, and a modest gap is normal rather than a fault. Google Ads reports a conversion against the date of the ad click, so a sale today can appear in last week's numbers and yesterday's totals will keep rising for days. Your accounting system counts the sale on the day the money moved, from every channel, including the customer who found you organically. Google also only counts what it can attribute, and declined consent, privacy settings and cross-device journeys all reduce that.

So treat platform conversions as a directional optimisation signal and your own books as the source of truth for profit. As a rough working rule, a gap under about 20 to 30 percent is usually attribution and timing. A much larger gap, or ad conversions exceeding real sales, points to a tagging fault: a conversion firing on page load, several actions counting one lead, test traffic being recorded, or spam form submissions. That last one is also why accounts with weak tracking tend to accumulate waste, a pattern we unpack in how to reduce Google Ads wasted spend.

Where to start

If you only have an afternoon: write down the one action that defines a sale for your business, confirm Google Ads is counting exactly that action and nothing else as primary, run one live test, and reconcile a week of conversions against your real enquiries or orders. That sequence finds most faults. Once the signal is trustworthy, the usual benchmarks and targets finally start to mean something, which is where what counts as a good ROAS becomes a usable conversation rather than guesswork, and clean measurement is the foundation under the analytics and growth work in Alpha Vault's services.

The broader point is that measurement is not admin to be done once and forgotten. It is the control system for your ad spend, and like any control system it needs calibration and periodic checking. Businesses that treat it that way compound an advantage over competitors who are bidding confidently on numbers nobody has ever verified.

If your conversion numbers have never quite matched your sales and you would like a second set of eyes on the setup, book a consultation and we will audit it with you.