Most Google Ads accounts do not have a traffic problem. They have a measurement problem. The dashboard reports conversions, but nobody can explain which actions fired, whether they fired once, or whether any of them became revenue.
Google Ads conversion tracking only helps when it records real business outcomes. Set it up badly and automated bidding will learn to find more of the wrong behaviour, faster.
Decide what Google Ads conversion tracking should measure
Start with the action that matters to the business: a qualified form submission, a booked meeting, a phone call that lasts long enough to be real, a purchase, or an imported closed sale. Write that definition before opening the setup screen.
Do not create a primary conversion for every click that looks promising. A visit to the contact page, a click on an email address, or a form-start event can be useful diagnostic data, but it is not equal to a completed lead.
Use primary actions for outcomes that should guide bidding. Keep softer actions secondary so you can observe them without telling Google to chase them.
The uncomfortable truth is that inflated conversion counts often survive because they make reports look better. They do not make the campaigns better.
Choose the right source for Google Ads conversion tracking
You can create a website conversion directly in Google Ads or import an event from GA4. Both approaches can work, but they are not interchangeable.
A native Google Ads conversion can provide faster reporting and is the usual choice when the event exists mainly to guide Google Ads bidding. It also supports features such as enhanced conversions when the implementation meets the requirements.
A GA4 import is useful when the same event definition must be shared across channels and analytics reporting. The trade-off is that imported data can arrive later and attribution settings may not match the numbers in other reports.
Do not create both versions as primary actions for the same form submission. That is one lead counted twice. If you need both for comparison, keep one secondary.
Create the conversion action with deliberate settings
In Google Ads, open Goals, choose Conversions, and create a new website conversion action. Enter the site domain and select the category that best represents the result, such as Submit lead form, Book appointment, Contact, or Purchase.
Use a name another person can understand without opening the settings. Website Lead – Contact Form – EN is useful. Conversion 4 is not.
Assign a value when the data supports it. An ecommerce purchase should use its actual transaction value. For leads, a fixed estimated value can be useful only when it reflects the economics of the business.
Choose the count setting carefully. Use One for lead generation because one person reloading a thank-you page should not become five leads. Use Every for purchases when separate transactions genuinely create separate value.
Set a conversion window that matches the buying cycle. A B2B service with a longer consideration period should not be judged with settings chosen for an impulse purchase.
Install the Google tag without duplicating it
Google Ads needs a base Google tag on the site and an event that records the conversion. The base tag may already be installed through Google Tag Manager, a WordPress plugin, GA4, or a previous Google Ads setup.
Check before adding anything. I have audited sites with the same tag hard-coded in the theme, inserted by Elementor, loaded by a plugin, and added again in Tag Manager.
Multiple installations can produce duplicate events, unreliable remarketing audiences, and confusing diagnostics. One clean implementation is better than four overlapping ones.
If you use Google Tag Manager, install the container once across the site and manage the Google tag and conversion event inside it. If a developer controls the site code, provide the exact tag IDs, event conditions, and test case instead of saying “please add tracking.”
Configure the event at the real completion point
The conversion should fire only after the desired action succeeds. For a form, that means a confirmed successful submission, not a click on the Submit button.
Button-click tracking is unreliable because the form may fail validation, the server may return an error, or the user may click several times. A click is an attempt. It is not proof of a lead.
A dedicated thank-you page is easy to test when the page is accessible only after submission and cannot be reached through normal navigation. For dynamic or embedded forms, use a verified success event pushed to the data layer.
For phone calls, decide whether you are measuring calls from ads, calls from the website, or clicks on a phone number. These are different actions and should not be mixed under one vague name.
Add enhanced conversions without ignoring consent
Enhanced conversions can send hashed first-party customer data, such as an email address, to improve matching and measurement. This can recover some attribution that standard browser-based tracking misses.
It does not replace the core conversion event. If the form success trigger is wrong, enhanced conversions will not repair it.
Implement consent collection and Consent Mode according to the markets you serve and the advice of your legal or privacy team. Do not copy a consent banner from another site and assume the technical behaviour is correct.
Test whether tags respect the visitor’s consent choice. A banner that changes its text but leaves every marketing tag firing is decoration, not consent management.
Test Google Ads conversion tracking before launch
Use Tag Assistant and Google Tag Manager preview mode, where applicable, to complete the action as a real user. Confirm that the base tag loads, the correct conversion event fires, and the event fires once.
Test success and failure paths. Submit the form with missing required fields, trigger validation errors, and abandon halfway through. None of those should record a completed lead.
Then complete a valid submission and verify that the lead reaches the CRM or inbox with the expected source data. Tracking that records a conversion while losing the actual lead is not working.
Google Ads may take time to show a new action as active or to display the first recorded conversion. Do not solve normal reporting delay by adding another tag.
Prevent duplicate and inflated conversions
Reload the thank-you page and use the browser back button. If the conversion fires again, add a safeguard such as a unique transaction ID, an event-once condition, or a server-side confirmation tied to the submission.
Review the Conversions column by conversion action, not only as a total. This exposes accounts where soft actions or duplicates are hiding inside a healthy-looking number.
Keep account-default goals under control. A new campaign can inherit conversion actions that belong to a different service, market, or objective.
Campaign-specific goals are useful when campaigns must optimize toward different outcomes. Use them deliberately, not as a patch for an account nobody has cleaned up.
Connect tracking to lead quality and revenue
A form submission is only the first measurement layer. The business still needs to know whether the lead was qualified, contacted, progressed, and sold.
Capture the Google click identifier and useful UTM data where permitted, then pass lead outcomes back from the CRM. Offline conversion imports can teach Google which clicks produced qualified opportunities or sales rather than any person willing to fill a form.
This matters in B2B accounts where lead quality varies sharply. Optimizing toward raw lead volume can send the campaign toward cheap students, job seekers, vendors, and tiny companies that will never buy.
Compare Google Ads data with CRM and finance records on a fixed schedule. The numbers will not always match exactly because attribution and timing differ, but large unexplained gaps require investigation.
Before you launch
- The primary conversion represents a real lead, sale, or other commercial result.
- A successful action fires once, while errors, reloads, and button clicks do not.
- Duplicate Google tags and duplicate primary conversion actions have been removed.
- Consent behaviour, enhanced conversions, and CRM data capture have been tested.
- The campaign is optimizing toward the intended goal, not every account-default action.
If you want someone to check whether your bidding is learning from clean data or expensive fiction, SociaLink’s advertising account audit covers the tracking setup as part of the account review.
