Wrongly set-up conversion tracking reports a success that doesn't exist and shifts the budget to the wrong place. Seven points to check before trusting the numbers.
The most dangerous number in ad reports is not zero but the wrong number. When you see zero, you know there's a problem; an inflated conversion count, on the other hand, pleases you and leads you to shift the budget to the wrong campaign. You can do the following seven checks in your own account, without needing anyone.
1. When was the last conversion recorded?
In Google Ads, on the Tools → Conversions screen, the last recorded date appears next to each conversion action. If this date is days ago, there are two possibilities: either conversions genuinely aren't coming, or tracking is broken. These are very different problems and both are urgent.
2. Is the same conversion counted twice?
This is the most common cause of inflation. When the code is added both site-wide and to the thank-you page separately, or set up both via a tag manager and via the theme, a single purchase is reported twice.
A simple test: make a test conversion (fill in the form or place a low-value order) and look at how many appear in the report. If you did one action and the report says two, the matter is clear.
3. Is what's being counted really a sale?
If "page view" or "all form submissions" is chosen as the conversion, the numbers come out high but are meaningless. Spam form submissions in particular inflate this number quickly. Make sure the counted event is the action that's genuinely valuable to the business: purchase, quote request, phone call.
4. Are phone customers measured?
In the service sector, a significant portion of customers don't fill in a form; they call directly. If clicks on the site's phone number aren't defined as a conversion, most of the business the ad brings never appears in the report. The campaign seems "not working" and gets shut off.
5. Is the conversion window suitable for your business?
The default tracking window is 30 days in most accounts. In businesses where the decision is made quickly (food ordering, emergency service), this period is longer than needed and can attribute sales to irrelevant clicks. In businesses where the decision takes months (architecture, enterprise software), 30 days is short and hides the real contribution.
6. Is the measurement defined in more than one place?
If the same event is counted both by Google Ads' own tag and imported via Analytics, the total figure can become double the reality. Choose a single source for an event; don't leave both open.
7. How much do ad blockers and cookie consent cut off?
A portion of visitors reject cookie consent or use an ad blocker; these visits don't show in browser-side measurement. This doesn't mean the setup is broken — but it explains part of the difference between platform figures and your own accounting records. Take this into account when comparing figures.
Why don't platform figures match each other?
If Google Ads says 40 sales, Meta says 25 sales, and your total sales are 50, no one is lying. Each platform credits to itself the sale of a user who touched its own ad. The same customer may have both seen the Instagram ad and searched from Google. So platform reports are not added up; your own sales records are the single true reference. Platform figures, meanwhile, are used to compare between campaigns.
How do you add a conversion in Google Ads?
To add a conversion in Google Ads, you create a new conversion action in the Conversions section of the Goals menu, choose its source, configure its settings and place the tag on your site. In the older interface, the same screen sits under the Tools menu.
- Choose the source. Website, app, phone calls or import. Import covers key events from GA4 or offline conversions from your sales team’s records.
- Set the action category. Categories such as purchase, lead form submission and phone call make reports easier to read.
- Assign a value. In e-commerce, the order value is passed dynamically. For service businesses, giving relative values to different types of enquiry helps the system tell valuable leads apart.
- Choose the count setting. The “Every” and “One” options, covered separately below.
- Set the conversion window and attribution model. As explained above, the window should fit your decision time.
- Install the tag. The Google tag is added to the site, and the event is tied through Google Tag Manager to a trigger such as a form submission or a thank-you page.
Once setup is done, review the list of conversion actions. If there are actions that were added over time and are no longer used, or several actions doing the same job, make them secondary or remove them. A cluttered conversion list makes reports harder to read and blurs what the automated bidding strategy is aiming for.
Primary vs secondary conversions, and what “Every” and “One” mean
Primary conversions are the actions Google Ads targets in its bidding strategy and counts in the Conversions column of reports; secondary conversions are for observation only and appear in the All conversions column. Get this split wrong and the automated bidding strategy mistakes a worthless action for a sale and allocates budget accordingly.
For example, if “contact page view” is left as a primary conversion, the system becomes very good at finding people who look at the page and leave, but it doesn’t bring in enquiries. Make intermediate steps like this secondary; keep only actions with real business value as primary.
How do you choose the count setting?
- Every: Every action after a click is counted separately. Suitable for e-commerce, where the same person can place two separate orders.
- One: Counted once no matter how many actions happen after a click. This is the right option for leads such as forms and calls; the same person submitting a form twice doesn’t mean two customers.
These two settings need to be considered together. If a form action is left as primary and “Every”, the same person resubmitting the form looks like two separate successes to the system. Phone calls have a similar setting: by setting a minimum call duration, you can stop short accidental calls from counting as conversions.
How do you test your conversion setup?
The most reliable way to test a conversion setup is to watch the tag fire live with preview tools, then see a real test action land in the platform’s report. Seeing “tag installed” isn’t enough; the tag may be firing on the wrong page or at the wrong moment.
- Google Tag Manager preview mode: With Tag Assistant you can see step by step which tag fires on which event as you browse the site. Submit the form and confirm the conversion tag fires only on a successful submission.
- GA4 DebugView: Shows the events from your visit in debug mode in real time.
- Google Ads conversion status: The status column on the Conversions screen shows whether the tag has sent data recently.
- Meta Events Manager: The Test Events tab lists events from the Pixel and Conversions API live and lets you see whether the same event is arriving from both sources.
During testing, also check that the conversion does not fire on a failed form submission (missing field, error message). This is why setups that count a button click as a conversion get inflated. Note your tests in a short table: date, action tested, which tool it showed up in. Repeating the same test after site updates means you spot a broken tag the same day, not weeks later. You can read about how measurement affects budget decisions in our where is your Google Ads budget going article, and how we approach account setup on our Google Ads management page.
Frequently asked questions
The conversion count dropped but sales are the same. What happened?
Usually there's been a change on the tracking side: the site was updated, the thank-you page address changed, or the cookie consent bar was renewed. First verify the measurement, then question the campaign.
Is server-side measurement necessary?
It's not mandatory. It's meaningful in high-volume accounts where browser-side losses are large enough to affect business decisions. In small accounts, the setup and maintenance cost can outweigh the extra accuracy it provides.
Does a test conversion spoil the report?
A few test records don't distort the statistics meaningfully; but note the day you tested and take that day's data into account when interpreting it.
What are enhanced conversions, and should I turn them on?
Enhanced conversions encrypt details such as the email or phone number entered in a form or order using hashing and send them to Google, helping measure some of the conversions that can’t be matched through cookies. Turning them on is worthwhile for most accounts; however, you need to state in your privacy notice that personal data is processed.
Why is the conversion count in GA4 different from Google Ads?
The two tools can record a conversion against different dates and different channels. Google Ads records the conversion on the date of the ad click, while GA4 records it on the date the event happened; attribution models can differ too. So small differences are normal, while large ones signal double counting or an incomplete setup.
When do you need offline conversion import?
You need it when the sale is closed not on the site but over the phone or in person. When information on whether a form enquiry actually became a sale is fed back from your CRM or sales records into Google Ads, the system tries to find people who become customers, not just people who fill in forms.
