In shortThe conversion count in Google Ads is often wrong. Seven common causes, from button clicks to broken tags, and how a business owner can fix each one.
Your Google Ads conversions are usually wrong because the account counts something that is not an enquiry: a page view, a button click, a phone tap, a bot's form fill or the same lead twice. Sometimes it counts too little, because a consent banner or a site change has stopped the tag firing. Each cause has a practical fix.
This matters more than it looks. The conversion number is not just a report for you. If your campaigns use automated bidding, it is the signal Google uses to decide which searches and people to pay for. Count junk as a lead and the system will go looking for more junk, and it will look like it is working. That is why checking conversions is the first thing I do in any account I manage under Google Ads management, before touching keywords or budgets.
Key Takeaways
- Count a lead only when your website has accepted the enquiry, not when a button is clicked or a page loads.
- Count calls over a minimum length, not taps on a phone number.
- Remove duplicate tags and count leads as "One" per ad click.
- Test a real enquiry after every website change.
- Feed won customers back to Google Ads and compare it with your real sales monthly.
Is Google Ads counting page views or button clicks as leads?
This is one of the most common faults in accounts I inherit. The conversion action is set to fire when someone loads the contact page, or when they click "Send" on the form. Neither is an enquiry.
A contact page view counts people who looked and left. A button click counts every attempt, including ones where the form failed or the visitor gave up at a required field. The account shows a healthy cost per lead while the inbox stays quiet. Thank-you pages cause a quieter version: anyone who refreshes or bookmarks one is counted again.
The fix: count only server-accepted enquiries. The conversion fires after your website's server has received the form, passed its checks and saved or sent it, and not a moment before. The confirmation message appears only once the submission has succeeded, and the tag fires from that confirmation. A web developer can set this up quickly. Google's conversion measurement overview explains the different ways a conversion can be recorded.
Could the same enquiry be counted twice?
Yes, and it happens in two common ways.
The first is duplicate tags. A plugin adds the Google tag, then someone adds it again through Google Tag Manager, and a single form fill sends two conversions. A similar double count happens when a Google Analytics key event is imported into Google Ads and a separate Google Ads conversion tag tracks the same form, with both set as primary actions. Google Ads adds them together.
The second is the counting setting. Every conversion action has a choice between counting "Every" conversion or "One" per ad click. "Every" suits online shops, where each purchase is a separate sale. For leads it inflates the number: if someone sends the form twice because they were not sure it went through, that is one enquiry, not two.
The fix: in the Goals section of Google Ads, look at which actions are marked primary. Keep one primary action per real outcome, set lead actions to count "One", and demote everything else to secondary so it is still visible but not used for bidding. Then send a single test enquiry and confirm the account records exactly one conversion.
Are spam and bot form fills being counted as leads?
If your form fires a conversion on submission, automated spam counts as a lead. So does the person selling search engine services through your quote form. Google has no way of knowing the difference unless you tell it.
Worse, automated bidding learns from what it is told succeeded, so an account fed on spam may drift towards the traffic that produces more of it.
The fix: the same server-accepted approach, with spam filtering built into the acceptance step. A hidden honeypot field, rate limiting and sensible validation stop most automated submissions. Only enquiries that pass trigger the conversion. Keep a log of rejections so you can check real people are not being blocked.
Is a phone tap being counted as a phone call?
Tracking "clicks on your phone number" records the tap, not the call. Someone can tap and hang up before it rings, or tap by accident while scrolling. Both count as conversions, and for phone-led businesses they can swamp the real total.
The fix: use Google forwarding numbers on your ads and website so Google Ads records actual calls, then set a minimum call length so only real conversations count. For Procon 24/7, a concrete supplier whose account I run, only calls of 60 seconds or more are counted. A brief ring is not an enquiry. Choose a threshold that fits how your own calls go: long enough to exclude wrong numbers, short enough not to miss a quick booking.
Could consent settings be hiding real conversions?
Yes. In the UK, advertising and analytics cookies generally need a visitor's consent, and your cookie banner decides whether the Google tag runs. If the banner blocks the tag and a visitor declines, their enquiry never reaches Google Ads. The account under-reports, and bidding learns from a smaller, less representative group.
Consent mode is Google's way of handling this. The tag reads the visitor's choice and adjusts what it does, and Google can model some conversions it cannot observe. The common gaps are a banner not connected to consent mode, so the tag never loads or loads for everyone, and a tag that fires before the visitor has chosen. The second is a compliance problem, not just a counting one.
The fix: check that your banner and consent mode are connected, and test both choices (accept and decline) to see what happens. Do not fix under-counting by ignoring consent. The ICO's UK GDPR guidance is the place to start on the privacy side.
Did a website change break your tracking?
This is the one that catches careful people out, because nothing looks broken. The site loads, forms send, the phone rings. The tag has simply stopped firing, and Google Ads reports a quiet week.
I know this from one client's experience. During the 2026 rebuild of Procon 24/7's website, a small scoping problem in the new code trapped the Google tag, and every Google Ads conversion silently stopped firing until it was caught. The account's numbers stayed trustworthy while that bug was live only because results are reported from Procon's own revenue reports, not from the ad platform. That choice was made early, and the dashboard later proved to be under-reporting on that account anyway: wrong in the opposite direction to most.
A second lesson I own from the same account: call tracking should have been in place from month one, not month four. Until it was, spend decisions were made without calls in the picture.
Measured from those revenue reports, Procon's account shows £1,337,693 of tracked revenue on £150,443 of ad spend from 2022 to September 2026, an 8.9x return, and 2,055 new customers. That is one client's result, not a forecast. The full method, including the measurement changes made in September 2026, is in the Procon 24/7 Google Ads case study.
The fix: after every website change, however small, send a real test enquiry and make a test call, then confirm each one appears in Google Ads. Google's Tag Assistant can show whether the tag fires. Check the status of each conversion action in the Goals section, which flags actions that have stopped recording.
How do you connect conversions to real sales?
Even perfectly counted enquiries are not sales. A quote request is a question, not a customer. To make Google Ads learn from what actually earns you money, you need to tell it which enquiries turned into work.
There are two related tools for this.
Enhanced conversions for leads
Enhanced conversions take details the person entered on your form, such as their email address or phone number, and scramble them one-way (hashing) before they are sent to Google. Later, when that lead becomes a customer, you upload the same details and Google matches the sale back to the ad click that started it. On Procon's quote forms this is now in place, so won customers can be matched back to the click rather than stopping at "someone filled in a form".
Tell visitors in your privacy notice how their details are used, and keep the data you upload to what is needed.
Offline conversion imports
Offline conversion imports let you upload outcomes that happen away from the website: a quote accepted, a job completed, a contract signed. You can include a value, so Google Ads sees that one enquiry was worth a small job and another a large one.
For most small businesses, a simple weekly or monthly upload from a spreadsheet or CRM is enough to start. It cannot promise better results, but it gives the bidding system, and you, a number much closer to real sales.
How often should you check Google Ads against real sales?
Monthly, as a minimum, and immediately after any change to the site, forms, phone numbers or cookie banner.
- Export last month's conversions from Google Ads, split by conversion action.
- Count the real enquiries you received in the same period from your inbox, call log or CRM.
- Compare them. A difference you cannot explain, in either direction, needs investigating before you change bids or budgets.
- Mark which enquiries became customers and upload them as offline conversions.
- Check each conversion action's status for any that have stopped recording.
| Cause | What it looks like | Fix |
|---|---|---|
| Page views or button clicks counted | More "leads" than enquiries | Count server-accepted enquiries only |
| Duplicate tags or "Every" counting | One test enquiry shows as two | One primary action, count "One" |
| Spam and bots | Leads you never hear from | Filter spam before the conversion fires |
| Phone taps counted | High call numbers, quiet phone | Forwarding numbers with a call length threshold |
| Consent gaps | Fewer conversions than enquiries | Connect the banner to consent mode, test both choices |
| Broken tag after a site change | Conversions drop to near zero | Test enquiry and call after every change |
When is it worth getting a second opinion?
If your Google Ads numbers and your real sales do not agree and you cannot see why, a fresh look usually finds the cause quickly. A free Google Ads audit checks the conversion setup first, because every other recommendation depends on it being right. Sometimes the honest outcome is to fix the tracking and hold spend until the numbers can be trusted.
If you would rather hand it over, Google Ads management with conversion tracking tied to real enquiries is run by me personally, in your own account, for a flat monthly fee given on the audit call and never a percentage of spend. If the landing pages are part of the problem, read the Google Ads landing page checklist next.
No setup can promise a perfect number. The aim is a conversion count you can check against your own sales, and a routine that catches it when the two drift apart.
