Call Tracking for Google Ads Without Losing the Lead
If customers call instead of filling in forms, most of your best leads are invisible. Here is how to track calls properly and feed them back into your bidding.

For a lot of businesses, the best leads never touch a form. They ring the number at the top of the page, book a job in ninety seconds, and disappear from your reporting entirely.
If phone calls matter to your business and you are not tracking them, your ad account is optimising toward the leads it can see, which are usually the weaker ones.
The three kinds of call, and how each gets tracked
Calls straight from the ad
Someone taps the call button in your ad without ever landing on the site. Google Ads tracks these natively through call extensions. Set a minimum call length so a two second misdial does not count as a lead.
This is the easiest one and it is usually already working.
Calls from your website
Someone clicks the ad, lands on your site, and rings the number displayed there. This is where most businesses lose the trail.
Google Ads offers a forwarding number that swaps the displayed number for a tracking one. It works, but it is limited to Google traffic and it replaces your real number, which some businesses dislike.
The more capable route is dynamic number insertion through a call tracking provider. Each visitor sees a different number, so the call can be tied back to the exact campaign, and it covers every channel rather than just Google.
Calls that turn into revenue later
Somebody calls, gets a quote, and books three weeks later. That sale is real and your ad account knows nothing about it.
This is where call tracking meets offline conversion tracking. You capture the click identifier when the visitor arrives, store it against the call record, and push the outcome back once the job is confirmed.
Setting up click to call tracking properly
A quick win first. Track clicks on your phone number link as a conversion. It is not a real call, because people click and hang up, but it gives you a directional signal within an hour of setup.
Treat it as a soft signal. Do not use it as a primary conversion for bidding, because you will end up optimising for people who tap a number and change their mind.
Making the data actually useful
Tracking that a call happened is only half of it. Whether it was any good is the important part.
Set a sensible minimum duration. Under thirty seconds is rarely a real enquiry. Sixty seconds is a reasonable starting point for most service businesses.
Separate first time callers from existing customers. Otherwise your support calls will look like a wildly successful campaign.
Record the outcome. Booked, quoted, wrong number, or not interested. This is what turns call data into something worth bidding on.
Send the good ones back as conversions, not every call. Optimising toward call volume gets you more wrong numbers.
Common mistakes
Counting every call as a conversion. Your bidding learns to find people who ring for directions.
Forgetting the phone number appears in more than one place. Header, footer, contact page, and any sticky mobile bar all need the same treatment.
Breaking the number on mobile. Always test that tapping it actually starts a call. A tracking setup that stops people calling you is worse than no tracking.
Ignoring calls from organic and social, which makes paid look like the only channel producing calls.
Is it worth it?
If more than about a fifth of your leads arrive by phone, yes, and usually quickly. Service businesses, clinics, trades, legal, property, and anything with a considered purchase see the biggest difference.
If you are ecommerce and almost nobody calls, skip it and spend the effort elsewhere.
Need this fixed properly?
If your best leads arrive by phone and none of them reach your ad reporting, that gap is worth closing. I set up and repair tracking for ecommerce stores and lead generation businesses, so the numbers in your ad accounts match the money in your bank account.
Have a look at what I do, or book a call and we can go through your setup together.






















