Half Your Leads Call Instead of Filling In the Form, and You Are Not Counting Them

For a large share of service businesses, the phone is the primary conversion and the contact form is the afterthought. People with a burst pipe, a toothache, or a legal problem at eleven at night do not fill in a form and wait. They call. If your conversion tracking only counts form submissions, you are optimising a campaign against the smaller and less urgent half of your results.

This one differs from most tracking failures because nothing is broken. The form conversion works perfectly. The tag fires, the data is clean, the reporting is accurate as far as it goes. The problem is the scope of what is being measured, which means no diagnostic tool will ever flag it. Everything it looks at is fine.

Why partial measurement is worse than it sounds

If calls were simply missing from a report, you could mentally add them back. The damage is that Smart Bidding cannot. It optimises toward the conversions it receives, and it has no way to know that a whole category of outcome is invisible to it.

So consider two campaigns. One attracts researchers who browse, compare, and eventually submit a form. The other attracts urgent, high-intent customers who land on the page and immediately call. With only form tracking, the first campaign looks like your star and the second looks like a money pit. The algorithm shifts budget toward the researchers, you cut the campaign that was producing the actual revenue, and every number in your reporting supports the decision.

That is the real cost here. Not a reporting gap, but a systematic bias that pushes spend away from your best customers, with full supporting evidence.

Three different things people call call tracking

Most confusion in this area comes from conflating three distinct mechanisms. They have different setups and different failure modes.

1. Calls directly from the ad

Somebody taps the call button on your ad without ever reaching your website. Google can count this because the call went through its own forwarding number, and it is the easiest of the three to get working.

2. Calls from a number on your website

Somebody clicks the ad, lands on your site, reads for a while, and then calls the number on the page. This is usually the largest group and the one most often untracked. Making it work requires the Google tag plus a website call event snippet, and the mechanism is that Google replaces the number displayed on your page with a forwarding number for visitors who arrived from an ad. The swap is what connects the call back to the click.

3. Imported calls

Calls recorded by a third-party call tracking platform or your CRM, uploaded back into Google Ads afterwards. This is the route when you need call outcomes rather than just call counts, for example counting only calls that turned into booked jobs.

Where website call tracking goes wrong

Setup two is where most of the failures live, because it depends on Google being able to find and replace a phone number on a page it did not build.

1. The number is not in a readable format. A phone number rendered inside an image, baked into a graphic in the header, or drawn in a way the script cannot read cannot be swapped. The page still shows a number. It is just always the original one.
2. The number is added late by JavaScript. If the number is injected after the swap script has already run, it misses it. This is increasingly common with single page applications and page builders, where content appears in stages.
3. Call reporting is not enabled. There is an account level setting governing this, and the website call snippet will not do its job without it.
4. The minimum call length is set badly. You choose how long a call must last before it counts as a conversion. Calls shorter than the threshold are not counted at all. Set it to sixty seconds and you discard genuine bookings that took forty seconds. Set it to five and you count every wrong number and hang-up.
5. Forwarding numbers are not available everywhere. Google publishes the list of countries where forwarding numbers can be used, and it changes. If your client operates outside that list, this mechanism is not available and a third-party platform is the route.
6. Two systems are fighting over the same number. If a dedicated call tracking platform is already swapping the number on the page and Google is also trying to, the result depends on which script wins, which can vary by page and by load order. Pick one system per number.

The double counting risk

Once you add call tracking, check the other direction too. It is easy to end up counting one human being twice.

• A customer who calls, does not get through, and then submits the form produces a call conversion and a form conversion. Two conversions, one person.
• A call counted both by Google’s native tracking and by an imported upload from your call platform is the same call recorded twice.

Neither is catastrophic, and both inflate your numbers in a way the client will eventually notice when they compare against reality. Decide which system owns which conversion and make sure the other one is not also reporting it.

How to verify it properly

Checking the setting is not verification. Verification is making a real phone call. Do this on an actual mobile phone, not a desktop.

1. Search for one of your own live ads and click it, so your visit genuinely originates from an ad click.
2. Look at the phone number on the page. Confirm it is different from the real number. If it has not changed, the swap did not happen and nothing downstream will work, whatever the settings say.
3. Call it. Stay on the line comfortably longer than your minimum call length, so the call qualifies.
4. Check that the call reaches the client’s actual phone. A forwarding number that does not forward is a far worse problem than untracked calls.
5. Wait, then confirm the conversion appears. Call conversions are not instant.
6. Repeat on each important landing page template, because the swap can work on one layout and fail on another.

Google also provides Tag Assistant for checking the implementation, and an inactive tag status there usually means the tag or the snippet is not installed correctly. Use it to diagnose, but let the real call be the thing that convinces you.

Native tracking or a dedicated platform?

Google’s own call conversion tracking is free, built in, and adequate for the common case: you want to count calls from ad traffic and feed them to bidding. For a single-location business with one phone number, start here.

A dedicated call tracking platform earns its cost when you need more than a count. Recordings for quality control, outcome tracking so only booked jobs count as conversions, multiple numbers across locations, or attribution across channels beyond Google. For a client where the phone is the business, that detail usually pays for itself quickly.

What you should not do is run both at full strength on the same number and hope. Choose the system that owns the number, and if you run both, make sure they are scoped to different numbers or different pages deliberately.

Set expectations honestly

Call attribution is inherently less precise than form attribution. Somebody can see an ad on their phone, write the number on paper, and call three days later from a landline. No system catches that. Somebody can call the number they already had saved from a previous visit.

Tracked calls will always undercount real calls. That is fine, as long as everyone knows it, and it is still enormously better than counting zero. Tell the client this up front, because the alternative is being asked later why tracked calls are lower than the number the receptionist counted, which is a much harder conversation after the fact than before it.

Where to start

If you manage lead generation accounts, the first question worth answering is not how to configure this. It is how many of your clients have a prominent phone number on the site and no call conversion action at all. For most agencies the answer is more than they expect, and each one is a campaign being judged on a fraction of its results.

Our free tracking scan reads the public source and the published tag manager container in about thirty seconds and shows you every tag and conversion mechanism installed on a site, which is a quick way to see whether call tracking is present at all. It is honest about the boundary: whether a forwarding number actually swaps and whether a real call records a conversion can only be settled by making the call, exactly as described above.

Call tracking also breaks quietly when a site is redesigned, when a number changes, or when a new plugin starts rendering the number differently. Taglert rechecks every connected account every thirty minutes and emails you when conversion tracking changes, so a template update does not silently stop counting your most valuable leads.