What call tracking actually attributes

Call tracking is sold as attribution and bought as certainty, and the gap between those two is where a lot of trades marketing budget goes to die. It is worth being precise about what the tool actually knows.

What it does know

Dynamic number insertion swaps the phone number on the page based on how the visitor arrived. A visitor from a Google ad sees one number, a visitor from organic search sees another, and a visitor who typed the domain directly sees a third. When the phone rings, the number dialed tells you the source.

That is genuinely useful, and it is the single highest value instrumentation on a trades site, because most trades leads arrive by phone and a form-only measurement setup is blind to the majority of them.

What it does not know

Whether the call was any good. The tool records that a call happened, how long it lasted, and which source produced it. It does not know whether the caller was in the service area, whether they booked, whether the job was profitable, or whether they were a supplier calling the number they found online.

A campaign that produces forty calls a month at a low cost per call looks better in the dashboard than one producing twelve. If the forty are mostly out of area and the twelve are all booked replacements, the dashboard is telling you to spend money in the wrong place.

Call duration is a weak proxy

The standard workaround is to count calls over a threshold, usually sixty or ninety seconds, as qualified. It is better than counting all calls, and it is still a proxy. A three minute call explaining that you do not serve that township counts as qualified. A forty second call from a repeat customer booking a routine job does not.

Use the threshold to filter obvious noise, not to decide budget.

Closing the loop

The number that should drive spend is cost per booked job, by service line. Getting there means the outcome of the call has to come back from wherever the job is actually recorded, which is the CRM or the scheduling system, and be joined to the source that produced it.

That join is the work. It is less interesting than the tracking setup and it is what makes the tracking setup worth having. Without it you have a very precise measurement of an intermediate step.

A practical order of operations

Install the tracking and get sources separated. Then get the calls into the CRM against the source. Then get the job outcome back onto the record. Each step is useful on its own, and skipping to the last one without the first two is how these projects stall.

Most contractors stop after step one, which is why so many can tell you their cost per call and not their cost per job.