What is first-touch attribution?

First touch answers one question well. What brings new people into contact with us at all.

For a business that needs awareness more than it needs closing, that is the useful question, and last-touch models systematically hide the answer.

The distortion is obvious once stated. A blog post read eight months ago gets the whole sale, and the phone call that closed it gets nothing.

It also depends on remembering that first contact, which most small-business tracking cannot do past a cleared browser or a change of device.

Say a pediatric dental office gets a new family in August. Here's the real path. In April the parents saw the practice's banner at a Little League field. Two months later the mom searched for a kids dentist on her phone, tapped the listing and browsed for two minutes. By August she had forgotten that visit, and she typed the practice name on her laptop, clicked through and called.

The banner left no record. June's visit set a cookie on her phone, but Safari deletes script-set cookies after seven days, and she switched devices anyway. So the August visit looks brand new. The system's first touch is a brand search, dated the same day as the call.

On the report, first touch and last touch are the same click. The practice concludes that name recognition drives new patients, which is true, and learns nothing about what built it. That $600 banner looks like it did nothing.

Asking callers is the repair, and the wording decides what you learn. How did you find us today gets you the last step: I searched your name. Ask how did you first hear about us and the answer reaches further back: my son's team plays at that field.

Use the second wording, and ask it early in the call while the caller is still in a mood to tell you things. Record the answer in their own words before sorting it into a category. Someone told me is vague, but someone at church told me is a lead on where your referrals come from.

Expect some people to say they don't remember, and don't push. A forced answer is worse than a blank, because most will name whatever they saw last.

Tally the answers quarterly alongside your tracked first-touch report. Where the two lists differ, the spoken one is usually closer to the truth about awareness. Your tracked one is closer on what people clicked.

Phone calls give you a chance the web does not: you can ask. A caller who says how they heard about you has just supplied a first touch no cookie would have caught.

That line sits in the transcript and the summary, so it is searchable later rather than lost at the end of the call.

Nothing about this is automatic. It works because the assistant asks the question you told it to ask.