Self-Reported Attribution: When You Just Ask the Buyer Where They Came From

A deal closes for real money, and you open the record to see where it came from. The source field says direct. Direct is supposed to mean the buyer typed your web address straight into their browser, with no ad or search or link in front of it. Nobody does that for a company they have never heard of. The label is accurate and useless in the same breath. Whatever pointed that person at you is the thing you actually want to know, and your analytics never saw it happen.

There is an older way to find out, and it is almost embarrassingly simple. You ask the buyer. On the form they fill in, or in the first conversation, you put one question in front of them: how did you hear about us. What they type or say back is self-reported attribution, and for a lot of businesses it is the closest thing to the truth they will ever get about where their customers come from.

How did you first hear about us?

Your source field

Direct

What the buyer typed

A podcast, then a colleague sent me your case study

What self-reported attribution actually is

Self-reported attribution is the buyer telling you, in their own words, what led them to you. No pixel or cookie doing the guessing on their behalf. You collect it by asking, usually with a short how did you hear about us field on a form or a question early in a sales call.

Everything else in the attribution world works from what your systems can observe: which link got clicked, which UTM tag rode along with it. It captures the small slice of the journey that touches your own site and links. Everything earlier, in the places your systems were never present for, is simply missing from it. Self-report comes from the one witness who sat through the whole thing, memory and all, and it can point at parts of the customer journey no tracking tool was ever going to catch.

Why the buyer's answer is often the most honest one you have

Most of a business-to-business buying decision now happens where you cannot see it. People research in places that never report back to you: a podcast on their commute, a Slack group for their industry, a peer who used you last year, a review on a site you have never logged into. This is what people mean by the dark funnel, the stretch of the buyer's journey that leaves no trace in your analytics. Most of it is over before a buyer ever fills in a form, and none of it shows up in a tracked source.

A tracked model can only credit what it can tag. When the real reason someone bought was a recommendation in a private community, the best your dashboard can do is credit the Google search they ran afterward to find your site. Ask them directly and you get the podcast, or the review that a tracked click was never going to record. That is why, for a dark-funnel journey, the buyer's own answer is often more honest than the model. We walk through how the tracked models work, and where they fall short, in marketing attribution. And as the tracking itself gets weaker, which we cover in cookieless attribution, asking people directly turns into one of the few reliable signals you have left.

How to ask the how did you hear about us question well

The wording carries more weight than it looks. An open text box gets you the real answer in the buyer's language, which is where the surprises live, but it is a chore to count later because forty people will describe the same podcast forty ways. A dropdown of fixed options is easy to tally and steers people toward whatever you listed, so you mostly learn what you already suspected. A middle path works best for most teams: a short list of the sources you actually care about, plus an other box people can type into freely.

Where you ask matters as much as how. Put the question on the forms that sit close to a buying decision, the demo request or the contact form, not on a newsletter signup where the answer is noise. Then ask it again out loud in the first call, because people write one thing on a form and say something fuller once a person is actually listening.

Whatever you collect is wasted if it lands in a free-text notes field nobody can filter. Give it a real home: a dedicated property on the contact or deal record in your CRM, so you can run a report on it. If you cannot pull up a clean list of what everyone who bought this quarter told you, the question is not really doing anything for you.

Reconciling what people say against what your tracking shows

You will end up with two versions of where a customer came from, and they will disagree. The buyer says they found you through a podcast. The source field says paid search. Both are recording something real. The podcast created the demand, and the search was how they navigated to you once they had already decided to look. You run into trouble when you decide one of them is the real source and discard the other.

Read them together instead. The tracked source can give you the last click and little beyond it. The self-report fills in the months before that, back when the buyer was actually making up their mind. When a lot of your closed deals name a channel your analytics barely registers, pay attention to that gap. It usually means demand is being created somewhere you are not measuring, which is often the most useful thing self-reported data will hand you.

Where it goes wrong

Self-report is human, so it carries every flaw of human memory, starting with the fact that people forget. They compress a six-month journey into whatever they can recall in the two seconds they spend on your form. Recency bias does most of the damage: someone genuinely discovered you through a conference talk in spring, but the last thing they did was Google you, so they type Google and mean it.

The podcast that actually changed their mind gets lost, because they do not remember it the way your attribution report needs them to. You also get vague answers like a friend, or online, or I do not remember, and a sample that is smaller and messier than a clean tracked dataset. All of which means self-reported attribution points you in a rough direction rather than handing you a clean number. Use it to see patterns and correct your assumptions. Do not wire sales commissions or a big budget swing to it on its own, because the numbers were never precise enough for anyone to stake that much on.

Where to start

You can have this running this week. Add one field to your main contact and demo forms, ask the same question in the first call, and store the answer where you can report on it. Once a month, read what buyers told you next to what your tracking recorded, and look for the channels that keep showing up in people's own words while staying invisible in your dashboard.

Reading the two side by side is something we build into measurement from the start inside our marketing operations work, because the two together tell a truer story than either one alone. None of it takes a project. Add the question, ask it again in the first call, and once a month read the answers next to what your tracking recorded. You will usually learn more about where customers come from than the dashboard managed to tell you all year.

Common questions