The Complete Guide to 'How Did You Hear About Us?' Surveys for B2B (2026)

"How did you hear about us?" is the most underrated question in B2B marketing. Most companies skip it entirely, bury it in a dropdown nobody thinks about, or collect responses that sit in a spreadsheet no one analyses. This is a mistake. Done right, HDYHAU surveys reveal the channels that actually create demand — the ones your analytics platform will never show you.

The problem is not that companies don't ask the question. Many do. The problem is how they ask it — and what they do with the answers. A dropdown menu with ten options and no text field produces data that confirms what you already believe. An open-text field with the right framing produces data that changes how you allocate your budget.

This guide covers everything: why most HDYHAU implementations fail, how to design one that works, when to ask, what answer options to include, how to analyse responses at scale with AI, and the benchmarks you should expect in B2B. It is written for marketing leaders who suspect their attribution data is lying to them. It probably is.

Why HDYHAU Actually Matters (When Done Right)

Let's start with why most implementations fail. The standard approach is a dropdown on the demo request form: Google, LinkedIn, Referral, Event, Other. The prospect clicks whatever is fastest — usually the most recent touchpoint they remember — and moves on. The data goes into a CRM field that nobody queries. Marketing uses it once a quarter to justify existing spend. Nothing changes.

Dropdowns are worthless because they force a single answer to a multi-touch question. A buyer who spent six months listening to a podcast, saw three LinkedIn posts, got a recommendation from a colleague, and then Googled your brand name will select "Google." Your dashboard will credit Google. Your podcast investment will look like it produces nothing. Your colleague-referral channel — the one that actually created the demand — will not appear anywhere in your data.

The magic happens when you switch to open text. When you ask "What first put us on your radar?" instead of "How did you hear about us?" and give people space to write a sentence, the responses transform. You stop getting channel labels. You start getting stories.

"My colleague Sarah mentioned you after your talk at SaaStock. I then listened to your episode on Lenny's Podcast and started following you on LinkedIn."

— An actual HDYHAU response. Try getting this from a dropdown.

That single response tells you more than a thousand dropdown selections. It tells you the conference talk created awareness, the podcast deepened trust, LinkedIn maintained presence, and a peer recommendation was the catalyst. No attribution model in the world would have connected those dots. But the buyer just told you, for free, in fifteen seconds.

The key principles for making HDYHAU work: ask about first awareness, not last touch. Use open text, not dropdowns. Frame it as "What first put us on your radar?" Follow up for specifics — which podcast, which person, which event. And actually analyse the responses. That last part is where most companies fall down, and where AI changes everything. More on that shortly.

When To Ask: The Timing Question

Where you place the HDYHAU question matters almost as much as how you phrase it. There are three viable options, each with trade-offs. The right choice depends on your sales motion, your conversion volume, and how much friction your form can tolerate.

For most B2B companies, placing the question on the demo request or signup form is the right starting point. Yes, it adds a field. But a single open-text field adds negligible friction — far less than the "company size" and "job title" fields most companies already require. And the data you get back is worth more than all of those fields combined.

How AI Transforms HDYHAU Analysis

Here is the historical problem with open-text HDYHAU responses: they are messy. People misspell podcast names. They write "a friend told me" without specifying which friend or in what context. They combine three channels into one run-on sentence. At 20 responses a month, a marketing manager can read and categorise them manually. At 200, it becomes a chore. At 2,000, nobody does it.

This is why most companies default to dropdowns — not because dropdowns are better, but because they produce data that is already categorised. The trade-off is that the categories are wrong. You get clean data that tells you the wrong story. Open text gives you the right story in a format that is hard to analyse. Until now.

AI language models change this equation completely. They can read a response like "my mate Dave who runs ops at Grafton mentioned you at drinks, then I heard the founder on Pavillion's podcast and thought I should check it out" and extract: peer recommendation (Dave, operations role, Grafton), podcast (Pavillion), social proof path. They handle misspellings. They identify specific sources. They spot patterns across thousands of responses that no human would catch — like the fact that 14% of your enterprise leads mention the same Slack community, or that a specific podcast episode drove a spike in demo requests three weeks after it aired.

This is what First Signals does. It takes the messy, rich, honest open-text responses that buyers give you and turns them into structured channel intelligence — automatically categorised, source-identified, and tracked over time. The result is that you get the accuracy of open text with the usability of a dashboard.

The gap between what HDYHAU reveals and what analytics dashboards show is not a rounding error. It is a fundamental misalignment between where you think your pipeline comes from and where it actually comes from. Companies that close this gap routinely reallocate 20-40% of their marketing budget — away from channels that capture demand they were going to get anyway, and toward channels that create new demand they would otherwise never see.

The Dark Funnel Problem

There is a name for the channels that create demand but don't show up in your analytics: the dark funnel. It includes every touchpoint that happens outside your tracking — conversations between peers, podcast episodes listened to on a morning commute, Slack community threads, conference hallway conversations, LinkedIn posts read but never clicked.

The dark funnel is not a small or marginal part of B2B buying. For most companies, it is the majority. When you ask buyers how they actually made their purchasing decision — not which ad they clicked, but what made them believe your company could solve their problem — the answer almost always involves a channel you cannot track. A trusted person told them. A podcast built credibility. A community thread gave them confidence.

Your attribution dashboard shows you who came to the party. HDYHAU tells you who sent the invitation. In B2B, the invitation is almost always more important than the door they walked through.

HDYHAU is the only scalable method for measuring the dark funnel. It is not perfect — buyers have imperfect memories, some will still default to the most recent touchpoint, and the data requires interpretation. But imperfect signal from the right channels is infinitely more valuable than precise measurement of the wrong ones.

Implementation: A Practical Checklist

If you are ready to implement or overhaul your HDYHAU survey, here is the sequence that works. This is not theory — it is the process we use with every First Signals customer, refined across hundreds of implementations.

Frequently Asked Questions

The Question You Should Be Asking Tomorrow

You do not need a six-month attribution project. You do not need to replace your analytics stack. You do not need to hire a data scientist. You need to add one open-text field to one form and start reading what your buyers tell you.

The companies that do this consistently find that their understanding of their own pipeline is wrong — not slightly wrong, but structurally wrong. They discover that channels they were ignoring drive the majority of their demand, and channels they were investing heavily in were capturing demand that would have arrived anyway. That discovery, uncomfortable as it is, is the beginning of marketing that actually compounds.

Start with the question. Make it open text. Frame it around first awareness. Follow up for specifics. And when the volume gets too high to read manually, let AI do the categorisation. The data is already there, sitting in the heads of every person who fills in your form. You just have to ask for it properly.