How to Add a “How Did You Hear About Us?” Survey

A “How did you hear about us?” survey is the cheapest attribution tool you will ever deploy — one field, no engineering, and it captures the dark-social and offline channels your pixel physically cannot see.

It matters because those invisible channels are large: in TrackRev platform data, Direct is the highest-converting channel of all at 7.1% click-to-paid (Q2 2026), and a big slice of that “direct” is really word-of-mouth, private shares, and podcast mentions that no tracked system can name.

A survey is the only instrument that can.

But placement and wording decide whether the survey returns signal or noise, and most implementations quietly destroy their own data by asking at the wrong moment with the wrong options.

This guide covers where to put the question, how to word it, whether to use a dropdown or open text, and how to analyse the answers without over-reading them.

Done well, it recovers attribution for the channels your pixel misses, so you can see which of them produces paying, retained customers. Done badly, it produces a column of “Google” and “a friend” that tells you nothing.

Why this matters for your revenue

The channels a survey captures are often the ones driving your best customers, and they are invisible to everything else you have.

Word-of-mouth referrals, community recommendations, and unlinked podcast mentions convert well precisely because they come from a trusted human — but because they leave no referrer and no trackable link, your pixel files them under “direct” and your ad platforms take no credit.

If you make budget decisions on tracked data alone, you will systematically underinvest in these channels, because your dashboard cannot see them working.

The survey closes that gap for the price of a single form field.

When 15% of new customers write “a colleague recommended you” or “I heard your founder on a podcast”, you suddenly have a reason to invest in the community programme or the podcast sponsorship that tracked attribution said produced nothing.

The revenue impact is asymmetric: the survey costs almost nothing to run, and the channels it reveals are frequently high-converting demand you were about to defund by accident.

Pair it with your tracked channel analytics and you cover both the measurable and the invisible sides of attribution. See dark-social attribution for what lives in that gap.

Where to put the survey

Placement is the single biggest determinant of data quality. Ask at the wrong moment and you either depress response rates or contaminate the answer. Three placements, in order of preference.

Post-signup or early onboarding (best)

The best moment is immediately after signup or in the first onboarding step — the account is already created, the customer is in a low-friction, cooperative frame of mind, and their memory of how they found you is fresh.

Asking here maximises both response rate and answer quality, because nothing is at stake for the customer and the discovery is recent.

A single optional question on the welcome screen, or as the first step of onboarding, is the standard high-signal placement.

Checkout (why it usually backfires)

Asking at checkout feels efficient — you catch paying customers — but it usually backfires.

The moment of payment is the highest-friction, highest-anxiety point in the funnel, and any extra field there risks depressing conversion for a piece of data you could have collected freely one step later.

Even when it does not hurt conversion, answers given while someone is trying to complete a purchase are rushed and low-quality. Attribution data is never worth a measurable dent in conversion.

Friction at the moment of payment

The rule is simple: never add an attribution question to the critical path of a purchase.

Every field between a customer and their completed payment costs you some conversions, and an optional “how did you hear about us?” is exactly the kind of non-essential field that should live after the transaction, not inside it.

If you want paying-customer attribution specifically, ask on the post-purchase confirmation screen or the first login — after the money is safely captured, never before.

In-app, after first value

A strong alternative for product-led motions is to ask once the customer has experienced first value — after they have completed the core action your product exists to deliver.

Engagement is high, goodwill is high, and the customer is invested enough to answer thoughtfully.

The trade-off is that some users never reach first value, so you miss them; but the answers you do get tend to be the highest quality of any placement, because they come from engaged customers who are happy to help.

How to word the question

Wording is the second lever. The goal is a question a non-marketer answers accurately without thinking about marketing.

Use the channel’s own name

Phrase the question and options in the customer’s language, not yours. “How did you hear about us?” works because it is plain; “What was your acquisition channel?” does not, because customers do not think in channels.

Options should name concrete things a person recognises — “A friend or colleague”, “A podcast”, “YouTube”, “Google search”, “Reddit or a community” — rather than marketing abstractions like “referral”, “organic”, or “paid”.

The closer the wording is to how the customer actually experienced finding you, the more accurate the answer.

Avoid jargon like “channel” and “referral”

Marketing jargon in a customer-facing survey does two kinds of damage: it confuses respondents, lowering answer quality, and it biases them toward whichever option they half-understand. “Referral” means a specific thing to you and something vague to them.

Strip every internal term. If an option cannot be phrased in words your least technical customer would use unprompted, rephrase it until it can.

The survey is an instrument pointed at customer memory, and it only reads cleanly if the customer understands the question exactly.

Answer format: dropdown vs open text

The format decides whether you get analysable data, rich detail, or a mess. The right answer is a hybrid.

The short dropdown

A short list of predefined options is what makes the data analysable — you can count “podcast” answers across a quarter only if “podcast” is a consistent option rather than free text spelled ten ways.

Keep the list short (six to eight options) and channel-shaped, ordered roughly by how common you expect each to be.

A list that is too long overwhelms the respondent and depresses answer quality; a tight list of the channels that actually matter to your business is both easier to answer and easier to analyse.

The open-text catch-all

A dropdown alone cannot capture the channel you did not anticipate, and those surprises are often the most valuable finding.

So pair the dropdown with a single optional open-text field — either a standalone “Anything else?” or an “Other” option that reveals a text box.

This is where you discover the newsletter you did not know was recommending you, or the specific podcast episode that drove a spike. The structured options give you countable trends; the open text gives you the discoveries.

Why “Other” needs a text box

An “Other” option with no text box is wasted data — it tells you a channel exists but not which one, which is the one thing you needed to know.

Always attach a text field to “Other”, and read those responses regularly, because a recurring write-in is a signal that your dropdown is missing a real channel.

When the same write-in appears often enough, promote it to a first-class dropdown option so you can count it going forward. The open text is how the survey’s option list stays current with how customers actually find you.

A starter option list

As a starting point, the option set below covers the channels most SaaS surveys need, phrased in customer language, with a note on what each one is really capturing and whether your pixel could have seen it.

Trim it to the channels that matter for your business, and let the open-text field surface the rest.

Placement options compared

The trade-off across placements is response rate against answer quality against risk to conversion.

How to analyse the responses

The final lever is interpretation, and it is where most teams go wrong by taking the numbers too literally.

Directional, not precise

Survey data is directional. It tells you a channel exists and roughly how much it matters, not its exact revenue contribution — because answers are last-touch memories, biased toward recent and prestigious touches, and given by only a fraction of customers.

Read the trends: which invisible channels show up at all, which are growing quarter over quarter, which spiked after a specific event.

Do not build a precise attribution model out of survey percentages; that is asking a memory to do a measurement’s job.

Reconciling with tracked data

The highest-value analysis is comparing survey answers against your tracked data. Where they agree — a channel both your pixel and your survey rank highly — you have strong confidence.

Where the survey names a channel your pixel filed under “direct”, you have found a piece of your dark-social demand and can size it.

Feed those write-ins and dropdown counts back into how you read your tracked channel revenue, using the survey to reallocate part of the direct bucket to its real sources. The two datasets are most useful side by side.

The honest limitation of survey data

Be clear-eyed about what a survey can and cannot do.

It cannot give you precise percentages, it cannot tie an answer to a specific revenue amount without extra work, and it is answered by a self-selected minority of customers whose memories are imperfect and biased.

It will over-credit whatever is memorable and under-credit whatever is quiet, and it compresses a multi-touch journey into a single answer. None of this makes it useless — it makes it directional.

The mistake is not running a survey; it is treating its output as measurement rather than as the only available read on channels no measurement can reach. Use it for discovery and sizing, and let your tracked data handle precision.

When NOT to use TrackRev

TrackRev is a tracked-attribution platform, not a survey tool — it will not host your “How did you hear about us?” question, and this guide is deliberately platform-neutral about where you build the survey itself.

What TrackRev provides is the tracked half of the picture the survey complements: a first-party pixel, a touchpoint log, and a billing join that make your measurable channels precise, so the survey only has to cover the genuinely invisible ones.

If your growth is almost entirely word-of-mouth with no measurable channels to attribute, a survey and a spreadsheet may be all you need for now, and a tracked platform is more than the problem requires.

A survey plugs the one gap tracked attribution cannot — but the tracked side only stays trustworthy when links, revenue, and affiliates share a single definition of a sale, which they do not when spread across tools.

The default stack pairs Bitly Growth (~$35/mo) for links with Rewardful Starter (~$49/mo) for affiliates — $84+/month for two systems whose numbers never reconcile.

TrackRev is $39/mo for link tracking, revenue attribution, and the affiliate programme on one first-party pixel, so your measurable channels are consistent and your survey has less “direct” to explain away.

Start on the free tier at /pricing, which covers 1,000 events.