The "how did you hear about us" examples on this page are written for an event registration
form: one question, a list of eight to twelve answer options that match the channels in your
budget, and an "Other" box with free text. The answers are self-reported attribution, meaning
the source the registrant names, kept beside the source your tracking recorded. Read side by
side, the two columns show which channels your tracked numbers undercount.

## What is self-reported attribution?

Self-reported attribution means asking the buyer where they heard about you and recording the
answer as a source. For an event, the buyer is the registrant and the place to ask is the form.

Tracked attribution is what software recorded: the campaign tag on the link, the referring
site, the email the click came from. It is precise about the last recorded step and blind to
anything that happened off the link. A colleague's recommendation, a chapter meeting, and a
printed mailer leave no tag. The registrant then types the event name into a search engine or
opens an old email, and that last step takes the credit.

Self-reported answers have the opposite profile. They capture the remembered reason and miss
the mechanics. People forget, pick the first plausible option, and name the most recent thing
they saw. The [event marketing attribution](https://eventiq.io/md/event-marketing-attribution) guide covers the
credit models that sit on top of tracked touches.

One answer per registrant is a single data point. Do not present it as multi-touch attribution.

## How should you word the question on a registration form?

Ask about the event, ask about the first time, and ask once. "Us" is ambiguous for an
organization with twelve programs, and "hear" without "first" invites the most recent touch,
which your tracking already has.
| Wording | What it measures | Use when |
| --- | --- | --- |
| How did you first hear about [event name]? | Discovery: the channel that introduced the event | Default. You want to know which channels reach new people |
| What prompted you to register today? | Trigger: the last push before the decision | You are testing deadlines, reminders, or offers |
| Did someone recommend [event name] to you? If so, who? | Referral, with a name | You run a referral or group program |
Pick one wording per form.

**Placement.** Put the question at the end of the attendee details step, before payment. On
the payment step it competes with card entry. In the
[post-event survey](https://eventiq.io/md/blog/post-event-survey-questions) it reaches only the people who attended
and chose to respond, weeks after the memory was fresh.

**Required or optional.** Make the question required only if the list includes "I don't
remember." A required question with no exit produces a confident-looking answer from someone
who guessed. If your form cannot offer that exit, leave the question optional and report the
response rate beside every share.

## Which answer options should you offer?

Offer options the registrant recognizes and that you can map to a line in the budget.
| Option on the form | Maps to | Note |
| --- | --- | --- |
| Email from [organizer name] | Email | Name yourself, so partner emails do not land here |
| A colleague or manager told me | No spend line: word of mouth | Tracking cannot see it |
| I attended a previous year | No spend line: retention | Separates returning demand from new reach |
| A speaker or presenter | Speaker promotion | Covers speaker posts and direct invitations |
| A sponsor or exhibitor invited me | Sponsor and exhibitor promotion | Evidence for the sponsor conversation as well |
| An association, chapter, or partner organization | Partner promotion | Add a text box for the name if partners are paid |
| LinkedIn or other social media | Paid social and organic social | People cannot reliably tell a sponsored post from an organic one |
| Search engine | Paid search and organic search | Often the last step before the form |
| Industry publication or newsletter | Media partnerships and advertising | Name the titles you pay, if there are few |
| An AI assistant or chatbot | No spend line yet | Keeps these answers out of "Other" |
| Other (please tell us) | Coded by hand | Free text. Give any recurring answer its own option |
| I don't remember | Its own row in the tally | The exit that keeps a required question honest |
Five rules keep the list usable:

- **Eight to twelve options.** A longer list gets skimmed, and the first few options absorb
  the answers.
- **Single select.** Multi-select produces totals that exceed registrations, and any rule for
  dividing them is an assumption.
- **Randomized order** where the form allows it, with "Other" and "I don't remember" last. A
  fixed list favors whatever sits on top.
- **No overlapping options.** "Social media" beside "LinkedIn" splits one channel in two.
- **The same list every edition.** A renamed option breaks the year-over-year comparison.

## How do you turn the answers into numbers?

Tally both columns over the same base, total registrations, so they compare row by row. Each
column carries its own unknown row.

**Response rate = Registrations with a usable answer ÷ Total registrations**

**Self-reported share (channel) = Answers naming the channel ÷ Total registrations**

**Tracked share (channel) = Registrations with that tracked source ÷ Total registrations**

**Gap (channel) = Self-reported share − Tracked share, in percentage points**

**Cost per self-reported registration = Channel spend ÷ Answers naming the channel**

**Cost per tracked registration = Channel spend ÷ Registrations with that tracked source**

A usable answer is any option other than a blank or "I don't remember," with "Other" coded
into a channel by hand. The tracked source comes from your registration platform's source
report or from web analytics. The spend comes from the channel lines in your
[event marketing budget](https://eventiq.io/md/templates/event-marketing-budget).

## How do you read self-reported answers against tracked sources?

Read the gap, channel by channel, against a threshold you set before the cycle opens. Five
percentage points is a reasonable start. Replace it once you have two editions of your own gaps.
| Pattern | What it usually means | What to do |
| --- | --- | --- |
| Self-reported share above tracked share by more than the threshold | The channel creates demand that arrives through another door | Do not cut it on tracked cost alone. Give it its own links or codes next cycle |
| Tracked share above self-reported share by more than the threshold | The channel is the last step for demand created elsewhere | Keep it working. Do not credit it with discovery |
| The two shares within the threshold | Both sources agree on the size of the channel | Use either share. On a small channel, still compare the two cost figures |
| A large self-reported answer with no tracked equivalent | People-driven demand: colleagues, returning attendees, speakers | Serve it with a referral ask or a group offer, and stop looking for it in ad reports |
Where the columns disagree, report cost per registration as a range between the tracked and
the self-reported figure. A decision that holds at both ends of the range is safe to make. One
that flips inside the range needs a test.

## Example: 1,200 registrations at an annual conference

Take a conference with 1,200 registrations and a required source question that includes "I
don't remember." All figures are hypothetical. After "Other" is coded by hand, 48 registrations
have no usable answer, so the response rate is 1,152 ÷ 1,200 = 96.0%.
| Source | Self-reported | Self-reported share | Tracked | Tracked share | Gap (points) |
| --- | --- | --- | --- | --- | --- |
| Email from the organizer | 312 | 26.0% | 504 | 42.0% | −16.0 |
| Colleague or manager | 264 | 22.0% | none |  |  |
| Attended a previous year | 228 | 19.0% | none |  |  |
| Speaker or presenter | 96 | 8.0% | none |  |  |
| Association, chapter, or partner | 72 | 6.0% | 36 | 3.0% | +3.0 |
| Sponsor or exhibitor | 60 | 5.0% | 24 | 2.0% | +3.0 |
| LinkedIn or other social media | 72 | 6.0% | 144 | 12.0% | −6.0 |
| Search engine | 48 | 4.0% | 156 | 13.0% | −9.0 |
| No answer or don't remember | 48 | 4.0% | none |  |  |
| Direct or unknown | none |  | 336 | 28.0% |  |
| Total | 1,200 | 100.0% | 1,200 | 100.0% |  |
Three rows have no tracked value. Colleagues, previous attendance, and speakers account for
264 + 228 + 96 = 588 registrations, which is 49.0% of the room. Tracking shows only 336
registrations as direct or unknown, so at least 588 − 336 = 252 of those people arrived through
a tagged link and were credited to it.

Email shows where they went. Tracking credits email with 504 registrations and 312 people named
it, a gap of −16.0 points. At least 504 − 312 = 192 of the people who clicked an organizer
email named a different source. Search has the same shape: 156 tracked, 48 named.

Now put spend beside the four channels that carry spend in this example.
| Channel | Spend | Tracked registrations | Cost per tracked registration | Self-reported registrations | Cost per self-reported registration |
| --- | --- | --- | --- | --- | --- |
| Email | $6,552 | 504 | $13 | 312 | $21 |
| Social, paid and organic | $28,800 | 144 | $200 | 72 | $400 |
| Search, paid and organic | $9,360 | 156 | $60 | 48 | $195 |
| Partner promotion | $7,200 | 36 | $200 | 72 | $100 |
On the tracked column alone, social and partner promotion look identical at $200 per
registration. The self-reported column separates them. Partners were named by twice as many
people as clicked a partner link, which puts their cost between $100 and $200. Their gap is
only +3.0 points, inside a five-point threshold, because the channel is small: a small gap in
points can still double a cost figure. Social was named by half as many people as clicked a
social link, which puts its cost between $200 and $400.

Suppose the budget case set a ceiling of $200 per registration. Partner promotion sits at or
under the ceiling at both ends of its range, so it stays. Social sits at the ceiling at one end
and at twice the ceiling at the other, so the decision flips inside the range. Test it next
cycle before cutting it.

The 588 people-driven registrations need a different response, because no ad budget produced
them. The 228 who named a previous year are a check on the returning-attendee count you build
from records, covered on the [attendee retention](https://eventiq.io/md/blog/attendee-retention) page. The 264 who
named a colleague justify a referral ask in the confirmation email. The 96 who named a speaker
justify a tagged link per speaker next cycle.

## Which mistakes make the answers unusable?

**Adding the columns.** Tracked plus self-reported registrations counts the room twice.

**Leaving group orders undefined.** One coordinator who registers ten colleagues answers once.
Decide whether that answer is copied to all ten registrations or counted once, and apply the
same rule to every edition.

**Leaving the answer on the form.** Export the answer with the registration record and the
email address, and map it to a field on the CRM contact.

## What to do this quarter

- Add one source question to the registration form, worded around the event and the first
  time the registrant heard of it.
- Build the answer list from your budget lines, add the people-driven options and "I don't
  remember," and write down which budget line each option maps to.
- Set the gap threshold and the group-order rule before registration opens.
- Tally both columns over total registrations at close, and report cost per registration as a
  range wherever they disagree.
- Map the answer to a CRM contact field.

## Common questions

### Is self-reported attribution more accurate than tracked attribution?

Neither. They measure different things. Tracking records the last link a person used, and the
answer records what the person remembers as the reason. Use tracking for the mechanics of
conversion and the answers for discovery, and report both.

### What response rate is enough to trust the shares?

There is no published threshold worth borrowing. Report the response rate beside every share.
As a working rule, treat shares as directional when fewer than half of registrants gave a
usable answer.

### Can the answers be used to credit pipeline or revenue?

Only if the answer travels with the person into the CRM. Then you can count opportunities by
self-reported source, next to the campaign the CRM recorded. It is one field on a contact and
does not divide a deal across touches.

## Where EventIQ fits

EventIQ replaces nothing. It connects on top of the platforms you already run: event platforms
(Cvent, Zoom, Swapcard), CRM (Salesforce, HubSpot, GoHighLevel), and marketing (Google Ads, Meta
Ads, LinkedIn Ads, Mailchimp, Google Analytics). Platforms with an API outside that list are
connected on request.

EventIQ does not collect the answer to "How did you hear about us?" The question is a field on
your registration form, and the answers stay in the registration platform's export, where you
tally them. What EventIQ holds is the record around the answer. Registrations from Cvent carry
their date, ticket type, and price where the platform provides them. Marketing spend, entered
by your team or imported from a CSV, sits by event and channel. Attendees are matched to CRM
contacts by exact email, and each record keeps the result: matched, unmatched, or no email.

The ad platform connections bring campaign name and status only, so spend by channel is a
figure your team enters, and tracked registrations by source are read in your registration
platform or web analytics. The cost columns on this page are yours to compute, because the
answer is not in EventIQ. EventIQ does no multi-touch attribution, and the model picker in the
product does not change the calculation yet. The link from an event to revenue is one
Salesforce campaign relationship that a person confirms, which brings deals with their stage,
amount, and close date into the event view.

[Book a demo](https://eventiq.io/#early-access) to see marketing spend by event and channel next to registrations
by date and ticket type on a sample event, in a 20-minute demo.

---

HTML version: https://eventiq.io/blog/how-did-you-hear-about-us