Working · 2025-12SecondaryGrowth opsAnalytics

'How did you hear about us?' self-reported attribution

Ask new signups or buyers where they heard about you, ideally in a free-text field, and use the answers next to software attribution to see channels that tracking misses.

Also called: HDYHAU, self-reported attribution, checkout survey

Claim and evidence

The claimthe average difference between HDYHAU and the installs reported by Meta on a daily basis is 95.

RevenueCat (consulting example) · revenuecat.com · 2025-12-16

What the numbers show

95

average daily difference between survey answers and Meta-reported installs for one subscription app (consultant example, via RevenueCat)

revenuecat.com
nearly 30%

of people leave the question blank when it is optional (Recast, vendor-reported)

getrecast.com

Why it works

Diagram, not data
Podcast mentionFriend'srecommendationPrivate shareAnswer to 'how did youhear?'Compare withtracked attribution
Podcast mentionFriend'srecommendationPrivate shareAnswer to 'how did you hear?'Compare with tracked attribution

Click tracking credits the last measurable touch, so podcasts, word of mouth and private sharing look like 'direct' or branded search. Asking people surfaces those influences, but memory is unreliable and only existing customers answer, so the data works as one input alongside experiments and interviews.

How to run it

  1. Add the question at signup, checkout or lead form
  2. Prefer free text and map answers to channels with keyword rules
  3. Compare weekly answer shares with platform-reported conversions
  4. Treat gaps as hypotheses to test with holdouts or lift tests
  5. Re-check after adding new channels, since answers follow your spend mix

Before / after

Before
Create your account
Email
Password
Sign up
Sign-up form with no source question
After
Create your account
Email
Password
How did you hear about us?
Sign up
Free-text question about the source

Generalised wireframe of the pattern, not a copy of any company's screen.

Ways to run it

Design options to test, not promised results.

  1. Option AFree text with keyword mapping

    Ask in an open field and map answers to channels with keyword rules.

  2. Option BCompare with ad platforms

    Set daily answers next to platform-reported installs to spot under-credited channels, as in RevenueCat's example.

  3. Option CHypotheses for lift tests

    Treat gaps between answers and tracking as ideas to test with holdouts, not as final truth.

Dated examples

  1. 2025

    RevenueCat (consulting example)

    A growth consultant compared daily HDYHAU answers for a subscription app with Meta-reported installs to spot under-attribution of social ads.

    The average daily gap between HDYHAU answers and Meta-reported installs was 95.

    “the average difference between HDYHAU and the installs reported by Meta on a daily basis is 95.”
    ✓ quote checked word-for-word on the page

    revenuecat.com · 2025-12-16

Does it still work?

Workingas of 2025-12

Widely used in 2025; practitioners warn it becomes less reliable as the channel mix grows. source

Risks

  • Recall is incomplete and biased toward memorable or recent touches
  • Only reaches people who converted, so it cannot reveal new channels
  • Answers drift with your spend mix
  • Optional fields get many blanks

Sources

  1. sparktoro.com · SparkToro blog · Rand Fishkin · 2023-09-20
    “Surveys like these can be useful. I’m not saying this is useless or to take it off your site.”
    ✓ quote checked word-for-word on the page
  2. getrecast.com · Recast · Shaifali Agrawal · 2023-05-31
    “Moreover, when you make the field optional, nearly 30% of people do not fill out the form.”
    ✓ quote checked word-for-word on the page
  3. revenuecat.com · RevenueCat blog · David Vargas · 2025-12-16
    “However, as apps scale and expand into additional channels, HDYHAU responses become increasingly unreliable as an attribution source”
    ✓ quote checked word-for-word on the page

Change log

  • 2026-11-20 — harvest v0.1 (pass2-expansion-ops)