A move-in coordinator asks the question at the end of a tour: “So — how did you hear about us?” The daughter pauses and says, “I think… I Googled you? Or maybe a friend mentioned you.”
That answer becomes a row in your CRM. It rolls up into a pie chart. That pie chart decides where next year’s marketing budget goes.
Here’s the problem: that daughter began her search four months ago, touched your community a dozen times across three devices and two siblings, and the very first thing she did — the thing that put you on her shortlist — was ask ChatGPT “best memory care near me for a parent with dementia.” She doesn’t remember that. She remembers the tour she just took.
“How did you hear about us?” doesn’t measure how families found you. It measures what they remember at the moment you ask — and in senior living, those are two completely different things.
See the first touch your form can’t: Check whether AI engines recommend your community — free → The AI-search touch that starts so many senior-living journeys leaves no click and no referrer, so it never reaches your form.
What “how did you hear about us?” actually is
It’s self-reported attribution — a single survey question, usually a dropdown, asked once at inquiry, tour, or move-in. Operators lean on it for an understandable reason: senior living lacks the clean click-trail of e-commerce, so a human asking a human feels like the most honest data you can get.
For a short, simple, single-person purchase, it’s a reasonable proxy. Senior living is none of those things — which is exactly why the question quietly misleads the operators who trust it most.
Why self-reported attribution breaks in senior living (5 reasons)
1. The 107-to-400-day memory gap
A senior-living decision unfolds over a 107-to-400-day cycle. By the time anyone fills out your form, the first touch — the one that caused the shortlist — is months in the past and long forgotten. Self-reported attribution doesn’t record the cause of a decision; it records the most recent or most memorable moment before someone asked.
2. The buyer isn’t the resident — and usually isn’t one person
An adult child (most often a daughter) does the research; siblings weigh in; the parent makes the final call. Your form asks one person, once, for a single answer about a multi-person, multi-touch journey. Whatever influenced the researcher gets collapsed into whatever the form-filler happens to say.
3. Recency bias over-credits the last touch
People attribute to what’s freshest in memory, not what set the decision in motion. The tour they took this morning, the friend who reminded them last week, “Google” as a catch-all — these win. The AI conversation that planted your name 90 days earlier loses. This is textbook recency bias, and a long buying cycle maximizes it.
4. Salient, nameable channels beat invisible ones
“A Place for Mom” and “a friend recommended you” are concrete, memorable, easy to say out loud. “I asked an AI and it suggested a few places” is vague and forgettable. So human and brand touches get over-credited, and digital first-touches get under-credited — not because they mattered less, but because they’re harder to name.
5. AI search is invisible by construction
When ChatGPT or Google’s AI Overview recommends your community, there is no click, no referrer, no UTM parameter. The family reads the answer, then navigates to you directly or via a branded search. So even a family who sincerely tries to report it has nothing concrete to point at — and the move-in gets logged as “direct,” “organic,” or “walk-in.”
What the form records vs. what actually happened
| What the form says | What likely happened |
|---|---|
| ”Google” / “online search” | An AI Overview or ChatGPT answer built the shortlist; a branded search just navigated to you later |
| ”A friend / word of mouth” | A friend confirmed a choice the family had already researched online for weeks |
| ”A Place for Mom” | The aggregator re-surfaced a community the family first discovered elsewhere — but now you owe a referral fee of 50–120% of first month’s rent (~$3,500–$12,000) |
| “Walk-in / just found you” | Months of invisible digital research that finally converted to a direct visit |
| ”I don’t remember” | The honest answer for most long-cycle journeys — and the one your pie chart can’t use |
The pattern is consistent and costly: aggregators and human referrals get over-credited; AI search and organic get under-credited. You keep paying referral fees for move-ins those channels didn’t truly originate, and you under-invest in the AI-visibility and organic channels that are quietly doing the work.
What to do instead: triangulate, don’t trust
The fix isn’t to throw out “how did you hear about us?” — it’s to stop treating one survey answer as truth. Use it as one signal among several in a closed-loop, multi-signal model: AI landing pages with call tracking, UTM + referrer capture, the HDYHAU answer, and CRM move-in records, matched into a directional model that’s honest about its confidence. (The full method.)
And if you only fix the form itself
- Ask more than once — at inquiry and at move-in — and watch how often the answer changes. That delta is the recency bias, measured.
- Make it multi-select. A real journey has several touches; a single-choice dropdown forces a lie.
- Add the option that’s missing: “Online search or AI assistant (ChatGPT, Google AI, Perplexity).” You can’t credit a channel your form never lists.
The bigger picture
Two blind spots compound into one expensive problem. You can’t see whether AI engines recommend your community (visibility), and you can’t prove which channel produced a move-in (attribution). Fixing both — measure the AI first touch, then attribute it to the dollar — is how senior-living marketing finally becomes accountable.
Start with the half you can check in 60 seconds: Is your community recommended by ChatGPT and Google AI? →