Most senior-living communities can tell you where their leads come from. Far fewer can tell you where their move-ins come from — and those are the only ones that pay rent. Tracking move-ins (not leads) by source is what lets you move budget toward the channels that actually fill beds.
The quick version: define your channels (make AI search its own) → tag every entry point with UTMs, call tracking, and landing pages → capture the source at inquiry and store it on the lead → connect your CRM so the move-in is the endpoint → match inquiry to move-in with a confidence level → report cost per move-in by channel.
Step 1 — Define your channels (including the invisible one)
List the real ways families reach you, and treat each as a distinct channel:
- AI search — ChatGPT, Perplexity, Google AI Overviews (make this its own channel; most CRMs hide it inside “organic” or “direct”)
- Organic search and Google Business Profile
- Paid — search and social ads
- Referral aggregators — A Place for Mom, Caring.com, Seniorly
- Professional referrals — hospitals, discharge planners, social workers
- Word of mouth / events / direct / walk-in
The single most common mistake is not giving AI search its own bucket — so the fastest-growing top-of-funnel channel stays invisible from day one.
Step 2 — Tag every entry point
You can only attribute what you instrument. Put a fingerprint on every way in:
- UTMs on every link you control — ads, email, social, your GBP website link, every directory listing.
- Call tracking (DNI — dynamic number insertion): a unique phone number per channel/landing page. A large share of senior-living inquiries arrive by phone, and an untracked call is an unattributed lead.
- Dedicated AI/organic landing pages: visitors who arrive without a UTM — which is exactly what AI-search and organic referrals do — still land somewhere you can identify.
Step 3 — Capture the source at inquiry, and write it onto the lead
This is the step almost everyone skips, and it’s the one that matters most. The channel data exists only at first contact. By move-in — often 107 to 400 days later — it’s gone, and no one remembers.
So: at the moment of inquiry, stamp the source onto the lead record in your CRM. If your web forms and call tracking don’t push source data into the lead automatically, fix that before anything else. Attribution that tries to reconstruct the source months later is just guessing.
Step 4 — Use “how did you hear about us?” as a backup, not the truth
Keep the question — but treat it as a corroborating signal, not your system of record. Make it multi-select, ask it at inquiry and again at move-in, and add an explicit “online search / AI assistant (ChatGPT, Google AI)” option. (Why it’s unreliable on its own.)
Step 5 — Connect your CRM so the move-in is the endpoint
Attribution that stops at the lead measures the wrong thing. Pull prospect → stage → move-in (and the move-in’s dollar value / resident LTV) from your CRM — WelcomeHome, Aline, Sherpa/Enquire, or even a CSV export. The move-in is the event that pays; it has to be the endpoint of every report.
Step 6 — Match inquiry to move-in, and attribute with confidence
Link each move-in back to its originating inquiry — by email, phone, or time window — and credit the channel captured back in Step 3. Carry a confidence level on each match, so you know how much to trust each attribution. This is the moment the loop closes: a channel finally gets credit for a resident, not just a lead.
Step 7 — Report cost per move-in by channel
Now the number ownership actually wants is computable: cost per move-in (CPMI) = channel spend ÷ move-ins that channel produced. Rank your channels by CPMI and resident LTV, and shift budget toward what fills beds. (The full attribution guide.)
The gap this playbook runs into — and how to close it
Follow every step above and you’ll attribute the channels you control beautifully. But the fastest-growing first touch — AI search — leaves no UTM and no referrer. When ChatGPT or Google’s AI Overview recommends your community, the family navigates to you directly weeks later, so even a flawless setup logs that move-in as “direct,” “organic,” or “walk-in.” Your tracking is perfect and the channel is still invisible.
Closing that last gap takes two moves your CRM can’t make on its own:
- Measure whether AI even recommends you — because you can’t attribute a channel you’re not in. Check your AI visibility free →
- Give AI/organic arrivals a tracked path (Step 2’s landing pages + call tracking), then match those inquiries to CRM move-ins — producing a directional credit for the AI-sourced move-ins you’re currently logging as “direct.”
Do it with a spreadsheet, or a platform
You can run this with a spreadsheet: a tab per channel, UTM and call-tracking exports, a monthly reconciliation against your CRM’s move-in list. It works at low volume and it’s a great way to learn where your blind spots are.
It breaks down when matching gets manual across hundreds of inquiries and a 100–400-day lag, and when you want the AI-source credit that requires landing-page + CRM matching with confidence scoring. That’s the point where a purpose-built attribution layer earns its keep — but the standard is identical either way: closed-loop, multi-signal, ending at the move-in, honest about confidence.