First, look at what the score is counting
The headline was 17 mentions across 100 generated ChatGPT questions for one Canadian retirement operator. Open the breakdown and you find that 13 of those mentions came from 13 questions that already named the operator. The other 87 questions produced four mentions. That means 76.5% of the mentions came from just 13% of the questions.
So I’d show the two rates separately: 100% for questions naming the operator, and about 4.6% for questions without its name. One tells you how AI talks about a known option. The other tells you whether it brings that option into the conversation. Your strategy needs to care about both.
This is one generated-question sample, checked on 3 November 2026. Its collection date wasn’t available. It doesn’t count real families, tell us whether each mention was positive, or show move-ins. It does give us a clear reason to open the questions behind the score.
| Question group | Questions | Mentions | Mention rate |
|---|---|---|---|
| 13 | 13 | 100% | |
| 87 | 4 | 4.6% | |
| 100 | 17 | 17% |
Why I care about the questions before your name
Search Influence surveyed 300 families researching senior living or who had searched in the previous year. Its October 2026 summary says 57% seriously considered only two or three communities; 77% considered four or fewer. Those are the publisher’s survey results, rather than a national admissions benchmark.
Imagine a family building that shortlist. If they haven’t heard of your community, they won’t start by asking for your name. They might ask about care near a relative, fees, or options for a couple. I’d want to know whether your business enters those answers when it’s a suitable option.
The survey and our ChatGPT sample cover different groups and places, so we can’t turn them into a lost-resident estimate. The strategic point is simpler: check whether you’re being introduced while the family still has options to compare.
Find the decisions where you’re missing
In this sample, five topics produced no operator mentions: costs, choosing a community, transitions, memory care, and respite. Together they covered 45 questions. That gives us a much better research brief than “improve awareness.” Each topic helps a family make a different decision.
The four discovery mentions came from amenities, independent living, and assisted living. I’d look at those answers too. What made the operator relevant there? Were the right services described? Which sources did the answer use? The overall 17% can’t tell you that.
Before recommending new content, I’d check whether the operator actually belonged in the missing answers. Service fit, location, wording, source quality, and variation between answers can all matter. A zero in a small test is a reason to investigate, not a diagnosis on its own.
| Topic without mentions | Sampled questions | Operator mentions |
|---|---|---|
| 17 | 0 | |
| 12 | 0 | |
| 8 | 0 | |
| 6 | 0 | |
| 2 | 0 | |
| 45 | 0 |
Start with costs if families can’t get a clear answer
The same family survey names pricing transparency as the most important comparison factor and the hardest information to find online. It says 53% use AI to compare costs, care levels, or amenities, and 46% use it to narrow a shortlist. That puts a practical question on the table: can families understand your fees before they call?
The sample’s 17 cost questions produced no mentions. I’d review the actual questions, then look at what your pages explain: accommodation, included care, likely extras, assessment, and how to get a current quote. You may not be able to publish one final price for everyone. You can still explain how the price works.
This doesn’t prove a missing fee page caused the result. It tells us that costs are important enough to check separately. I’d rather give your team one specific pricing question to answer well than a vague instruction to publish more AI content.
Check brand and discovery inside each topic too
There’s another trap in the data. The brand-services topic has 13 questions and 11 mentions. Those aren’t the same 13 questions as the brand-name group. Within that topic, 11 questions named the operator and all 11 mentioned it. The other two didn’t name it and produced no mentions.
Assisted living shows the same issue. It had three mentions across 23 questions. Two came from the two questions naming the operator; one came from the other 21. If I only showed the topic total, you’d miss what was driving it.
I’d keep two labels on each tracked question: what decision it’s about, and whether it names your business. Then you can ask a useful question about a service: are we being introduced here, or only discussed when someone already knows us?
Don’t turn question groups into a customer funnel
The report also groups questions by intent. Research produced two mentions from 24 questions, comparison ten from 49, and buying five from 27. That works out to about 8.3%, 20.4%, and 18.5%.
Those aren’t the same people moving from research to purchase. They’re separate generated questions. We also don’t have the brand-name breakdown within each intent group. So the higher comparison rate doesn’t prove families become more likely to choose this operator as they progress.
Use these groups to check the answers available at different decisions. To understand how people move towards a visit or move-in, follow actual inquiries in your admissions records. Keep the two views connected, with their labels clear.
Use other providers’ results to ask better questions
Search Influence’s Liberty case reports 98.7% average presence for questions about a named community’s services, and 90.5% for a city-based assisted-living question among communities offering memory care. Those are the agency’s reported results. It hasn’t published the full underlying set of answers.
I wouldn’t put those figures beside our operator’s 4.6% and call it a ranking. The questions, locations, eligible services, dates, engines, and groups counted differ. Ask what was measured before deciding what another provider’s number means for you.
What I would borrow is the focus on a specific question and a suitable community. Choose the decisions you want to compete for, keep the test consistent, and compare your own results over time. That makes a case study useful without turning it into your forecast.
Now open the sources behind the answers
Being mentioned and being cited are different. AI can name your operator while a directory explains the service. It can also cite your page without recommending your community. Searchable separates brand mentions from the source links so you can inspect both.
I’d check which pages are shaping the family’s understanding of costs, care, and quality. Your own community page may explain the current service; an official record may support a rating; a comparison page may frame the alternatives. These pages have different jobs.
Our sample gives us mention counts, not a complete source map, so I can’t name which rival or directory owns the missing answers. That would be the next piece of research. Check where you’re absent, where the facts are wrong, and where someone else is explaining your offer.
Here’s how I’d use this with your team
Pick a community and pull up the questions behind its score. Split brand from discovery, then break discovery down by the decisions families need to make. Start with the questions where your community is a fit and the answer would change whether a family considers you.
Save the answers and their sources, rather than the score alone. Keep the topic, location, engine, date, and brand-name label with each one. Our 100-question snapshot shows why: change the mix and you can change the headline without improving discovery.
Then choose an action your team can check. Clarify a fee explanation, correct a community fact, or answer a comparison question. Revisit the same question set after the change has had time to be found. The family survey and provider cases give context; your own answers tell you where to work.
I can help you build that question set in Searchable and turn it into a short action plan. Once you can see the discovery gap, the next job is measurement: make sure a better-looking chart means the answers actually improved.
Sources and details
Sources reviewed: November 3, 2026. I work at Searchable. The examples and recommendations are mine; reported results are credited to their sources.
1
Searchable