Home › Index › Health & Wellness
Who AI recommends in Health & Wellness
The commercial questions of the sector, run across the ten measured engines, in the same window and by the same method. Every percentage comes out with the number of checks that produced it.
1.810
checks in the period
6
brands measured
10
engines
30
days of window
Presence ranking
Percentage of answers in which the brand was cited, over its total checks in the period.
| # | Brand | Presence | Denominator | Engines where it appears | |
|---|---|---|---|---|---|
| 1 | Clinic chain A | 62.9% | in 340 checks | GoogleAIOBingChatGPTGeminiPerplexity | |
| 2 | Hospital group B | 49.4% | in 340 checks | GoogleAIOBingChatGPTGeminiPerplexity | |
| 3 | Clinic C | 40.3% | in 300 checks | GoogleAIOBingChatGPTGeminiPerplexity | |
| 4 | Network D | 30.3% | in 290 checks | GoogleAIOBingChatGPTGeminiPerplexity | |
| 5 | Medical centre E | 16.8% | in 280 checks | GoogleAIOBingChatGPTGeminiPerplexity | |
| 6 | Clinic F | 11.9% | in 260 checks | GoogleAIOBingChatGPTGeminiPerplexity |
A brand with fewer than 30 checks in the period does not enter the ranking: with a small sample, the percentage swings by chance.
Who is cited instead
The domains the answers used as a source when they talked about this sector. Not all of them are competitors: some are press, associations and marketplaces, and knowing that changes the plan.
A domain with fewer than 5 citations does not make the list.
Methodology
- The questions. The sector commercial questions, the same for every brand, published alongside the edition.
- The engines. Google and AI Overviews weekly; ChatGPT, Gemini, Perplexity and Bing organic every 15 days.
- The window. 30 consecutive days, the same for every brand. A brand that entered mid-period is not compared with one that was there the whole time.
- What does not count. A failed query is not a check: counting it inflates the denominator and drives the percentage down for no real reason.
- The floors. 200 checks and 5 brands for a sector to qualify; 30 checks for a brand to qualify; 5 citations for a domain to qualify.
Models are probabilistic. Every run records what happened in that query, with the question, the engine and the window stored. The index does not promise that the same question will return the same answer tomorrow, and that is why it is published on a schedule rather than once.