Sample Category Study · Sleep supplements
A sleep supplement brand, named in 5% of one leading AI assistant’s answers
400 simulated buyer conversations with two leading AI assistants, on questions that name no brand. Nature Made, the brand named most, led it for every buyer type, and neither assistant named the brand for a buyer shopping for a child.
Supracited AI engine
Each seat is one simulated conversation, inner rows with Claude, the outer row with ChatGPT.
Nature Made leads the brand most for someone 65 and over on a tight budget: 20% against 5%.
Nature Made leads the brand most for someone 45 to 64 comparing carefully: 15% against 0%.
Thorne leads the brand most for someone 45 to 64 comparing carefully: 17% against 5%.
Thorne in 0% of conversations, the brand in 0%. Neither was named for any buyer.
Pure Encapsulations leads the brand most for someone 45 to 64 comparing carefully: 15% against 5%.
Pure Encapsulations in 0% of conversations, the brand in 0%. Neither was named for any buyer.
Doctor’s Best leads the brand most for someone 65 and over on a tight budget: 10% against 5%.
Doctor’s Best in 0% of conversations, the brand in 0%. Neither was named for any buyer.
ZzzQuil leads the brand most for someone 45 to 64 buying for an older parent: 12% against 7%.
ZzzQuil in 0% of conversations, the brand in 0%. Neither was named for any buyer.
| Buyer | The brand | Nature Made | Both |
|---|---|---|---|
| 18 to 29, in a hurry | 7% | 15% | 3 of 60 |
| 65 and over, tight budget | 5% | 20% | 3 of 60 |
| 45 to 64, comparing carefully | 5% | 17% | 3 of 60 |
| 30 to 44, buying for a child | 0% | 2% | 0 of 60 |
| 45 to 64, for an older parent | 7% | 18% | 2 of 60 |
| All five buyers | 5% | 14% | 11 of 300 |
| Buyer | The brand | Nature Made | Both |
|---|---|---|---|
| 18 to 29, in a hurry | 0% | 10% | 0 of 20 |
| 65 and over, tight budget | 0% | 5% | 0 of 20 |
| 45 to 64, comparing carefully | 0% | 15% | 0 of 20 |
| 30 to 44, buying for a child | 0% | 5% | 0 of 20 |
| 45 to 64, for an older parent | 0% | 10% | 0 of 20 |
| All five buyers | 0% | 9% | 0 of 100 |
| Buyer | The brand | Thorne | Both |
|---|---|---|---|
| 18 to 29, in a hurry | 7% | 15% | 0 of 60 |
| 65 and over, tight budget | 5% | 3% | 0 of 60 |
| 45 to 64, comparing carefully | 5% | 17% | 1 of 60 |
| 30 to 44, buying for a child | 0% | 3% | 0 of 60 |
| 45 to 64, for an older parent | 7% | 18% | 2 of 60 |
| All five buyers | 5% | 11% | 3 of 300 |
| Buyer | The brand | Thorne | Both |
|---|---|---|---|
| 18 to 29, in a hurry | 0% | 0% | 0 of 20 |
| 65 and over, tight budget | 0% | 0% | 0 of 20 |
| 45 to 64, comparing carefully | 0% | 0% | 0 of 20 |
| 30 to 44, buying for a child | 0% | 0% | 0 of 20 |
| 45 to 64, for an older parent | 0% | 0% | 0 of 20 |
| All five buyers | 0% | 0% | 0 of 100 |
| Buyer | The brand | Pure Encapsulations | Both |
|---|---|---|---|
| 18 to 29, in a hurry | 7% | 12% | 2 of 60 |
| 65 and over, tight budget | 5% | 7% | 0 of 60 |
| 45 to 64, comparing carefully | 5% | 15% | 3 of 60 |
| 30 to 44, buying for a child | 0% | 2% | 0 of 60 |
| 45 to 64, for an older parent | 7% | 13% | 2 of 60 |
| All five buyers | 5% | 10% | 7 of 300 |
| Buyer | The brand | Pure Encapsulations | Both |
|---|---|---|---|
| 18 to 29, in a hurry | 0% | 0% | 0 of 20 |
| 65 and over, tight budget | 0% | 0% | 0 of 20 |
| 45 to 64, comparing carefully | 0% | 0% | 0 of 20 |
| 30 to 44, buying for a child | 0% | 0% | 0 of 20 |
| 45 to 64, for an older parent | 0% | 0% | 0 of 20 |
| All five buyers | 0% | 0% | 0 of 100 |
| Buyer | The brand | Doctor’s Best | Both |
|---|---|---|---|
| 18 to 29, in a hurry | 7% | 3% | 0 of 60 |
| 65 and over, tight budget | 5% | 10% | 0 of 60 |
| 45 to 64, comparing carefully | 5% | 5% | 0 of 60 |
| 30 to 44, buying for a child | 0% | 2% | 0 of 60 |
| 45 to 64, for an older parent | 7% | 5% | 0 of 60 |
| All five buyers | 5% | 5% | 0 of 300 |
| Buyer | The brand | Doctor’s Best | Both |
|---|---|---|---|
| 18 to 29, in a hurry | 0% | 0% | 0 of 20 |
| 65 and over, tight budget | 0% | 0% | 0 of 20 |
| 45 to 64, comparing carefully | 0% | 0% | 0 of 20 |
| 30 to 44, buying for a child | 0% | 0% | 0 of 20 |
| 45 to 64, for an older parent | 0% | 0% | 0 of 20 |
| All five buyers | 0% | 0% | 0 of 100 |
| Buyer | The brand | ZzzQuil | Both |
|---|---|---|---|
| 18 to 29, in a hurry | 7% | 2% | 0 of 60 |
| 65 and over, tight budget | 5% | 7% | 0 of 60 |
| 45 to 64, comparing carefully | 5% | 3% | 0 of 60 |
| 30 to 44, buying for a child | 0% | 2% | 0 of 60 |
| 45 to 64, for an older parent | 7% | 12% | 0 of 60 |
| All five buyers | 5% | 5% | 0 of 300 |
| Buyer | The brand | ZzzQuil | Both |
|---|---|---|---|
| 18 to 29, in a hurry | 0% | 0% | 0 of 20 |
| 65 and over, tight budget | 0% | 0% | 0 of 20 |
| 45 to 64, comparing carefully | 0% | 0% | 0 of 20 |
| 30 to 44, buying for a child | 0% | 0% | 0 of 20 |
| 45 to 64, for an older parent | 0% | 0% | 0 of 20 |
| All five buyers | 0% | 0% | 0 of 100 |
0%
The brand, with ChatGPT
Never named it in its answers.
27%
Named one of the five rivals, with Claude
Any of them, in any position.
93%
Of the brand’s mentions came beside a rival
It almost never came up alone.
Who answered: Claude, 300 conversations, and ChatGPT, 100, counted on their own · United States · recorded 17 October 2026, Pacific · 0 failed · 6,000+ data points measured by the Supracited AI engine.
Every conversation, buyer by buyer
Supracited AI engine
- 18 to 29, in a hurry7%0%15%10%15%0%12%0%3%0%2%0%
- 65 and over, tight budget5%0%20%5%3%0%7%0%10%0%7%0%
- 45 to 64, comparing carefully5%0%17%15%17%0%15%0%5%0%3%0%
- 30 to 44, buying for a child0%0%2%5%3%0%2%0%2%0%2%0%
- 45 to 64, for an older parent7%0%18%10%18%0%13%0%5%0%12%0%
The brand
5%
0%
Nature Made
14%
Nature Made
9%
Thorne
11%
Thorne
0%
Pure Encapsulations
10%
Pure Encapsulations
0%
Doctor’s Best
5%
Doctor’s Best
0%
ZzzQuil
5%
ZzzQuil
0%
Nature Made leads the brand most for someone 65 and over on a tight budget: 20% against 5%.
Nature Made leads the brand most for someone 45 to 64 comparing carefully: 15% against 0%.
Thorne leads the brand most for someone 45 to 64 comparing carefully: 17% against 5%.
Thorne in 0% of conversations, the brand in 0%. Neither was named for any buyer.
Pure Encapsulations leads the brand most for someone 45 to 64 comparing carefully: 15% against 5%.
Pure Encapsulations in 0% of conversations, the brand in 0%. Neither was named for any buyer.
Doctor’s Best leads the brand most for someone 65 and over on a tight budget: 10% against 5%.
Doctor’s Best in 0% of conversations, the brand in 0%. Neither was named for any buyer.
ZzzQuil leads the brand most for someone 45 to 64 buying for an older parent: 12% against 7%.
ZzzQuil in 0% of conversations, the brand in 0%. Neither was named for any buyer.
Quoted from a recorded AI answer. Not medical advice and not our recommendation.
Each card shows the brands an answer named, in the order it named them. We quote only sentences we have checked as plain fact, such as what a brand is or where it is sold, and none of these answers has one.
| Buyer | The brand, Claude | The brand, ChatGPT | Nature Made, Claude | Nature Made, ChatGPT | Thorne, Claude | Thorne, ChatGPT | Pure Encapsulations, Claude | Pure Encapsulations, ChatGPT | Doctor’s Best, Claude | Doctor’s Best, ChatGPT | ZzzQuil, Claude | ZzzQuil, ChatGPT |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 18 to 29, in a hurry | 7% | 0% | 15% | 10% | 15% | 0% | 12% | 0% | 3% | 0% | 2% | 0% |
| 65 and over, tight budget | 5% | 0% | 20% | 5% | 3% | 0% | 7% | 0% | 10% | 0% | 7% | 0% |
| 45 to 64, comparing carefully | 5% | 0% | 17% | 15% | 17% | 0% | 15% | 0% | 5% | 0% | 3% | 0% |
| 30 to 44, buying for a child | 0% | 0% | 2% | 5% | 3% | 0% | 2% | 0% | 2% | 0% | 2% | 0% |
| 45 to 64, for an older parent | 7% | 0% | 18% | 10% | 18% | 0% | 13% | 0% | 5% | 0% | 12% | 0% |
Four fixes, graded by evidence
Each grade names the evidence behind the technique in general, not proof that it will move this brand’s count; the re-check is what tests that.
Fix 1Controlled study
Add label facts and named review quotes to your key pages
First step: Pick the product page you most want recommended. Make sure its text, not only a label image, states each ingredient with its amount per serving, the form, the count and the price, as the label prints them, and add one quoted line from a named review about a label fact or about testing, publication credited, once your label reviewer has approved it. Add no wording about what the product does.
Fix 2Large correlational study
Refresh your key pages on a schedule
First step: List the 20 pages that matter most (best sellers and comparison pages). Refresh 10 of them now with a substantive change (a price, a label fact that changed, a comparison of label facts, a new review quote your label reviewer approved) and leave the other 10 as they are, so the day-90 re-check can tell whether the refresh helped. Update the visible last-updated date only where the content changed.
Fix 3Large correlational study (three separate studies converge)
Get onto the buying guides the AI already cites
First step: Open each guide page the private report lists and check whether it already covers the brand; where it does not, send its pitch to the author it names, or to the site’s contact page.
Fix 4Confirmed mechanism (direct HTTP observation)
Check whether AI crawlers visit your site at all
First step: Get the last 30 days of web server access logs from your host (on a hosted store without raw logs, ask the host or use your CDN’s crawler report) and count the AI crawlers by name.
We measure and source. We do not write, edit or suggest any statement about what a product does; that stays as your label states it.
The rest of the study
The brands named most, the sources, and the sections a Category Study adds to the fix list. Each opens in place.
The brands named mosteach assistant on its own, with a range on each count
Each assistant on its own, never added together. Range: the Wilson interval on each count, an indicative range rather than a 95 percent guarantee. Bars run from 0 to 20%; the tick marks the share.
With Claude 300 conversations
- Nature Made14%range 11 to 19%
- Thorne11%range 8 to 15%
- Pure Encapsulations10%range 7 to 14%
- Doctor’s Best5%range 3 to 8%
- ZzzQuil5%range 3 to 8%
- The brand (this study)5%range 3 to 8%
- NOW Foods3%range 2 to 6%
- Natrol3%range 2 to 6%
- Olly2%range 1 to 5%
- Nature’s Bounty1%range 1 to 3%
- Walgreens1%range 1 to 3%
With ChatGPT 100 conversations
- Nature Made9%range 5 to 16%
Named in none of its 100: Thorne, Pure Encapsulations, Doctor’s Best, ZzzQuil, NOW Foods, Natrol, Olly, Nature’s Bounty, Walgreens and the brand (this study).
Exploratory: 30 buyer scenarios. An industry framework (IAB, August 2026) counts a program with fewer than 50 buyer scenarios as exploratory rather than directional; read these ranges as a guide, not a precise estimate. Known limits.
Sourcesthe 15 sites cited most, across both assistants
Of the 284 domains the 400 answers cited, 1,260 times in all, 40% carry at least one flag, covering 44% of the citations. Every flag is a reason for a human look, not a finding about the site.
Narrow screen: scroll the table sideways for more columns.
| Domain | Answers citing it, both assistants | Registered | Popularity rank | Flag |
|---|---|---|---|---|
| nih.gov | 142 | 1997 | 50 | No flag; one withdrawn on a human read |
| yahoo.com | 44 | 1995 | 51 | No flag |
| sleepfoundation.org | 33 | 1996 | 1,548 | No flag |
| smartersupplementreviews.com | 33 | 2025 | Not in top 1M | Flagged registered 2025 or later |
| ubiehealth.com | 33 | 2022 | 83,237 | No flag |
| healthline.com | 30 | 2004 | 354 | No flag |
| healthychildren.org | 25 | 2002 | 2,311 | No flag |
| bioabsorbnutraceuticals.com | 21 | 2020 | Not in top 1M | Flagged off-topic homepage, low popularity |
| goodrx.com | 20 | 2011 | 4,687 | No flag |
| gmtoday.com | 18 | 1999 | 56,705 | No flag |
| takeroon.com | 17 | 2026 | Not in top 1M | Flagged off-topic homepage, low popularity; registered 2025 or later |
| vitalspage.com | 17 | 2026 | Not in top 1M | Flagged registered 2025 or later |
| harvard.edu | 16 | 1985 | 115 | No flag |
| seniorcitizenwellbeing.com | 16 | 2022 | Not in top 1M | Flagged off-topic homepage, low popularity |
| liveworksleep.com | 14 | 2025 | Not in top 1M | Flagged registered 2025 or later |
Popularity ranks: the Majestic Million (majestic.com), licensed under Creative Commons Attribution 3.0.
Content gap, outreach kit and page briefswhich rivals the cited buying guides name
How many of the buying guides the assistants cited most name each rival (9 guide pages read), and how often each assistant named it.
Narrow screen: scroll the table sideways for more columns.
| Brand | Guide pages naming it | With Claude, of 300 | With ChatGPT, of 100 |
|---|---|---|---|
| Nature Made | 3 of 9 | 14% (43 of 300) | 9% (9 of 100) |
| Thorne | 3 of 9 | 11% (34 of 300) | 0% (0 of 100) |
| Pure Encapsulations | 2 of 9 | 10% (29 of 300) | 0% (0 of 100) |
| Doctor’s Best | 0 of 9 | 5% (15 of 300) | 0% (0 of 100) |
| Natrol | 1 of 9 | 3% (10 of 300) | 0% (0 of 100) |
The brand’s own row, the four pitches and the page briefs name the brand’s pages or the guides that cover it, so they stay in the private report. No pitch carries wording about what a product does.
We measure and source. We do not write, edit or suggest any statement about what a product does; that stays as your label states it.
90-day roadmapthe fixes in order, then the re-check
Every fix in the order to start it, who does it, and the evidence grade: 11 to 32 hours over the 90 days where the research gives hours, with the re-check on or after 15 January 2027.
Narrow screen: scroll the table sideways for more columns.
| # | Item | Who | Evidence grade |
|---|---|---|---|
| 1 | Fix 3: get onto the buying guides the AI already cites (the four pitches in the private report) | Client | Large correlational study (three separate studies converge) |
| 2 | Fix 1: add label facts and named review quotes to your key pages (through the page briefs in the private report) | Client | Controlled study |
| 3 | Fix 2: refresh your key pages on a schedule (briefed pages can count toward the 10) | Client | Large correlational study |
| 4 | Fix 4: check whether AI crawlers visit your site at all | Us, guided | Confirmed mechanism (direct HTTP observation) |
| 5 | Re-check at day 90 | Us, guided | Measurement, not a technique |
In a client’s study the re-check asks the same buyers and buyer scenarios of the same assistants again and marks each count moved, held or within noise. An assistant that changes version in the meantime is labelled “version changed”, never credited to the fixes.
Method notehow to read the counts, every range and the 30 buyer scenarios
Each conversation is one simulated buyer asking one buyer scenario of one assistant, and each answer is one draw, which is why every count carries a range. Both assistants were asked with web search on, and the two are never added together. Location was not tested: every conversation ran from the United States.
Narrow screen: scroll the table sideways for more columns.
| Simulated buyer | With Claude, of 60 each | With ChatGPT, of 20 each |
|---|---|---|
| 18 to 29, in a hurry | 7%range 3 to 16% | 0%range 0 to 16% |
| 65 and over, tight budget | 5%range 2 to 14% | 0%range 0 to 16% |
| 45 to 64, comparing carefully | 5%range 2 to 14% | 0%range 0 to 16% |
| 30 to 44, buying for a child | 0%range 0 to 6% | 0%range 0 to 16% |
| 45 to 64, for an older parent | 7%range 3 to 16% | 0%range 0 to 16% |
| All buyers | 5%range 3 to 8% | 0%range 0 to 4% |
The 30 buyer scenarios, as asked
Hover, focus or tap a scenario to read the question as asked.
This is the shape of a Category Study.
This sample ran 400 simulated buyer conversations with two assistants. A Category Study runs 2,400+, with the day-90 Re-check included; a Brand Study stops at the graded fix list.