How a number gets printed

One answer is one draw from how an assistant answered that day. It says nothing durable about the brand. A brand named in half of 20 answers to the same question still carries about ±20 points of uncertainty.

Census tables print a cluster-bootstrap range as the headline figure, since it allows for the same question being asked again, not just for more answers, which widens it. A narrower Wilson range sits beside it in each full table, labelled: Wilson assumes every answer is an independent draw, which is the more optimistic of the two, so we do not lead with it.

Two 2026 papers argue for bootstrap intervals and adaptive sampling on exactly this problem. We print the interval on every table on this site.

At the six to eight buyer scenarios our published censuses ask today, that range is exploratory: a guide to where the share would likely land, not a 95 percent guarantee. A program of 50 or more buyer scenarios, our own included once it runs, is where the stronger reading holds. See known limits.

A keyword count says a name appeared, not how. In the pain relief census, store brands were mentioned 39 times. We read each mention in its sentence and coded it either as a product the answer told the reader to buy or as a qualifier: a place to buy, or one example in a list. Of those, 97% were qualifiers (38 of 39) and one was a product. Two mentions could be read either way; the one closer to naming a product was counted as a product, against our own finding.

Before a domain an assistant cites counts toward anything on this site, it gets three checks: its popularity rank in the Majestic Million (majestic.com, licensed under Creative Commons Attribution 3.0), whether its homepage actually matches the category it is being cited for, and whether it shows signs of a recent change of owner (a domain that changed hands and now hosts something unrelated to its history). A domain that fails is marked flagged in words, never by colour. A domain that passes is marked verified. A domain we have not yet checked is marked unranked.

Narrow screen: scroll the table sideways for more columns.

Roughly one in seven of the citations one leading assistant gave in our research got flagged for a human look this way; on a human read, about a third of flags turned out to be on-topic after all. That is why a flag is a reason for a human look, never a finding about the site. The precision figure stays out of client reports until flags are hand-labelled at scale, which is also why this page, not a client report, is where the caveat lives.

Every number on this site is affected by more than the sampling error its range covers. The major ones:

Version changes. An assistant updates without notice. A published number reflects the version live on its collection date; a later update is not re-measured until the next pass.

Personalization and location. Answers come from one signed-in account, one region and one device. Memory, saved instructions, a different location or a different device can change what an assistant says.

Scenario selection. A census asks the buyer scenarios in its own library, not a full market survey. Which ones made the list is a judgment call, named on each category’s own page.

AI-assisted coding. Some of the judgment behind these numbers, such as which mentions count or which cited sites get flagged, is made by AI rather than a person, and does not yet carry a second human check.

A range is only useful if it holds the true share as often as it says it does. For each published study we simulated the study a few thousand times at its own depth: its buyer scenarios, the answers to each, and how alike those answers are (the ICC), measured from its own data. Coverage is how often the printed range held the true share; a range that kept a 95 percent promise would score 95%.

Narrow screen: scroll the table sideways for more columns.

Every study here carries the exploratory label, so its ranges read as a guide. In the censuses, the lowest scores belong to names that came up in only one buyer scenario. The sample study prints Wilson ranges, which treat every answer as a separate draw, so they run narrow. A range printed without the exploratory label must score 94% or more here, or the site does not build.

Every mention rate on this site comes with the count behind it and a date, on a page that says whether one assistant answered or a second is being compared, which is the discipline this page exists to enforce. That discipline rules out the shortcuts that make a number easy to sell before it is earned: a ranking promised in advance, a score with no count behind it, or a consensus that was manufactured (bought reviews, hidden crawler text, incentivised posts) rather than measured. A number that cannot carry its assistant and its count does not get printed, on this page or anywhere else on the site.

How an answer is collected: each answer comes from asking one buyer question, in a fresh session, to one AI assistant with its own web search turned on and outside its consumer app, through the Supracited AI engine. We ask each question several times and keep every answer exactly as it came back, never edited afterward.

The table below says where each kind of detail lives: on this site for anyone to read, inside a client’s own report for the study it is for, or nowhere outside the Supracited AI engine itself. We publish the tasting; the recipe stays ours.

Narrow screen: scroll the table sideways for more columns.

robots.txt here allows the major search and AI crawlers, among them Googlebot, Bingbot, OAI-SearchBot, GPTBot and Google-Extended. Google-Extended specifically controls whether Google’s AI apps can train on and ground their answers in this content. It does not control whether this site shows up in Google Search or in AI Overviews, which Googlebot governs on its own. We leave it allowed anyway: a census that cannot be read is not a census, and the trade-off is that anything readable here can also end up reused by AI companies we have no agreement with.

Mention rule. A brand counts as named when one of its listed names appears in the body of the answer, matched as a whole word and regardless of case. The answer’s list of sources and every link are removed first, so a link to a brand’s own site is not a mention. A brand that is not on the tracked list cannot be counted.

Denominator. Each census table prints its denominator: the answers it counts, the questions and how many times each was asked. In a study, every brand shares one denominator, the answers to buyer scenarios that name no tracked brand, so a scenario that names a brand never lifts that brand’s rate.

Ranges. Census tables lead with a two-stage cluster bootstrap (questions drawn with replacement, then answers within each question) and show a Wilson score interval on the pooled count beside it. A study prints the span of the two, which is never narrower than either.

Position. The order in which tracked names first appear in the answer body. “In the first three” means among the first three tracked names.

Failed answers. Timeouts, errors, empty answers and block pages are counted in the sampling record and left out of every rate.

Sources. Taken from the assistant’s own citation record where it returns one, and otherwise from the sources the answer lists, which is the assistant’s own account of where it looked.

Hand coding. Where a count needs judgment, such as whether a store-brand mention names a product to buy, each mention is read in its sentence, and a case that could go either way is coded against our own finding.

What stays private. The full question library and its phrasings, how simulated buyers are written, how questions are assigned across buyers and assistants, the instructions each assistant is given, flag thresholds and the fix playbook. A client’s own analyst may inspect the setup of their study under a non-disclosure agreement.