Skip to content

The measurement, and the evidence under every number.

A Category Audit or a Diagnostic arrives in an account, not as a slide. Open any count and you see the answer the engine gave and the pages it cited. Every view below is open on a fictional sample company before you buy anything.

Counts before points. Every rate carries its base and a 95 percent interval.

Nothing in the account is a single composite score. A question is named or not named by one engine on one run, and the account keeps mentions, recommendations, position and citations apart rather than folding them together. Where two runs cannot be separated, the account says so instead of drawing a line between them.

Data table: View, What it answers, What is in it
ViewWhat it answersWhat is in it
Results and receiptsFor each buyer question, did the engine name us, and what did it actually say?Every count opens the observed answer text and the URLs that answer cited.
HistoryHas anything moved since the last scheduled run?One point per scheduled run, each with its count, its base and a 95 percent interval. A move is printed only where two intervals separate.
Named insteadIf the engine did not name us, who did it name?Per question and per engine, the companies named in our place, with checked and not found kept apart.
Source queueWhich pages do the engines read in our category, and what are we doing about each one?Every cited page with how often it was cited, which engines cited it and the state of our work on it.
ProofDid the change we shipped cause the movement, or did the category move anyway?Controlled lift: target questions against control questions from the same bank, four counts, and one of three fixed verdicts.
BattlecardsHow do we read against one named competitor, engine by engine?Quoted snippets from the observed answers with the receipt behind each one.
AlertsTell me when something moves, without my logging in.Movements between two consecutive comparable scheduled runs, sent to email, Slack or Teams, each with its receipts link.
Share linksHow do I show this to someone who has no account?A read only page for one view, revocable, asking search engines not to index it.
Each view is open on the public sample. The links are in the list below.

Kalvenor Systems is a fictional company. Everything in the sample account is SAMPLE DATA and is there so you can see the views before you buy.

The sample account belongs to Kalvenor Systems, a fictional field service management software company. It is SAMPLE DATA. It is there because the views are the product, and reading them is the only honest way to judge them.

History. Scheduled runs in May, June and August 2026 on the same ten questions and five engines: 20 of 50 answers named in May, 3 of 50 in June, 7 of 50 in August, each with its own 95 percent interval. The three July Diagnostic runs sit beside them on their own series.
Source queue. 405 cited pages, of which 1 is placed, 1 pitched and 2 declined. The rest are not started. The queue is a list of pages the engines actually cited, not a list of pages somebody would like to be cited.
Proof. Two experiments, and the second is the one worth reading. A comparison page for scheduling and offline questions moved the target questions from 0 of 25 answers to 12 of 23 while the control questions stayed at 3 of 25, a controlled lift of 52.2 points with a 95 percent interval of 21.7 to 79.2: Lift shown. A second change, maintenance contract detail on the dispatch page, reads No gain observed, and stays published beside the first.
Alerts. Three movements, each carrying the receipts behind it.

Every figure in this section was read from the served sample on 21 September 2026 and is fictional.

The question set, the citation logs, the observed answers and the reports are yours. They are returned in an open, machine readable format within five business days of a request, at no charge, whenever an engagement ends and for any reason.

A one pager for a CMO is generated from the same numbers, so the summary and the evidence can never drift apart. A share link opens one view for somebody with no account, and is revocable.

Read the account with your own tools: a read only API, webhooks, and a connector for Claude and ChatGPT.

Read your measurements where you already work. Every integration is read only: it writes nothing, changes nothing and starts no measurement.
Data table: Integration, What it does
IntegrationWhat it does
SmitheryThe connector, published and verified on the Smithery marketplace, with all fifteen tools listed.
Claude and ChatGPTAdd the connector to an assistant and ask about your own measurements in plain language.
The public sampleThe same fifteen tools over a fictional data set named Kalvenor, with no sign in. The endpoint is app.broadcastwell.com/mcp/demo/kalvenor.
The read only APIOpenAPI 3.1, with account keys you create and revoke yourself.
Only integrations that are live are listed here.

Nothing here needs an account to try. The public sample reads a fictional data set named Kalvenor with no sign in at all.

What is in the account after a Category Audit or a Diagnostic?

The results, and under every count the observed answer text and the URLs that answer cited. From there: who was named instead of you, the pages the engines cited in your category, the fixes drawn from the findings, and a one pager you can hand to a CMO. Counts are shown before points, and every rate carries its base and a 95 percent interval.

Can I see it before I buy?

Yes. The whole account is open on a fictional company, Kalvenor Systems, at app.broadcastwell.com/sample. It is SAMPLE DATA, labelled on every view, and no account or email is needed to read it.

How often do the numbers change?

Re-measurement is monthly. Each scheduled run is its own point in History with its own base and interval, and a move is printed only where two intervals separate. Runs are never averaged together and adaptive runs are kept out of the headline figures.

What does proof mean here?

Controlled lift. Target questions and control questions are drawn from the same bank, assignment is locked before the first follow up run and the lock time is printed, and the result is one of three verdicts: Lift shown, Gain on target questions not separable from drift, or No gain observed. Repeat runs of one question are correlated, so the true uncertainty is wider than the interval shown.

Can my own AI read the account?

Yes. There is a connector for Claude and ChatGPT with fifteen read only tools, and a read only API with an OpenAPI description. Both are covered on the developers page.

Is anything guaranteed?

No inclusion, no ranking, no revenue. No engine sells placement and nobody can promise it. What is promised is the measurement, the receipts under it, and the terms written on the terms page.

Start with a measured baseline

See which buyer questions leave you out, the sources visible in those answers, and the three changes worth testing first. The Category Audit covers ten questions; the Diagnostic extends to 35 questions and 525 scheduled observed answers, adaptive runs extra. AI answers vary, so repeated observations and disagreements are reported.

Start with the $490 Category Audit

Results appear in your account; a sign-in link arrives by email. Credits in full against the Diagnostic within 30 days.

Get the Diagnostic, $990

Results appear in your account; a sign-in link arrives by email. Credits in full against the program.

Every figure comes from the published method, v1.1.