How we measure AI visibility

Why we publish the method

We publish this because a score whose method is secret is a marketing graphic. It is also the page a buyer’s diligence asks for, and we would rather it already existed.

1. 01

   Ask

   What’s the best invoicing tool for small agencies?

   1 API call45 s timeout1 retry

   Your question, word for word, sent to the engine we hold a key for.

2. 02

   Engine answers

   Perplexity · live

   Fernhub is the usual pick 1.

   Kalye is a strong alternative 2,

   and Tallybook suits freelancers 3.

   A failed call is stored as a failure and left out, never as “not named”.

   The full answer and every source it cites are stored.

3. 03

   Names and citations found

   • ✓Named: KalyeWhole word, links blanked first
   • #2nd of 3 brands namedPosition score 0.75
   • ↗1 of 3 sources yourskalye.io, counted apart from the name
   • ~Tone readShown, not scored

   Rivals are found the same way, so you see who was named instead.

4. 04

   Scored

   75

   • Named, 50%50
   • Cited, 30%10
   • Rank, 20%15

   This answer alone scores 75. A run averages every answer it analysed.

Perplexity, and only Perplexity, todaySame questions, same arithmetic, every check.

The three inputs

Method / Visibility scoreWeight

Mention rateAnswers naming the brand over answers analysed50%

Citation shareCited sources on your domain over all cited sources30%

Position componentWhere you come among the brands named20%

How the position component works

Being the first brand named is worth far more than being the fourth, so position is scored on a scale rather than counted:

The scores are summed and divided by every answer analysed, not only the answers that named you, so one lucky first place cannot carry the score.

Average position is reported separately and is averaged only over answers where you were named. When you were never named it is reported as none, not as zero — an undefined value, honestly labelled.

AnswerPerplexity

What’s the best invoicing tool for small agencies?A

For small agencies, FernhubB is the usual pick for recurring retainers 1. KalyeC is a strong alternative with simpler pricingD 2, and Tallybook suits solo freelancers 3.

1. 1fernhub.com
2. 2kalye.ioEYours
3. 3reviewsite.example

This answer alone75(1 × 0.5 + 1/3 × 0.3 + 0.75 × 0.2) × 100

1. AThe question, word for wordAsked the same way every week and stored with the result, so the number keeps its meaning.
2. BA rival, found the same wayCompetitors are matched as whole words too. That is how you see who was named instead.
3. CNamed, and whereA whole-word match outside any link counts as named. Kalye is the 2nd brand named: 0.75.
4. DTone: shown, not scoredThe sentence around the name is read as positive, neutral or negative, and kept out of the score.
5. EA link is a citationEvery source domain is recorded. kalye.io is the brand’s own, so 1 of these 3 sources is yours.

Worked example one: our own zero

From our September 2026 audit: eight questions, one engine, eight answers analysed.

Bungad / Own auditPerplexity

Was Bungad named?Across every answer in the auditNot named

Did any cited source point at bungad.com?Across every cited sourceNone

• Mention rate0 ÷ 8 = 0
• Citation share0 ÷ 90 = 0
• Position component0 ÷ 8 = 0
• Score(0 + 0 + 0) × 100 = 0

All three inputs were zero, so there is nothing to round and nothing to argue about. Zero is the floor, and it is where we started.

Worked example two: how a real score is built

Illustrative numbers, not a client’s, so you can follow the arithmetic. Eight questions; the brand is named in three of them, at positions 2, 1 and 4; four of the 60 cited sources are on its own domain.

Notice what that says. A brand named in nearly four answers in ten, once in first place, still scores 26 — because it is almost never the cited source. Citation share is where that brand leaves points on the table, and it is the input you control fastest.

Visibility score

26Occasionally surfaced

0100

• Named · mention rate3 ÷ 8 = 0.375 × 5018.75
• Cited · citation share4 ÷ 60 = 0.0667 × 302
• Rank · position component2 ÷ 8 = 0.25 × 205

Score = round(25.75) = 26, clamped to 0–100. Position scores: 1st 1.0, 2nd 0.75, 3rd 0.5, 4th 0.25, 5th or later 0, averaged over all eight answers.

What the bands mean

The honesty rules built into the run

Enforced in the code that produces the number, not in a policy document. Each one, broken, would quietly flatter us.

Method / Run rulesIn code

An engine errorCounted in its own column, never as not namedExcluded

A run with any failureMarked incomplete and stored separatelyFlagged

• An engine error is never “not named”. Failed and empty answers are counted in their own column and never reach the scoring function. Otherwise the score would measure our own timeouts.
• A run with any failure is marked incomplete and stored under a separate filename, so it can never become the baseline for a trend.
• Records are written atomically and never overwrite an existing file. A published number that can be silently edited is not evidence.
• The question list is stored verbatim in every record. Changing a prompt changes what the number means, so the prompt set travels with the result.
• Nothing resembling an API key can enter a stored record. The records are meant to be publishable.
• Each run has a schema version, so old records stay readable when the shape changes.
• Each question gets a 45-second timeout and one retry before it is recorded as a failure.

Run / Honesty rulesEvery run

An engine errorCounted in its own column, never as “not named”Excluded

A run with any failureMarked incomplete, stored under a separate filenameIncomplete

Each question45-second timeout and one retry45 s

The question listStored verbatim in every recordKept

Non-determinism, measured

Engines do not return the same answer twice, so we ran our own audit twice, about a minute apart, and kept both records.

The conclusion we draw, and publish: headline metrics and brand tallies reproduce; per-domain citation counts do not. Treat a single run’s per-domain figures as estimates. Treat a single run of any kind as an anecdote — the signal is the trend across months on a fixed question set.

How to run this yourself

Nothing here is proprietary. If you would rather not pay anyone, this is the method.

The seven steps

1. Fix a question set and write it down. The questions a buyer actually asks before choosing someone like you, in their phrasing. At least two where price is the question and two you expect to lose. Store it verbatim — it is the denominator of every number that follows.
2. Name your competitors explicitly, with their domains. Without that list you learn whether you were named, but not who beat you, which is the more actionable half.
3. Ask every question to every engine you can reach, through an API rather than a chat window, so the run is reproducible and free of personalisation. Set a timeout and allow a small number of retries.
4. Record failures separately and never as absences. Count errors and empty responses in their own column, keep them out of the metrics, and mark a run containing one as incomplete so it cannot become a baseline.
5. Parse each answer for three things: whether your brand is named, its position among the brands named, and every cited source domain with a flag for whether you own it.
6. Compute the three metrics and the weighted score using the arithmetic above.
7. Store the whole run and repeat monthly, written atomically and never overwritten, comparing only against your own previous runs on the same question set.

Our own run is a script in our repository rather than a spreadsheet, and the stored records behind our published score are the raw output of it.

What this method does not measure

Stated here so a buyer does not have to find them.

• It is a sample, not a census. It reports what an engine says when asked your questions. It cannot count every time you were cited to somebody.
• Questions are not weighted by real demand. No engine publishes how often a question is actually asked, so every question in the set counts equally, which is a simplification.
• One engine automatically, today. Perplexity is the only engine we hold an API key for. ChatGPT, Claude and Gemini adapters are built and awaiting keys. Google AI Overviews and Copilot have no public consumer API, and we do not track them on any plan. A score from one engine says nothing about the others.
• No personalisation or memory effects. API answers are not the answers a logged-in user with chat history receives. Those may be better or worse for you and we cannot see them.
• Sentiment is recorded but excluded from the score. Being named badly and being named well both count as a mention. We keep sentiment as a separate field rather than smuggling a judgement into a headline number.
• English-language phrasing. The same question in another language can return a different set of brands entirely.
• Scores are not comparable between vendors. Different weights, different questions, different engines. Compare yours only with your own.

Common questions