Original Measurement, Published

This page describes a single piece of work: a company producing a measurement from its own data, and publishing that measurement in a form a third party can take up. Measurement here does not mean promotional material. A measurement is primary data that carries a date, states its method in writing, declares its limits openly and sits at a permanent address.

We applied this to ourselves before writing it up as a service. In July 2026 we measured our own server logs and published the result in our measurement series: across the 17 days between 15 and 31 July, 326,764 requests, of which 15,278 came from AI crawlers and 11,585 from classic search engine crawlers. The method, the number of lines that could not be parsed and what the measurement does not show are written on the same page. Having tested it on ourselves before selling it is not a credential; it is a minimum condition.

Why measure, why publish

The share held by corporate marketing pages in answer engines has been measured, and it is small. A vendor's own study, the Omniscient Digital analysis (8 January 2026; 240 branded queries, five answer engines, 23,387 citations), reports that about pages take 1.92 per cent of citations and home pages 1.82 per cent. The classic addresses on which a company describes itself therefore occupy a small place in the citation economy.

Pages carrying primary research point the other way. The analysis by Indig and Johnson published in Search Engine Land on 8 July 2026, covering 301 pages, 316 prompts and 1,075 citations, reports 11.3 citations per page for pages containing primary research against 3.4 citations for other content. We write in the same sentence that this ratio rests on a base of only eight pages; it is an observation on a small base and cannot be used as a fixed multiplier. We also keep the measured object as it stands: not original content in general, but primary research.

Read together, the two produce the following. What a company says about itself is already abundant. What is scarce is data that exists nowhere else and can be verified. The party that gets cited is often not the one that explains best, but the one that is the only source.

Three conditions that turn a measurement into a publication

A published number is no more than a claim unless it can be verified on its own. For a measurement to count as citable, all three are needed together.

• A statement of methodology. Where the data came from, what was counted, what was not counted and which lines were left out are written on the same page. If the method is not written down, the number cannot be reproduced.
• A permanent address. The measurement sits at its own address and that address does not change. A number embedded in a slide, a newsletter or a social media post cannot be cited, because there is nowhere to point to.
• A machine-readable record. The measurement is marked up with Dataset structured data; the period, the measured variables, the measurement technique and the licence fields are filled in. The table a person reads and the record a machine reads sit on the same page.

A number published without all three is an unverifiable claim. That is why we keep the three of them on the same page in our own measurement.

The work we do

The work starts from data the company already holds. Most companies have measurable primary data that nobody has looked at as data: server logs, order records, service calls, production waste, delivery times, fault reports, application and rejection rates.

• We go through the records you hold and work out which of them could become a publishable measurement.
• We put the counting rule and the classification in writing. What is not counted is written down as carefully as what is counted.
• We set up the script that produces the measurement, so the same measurement can be produced again in the next period.
• We prepare the result for publication together with the statement of methodology and the section on limits.
• We build the Dataset structured data and link it to the catalogue record of the series.
• We hand the production path over to you, so the next period can be produced with a single command.

A one-off measurement is a snapshot. Its value comes from repetition: a second and a third period measured by the same method make the number comparable. We therefore set the work up as a series, not as a single page.

Why the method stays fixed from period to period

The value of a measurement comes from being comparable. If the classification list, the data source or the counting rule is changed during the period, the difference reflects the method rather than the measurement. The method is fixed for that reason, and on the day it changes the change is stated openly in the period concerned.

What we do not promise

• We do not guarantee citations, rankings or visibility. Publishing a measurement is an input, not a commitment to a result.
• We do not say in advance what the measurement will show. Which way the number will go cannot be known before it is measured.
• If the result is not what you wanted, it is published in the same form. Any edit that turns a measurement into a text published because its result was liked destroys the only value the publication has. If you do not accept this condition, this service is not for you.
• We do not produce data. If there is no measurable record, there is no measurement; in that case we tell you which record needs to be kept from now on, and we stop there.
• We do not publish data that cannot be made public. No field carrying a trade secret, personal data or customer identity goes into publication; if the measurement cannot keep its meaning once those fields are removed, it is not published.
• We do not present another property's data as your measurement. Every published number comes from your own records.

Where to start

The first step is to see whether you hold a measurable record. To talk this through, you can reach us from the contact page. To see how a published example looks, you can read our own measurement series; what this service produces is exactly that form.

The evidence side of this work is covered on the evidence surfaces page, and our measurement and reporting method on the method page. The scope and the limits of the work are set out on the scope and exclusions page.