Policy

Conflict of interest

Martenfield publishes measurements of a category and sells work to companies in that category. That is a real conflict and we are not going to describe it as anything else.

Written , before any census edition has published

Here is what we do about it

The prompt pack is frozen and published before collection.
Prompt selection is the one place where a client could be quietly favoured without anyone outside being able to see it. Every edition’s pack is published with a freeze date that precedes the first collection call.
No paid placement, ever.
There is no advertising product, no sponsored row, no featured tier. No position in any table is influenced by payment. This applies identically to clients and non-clients.
Client rows are marked in the table.
Every edition, permanently, in the table itself. If a client’s position improves across editions, that is published as an improvement and marked as a client’s improvement.
Roster rules are published, with exclusions.
Every company considered but excluded is listed with the dated rule that excluded it.
Martenfield does not appear in any table it publishes.
We are not a vendor in these categories, so there is no position for us to hold.

No edition has published yet, so none of these rules has been tested against a real table. They are written first so that they apply to the first one. The method page sets out how the measurement itself works.

Client names

That a company is our client is public, and its name may be published. Publication rights are written into the contract at signature.

Deal specifics stay private: what the client paid, its internal data, its strategy, and anything else the contract protects.

Where a client appears in a table we publish, its row is marked as a client in the table itself, in every edition, permanently. Not in a footnote, and not on a separate disclosure page.

What we do not claim

We are not independent. A vendor that scores badly in our measurements is a prospect for the service that fixes it, and the scoreboard creates demand we then sell into. That is the structure, stated plainly.

What we claim instead is reproducibility. The method is published, the pack is frozen, the runs are logged, and the raw payloads are kept. Anyone can run it again and get their own numbers. That is checkable. Disinterest is not, and we do not have it.

Start here

Find out whether the models cite you.

The diagnosis measures 50 buyer-intent queries across 5 engines at 3 runs each, maps the sources the models draw from, and hands you the raw runs and the order of work. 2 weeks, $1,200.

See what the diagnosis covers → Email us →