How we decide what goes up.

We publish explanations of the work we do, industry examples, short essays and reference pages. The aim is to be useful before someone hires us. We do not publish invented client stories, reviews, ratings, awards, user counts or performance claims.

Every build example is a scenario unless the page says a client agreed to be named. The label appears on the page, not only in a footnote.

Numbers and factual claims are linked to a primary source where one exists: a study, a regulator, a standards body or the organisation that published the data. When we rely on a secondary summary, we say so. The sources page keeps the full list, including what each study does not show.

We do not imply affiliation, endorsement or partnership with a source, a vendor or a client unless that relationship is confirmed in writing.

Articles show the month they were last substantially edited. We change that date when the argument or facts change, not when a page is reformatted or a menu moves.

If a source stops resolving, or a claim turns out to be wrong, we correct the page and note the change. Email us with the page and what looks wrong; we will check it against the source.

The studio agent on this site answers from published studio material and labels estimates as indicative. It is not a person and it can be wrong. A human reads enquiries before agreeing to anything.

We use software, including AI tools, to help draft and check our own work. A person is responsible for every published page, every source link and every claim. We never use a client's customer data to train a model.

The examples show the shape of a build: the problem, the workflow, the safeguards, the numbers and the limits. They are not a promise that your business will get the same result.

If we publish real client work in future, it will include the measurement period, the method, the limits and the client's written permission. Until then, the site says plainly that it has no published client results.