Editorial standard
How Katalyst researches and publishes commercial insight.
Every article on this site makes claims about how hospitality businesses lose and recover money. This page sets out where those claims come from, how they are separated by strength, what happens when one turns out to be wrong, and why some questions are left deliberately unquantified.
Who publishes this.
Katalyst Labs is the author of record. Articles are published under the organisation rather than a personal byline — a deliberate choice, not an omission. The work draws on commercial roles held inside hotels, hospitality groups, F&B operations and short-stay portfolios, and the argument should stand on the evidence and reasoning shown rather than on whose name is at the top.
That experience is used in two specific ways: to decide which questions are worth answering, and to describe operating reality that is not documented anywhere public. It is not used as evidence for a number. Where an article needs a figure, the figure comes from a named source or is labelled as an estimate with its assumptions on the page.
The organisations named under Selected Operator Experience reflect prior employment and operating roles. They are not clients, partners or endorsements, and no article implies otherwise.
Source hierarchy.
Sources are ranked, and the rank is visible in how a claim is worded. A platform’s own documentation and an interested vendor’s survey are not treated as equivalent.
Primary platform and official documentation
Published documentation from the platform or authority whose behaviour is being described — Google Search Central, Google Hotel Center Help, OpenAI's crawler documentation, official statistics offices. Quoted rather than paraphrased into something stronger, and re-checked on the day of publication.
Official statistics and regulatory publications
Government and regulator data, carried with its market, period and definition. A figure for one market is never restated as a general fact.
Vendor research and industry surveys
Useful and structurally interested. Always labelled as vendor-published, always carried with sample size and market where disclosed — and left out entirely where the sample is undisclosed, because an unsourced percentage is decoration.
Operator experience
Direct experience from commercial roles inside hotels, groups, F&B operations and short-stay portfolios. Presented as operating judgement, not as measurement, and never as a client result.
Structured observation
Booking paths walked, listings and profiles inspected, prompts run and recorded with date, platform and sample size. Pattern evidence, labelled as pattern evidence.
Facts, inference and estimates are not the same thing.
Most misleading commercial writing is not false. It presents a reasonable inference in the grammar of a fact. These five categories are kept visibly apart.
| Type | How it is handled |
|---|---|
| Fact | Verifiable against a named source, with the source beside the claim rather than in a footnote. If the source cannot be linked or named, the claim does not run. |
| Inference | A conclusion drawn from facts, signalled as reasoning rather than evidence. The reader should be able to disagree with the inference while accepting the facts under it. |
| Estimate | A quantified judgement with its assumptions visible. Every estimate states what it assumes, and the arithmetic is shown so it can be re-run with different assumptions. |
| Illustrative calculation | Worked arithmetic on invented but realistic inputs, labelled illustrative on the same page as the numbers. Never a Katalyst result and never a benchmark. |
| Anonymised scenario | A recognisable shape of problem, described without identifying anyone. Not a case study, and never implied to be a disclosed engagement. |
External statistics carry their market and date beside the number, on the same page rather than in a source list at the bottom. A figure whose market is not stated is a figure that will be misapplied.
Verification, corrections and updates.
Human review. AI tools are used for research, drafting and editing, in the same way they are used in the client work. Every claim, figure and source is checked by a person before publication, and accountability for what is published sits with Katalyst regardless of how a sentence was drafted.
Re-verification. Platform behaviour changes. Claims that depend on a platform’s current documentation are re-checked when the article is substantively reviewed, and the review date is visible on the article.
Corrections. A factual error is corrected in place and noted on the article. Corrections are not made silently, and an article is never quietly deleted to resolve one.
Substantive updates. The review date moves only when the argument, evidence or recommendation actually changes. It is not refreshed to make a page look current — a date that moves without the content moving is a small lie with SEO benefits.
What is deliberately not published.
- Client results, named or implied. Katalyst has no disclosed client results, and the articles say so rather than implying otherwise.
- Benchmarks without a stated sample, named market and visible methodology.
- Adoption statistics that cannot be traced to a primary source, however widely they are repeated.
- Opaque scores, indices, badges or certifications.
- Guaranteed or implied rankings on any search or AI surface.
- Vendor comparisons that a commercial relationship would make unreliable.
- Content published to maintain frequency. A skipped month is preferable to filler.
Some commercially important questions stay unquantified here, and that is a decision rather than a gap. Where a credible number would need a sample Katalyst does not have, the mechanism is explained and the number is left out. An invented figure is more damaging to a reader than an honest absence.
Commercial relationships.
Katalyst currently holds no technology partnerships, reseller agreements or referral arrangements, and no vendor pays for mention or placement in any article. If that changes, the relationship is disclosed on the article and recommendation that relies on it. The full position is published at the technology recommendation policy.
Articles do carry calls to action, because Katalyst is a commercial business. The standard applied is that every article must be useful to a reader who never contacts us — if the argument only works as a sales premise, it is not published.
Questions or corrections: hello@thekatalystlabs.com. Read the Journal.