Editorial Charter
How BvLogic research is governed: eight rules with what each one costs, the conflict of interest stated first rather than disclaimed last, and the procedure for correcting an error in public.
Version 1.0, published 2026-08-11. Eight rules, each with what it costs, and the procedure for when we get something wrong.
BvLogic is not a neutral party. It intends to sell enterprise services, it publishes a specification it wrote, and it benchmarks its own system. Each of those is a reason to discount what we publish, and stating them is not a disclaimer at the bottom of a page: it is the first thing a reader should know. The response to a structural conflict is not to claim independence, it is to publish the method so the conclusion can be checked without trusting the author.
Why a charter at all
Research published by a company that sells into the market it researches has one structural problem, and everybody in the market knows it: the findings that help the seller are the ones that get published. There is no way to prove that has not happened. What can be done is to write down in advance what would have to be given up, put the rules where a reader can hold us to them, and record every correction in public rather than editing the original quietly.
It is published now, before the research programme it governs has anything at stake, because a charter written after the first inconvenient finding is a negotiation rather than a rule.
The rules
Each carries what it costs. A rule with no cost attached is a description of what we were going to do anyway, and those are not worth publishing.
No payment, contract or commercial relationship influences whether something is researched, what is found, or whether it is published.
In practice. There is no paid placement, no sponsored finding, no vendor tier that buys inclusion, and no arrangement in which a subject reviews a conclusion before publication. A subject may correct a fact; nobody outside the editorial line may change a judgement.
What this costs us. It rules out the revenue model most research businesses in this market actually run on.
A finding is published when it is against our own interest, on the same terms as one that flatters us.
In practice. Our benchmark score has already been published going down, from 66.7% to 58.3%, while the cause was still unknown. The claim that we had five customer leads was corrected in public to zero after the records were read individually. The Trust Center states we hold no certification.
What this costs us. Every one of those is a sentence a competitor can quote.
Every published figure carries its method or declares that it has none.
In practice. A number states what was measured, against what, and when. Where there is no base rate the finding says so and argues structurally instead. This is enforced in code, not intention: tools/evidence.py is the single implementation and it is mutation-tested.
What this costs us. It makes findings longer and less quotable than a bare statistic.
We do not score, rate or rank a product we have not run ourselves.
In practice. No competitor rows on the benchmark. No vendor scores on the procurement tool, where the buyer scores and we supply the criteria. No technology rating system, despite it being repeatedly proposed, until there is a methodology and a system actually exercised against it.
What this costs us. A vendor rating table would be the single most linkable thing this business could publish.
No median, quartile or peer comparison is published without a real population, its size, and how it was collected.
In practice. The Technology Index has no peer column and the page explains why. Published industry statistics appear as reference points, labelled as a different population, because calling an external figure 'people like you' is fabrication in a quieter voice.
What this costs us. Peer comparison is the feature buyers ask for most.
Errors are corrected on the page that carried them, dated, with the original claim still legible.
In practice. A correction is not a silent edit. Where a published prediction is amended, the amendment and its reason render on the page alongside the original hash, and the build refuses to ship a frozen prediction whose text changed without one.
What this costs us. It leaves a permanent, public record of having been wrong.
A citation must point at something a reader can open. A source that cannot be checked is not a source.
In practice. Internal links are verified to resolve at build time. A citation to a page we never published fails the build, because a broken citation looks verifiable and is not, which is worse than none.
What this costs us. Some genuinely known things cannot be cited and therefore cannot be stated as fact.
Commercial work and research use separate material. A client engagement never becomes a published finding without written permission, and a published finding is never shaped to suit an engagement.
In practice. No customer is named, quoted or characterised anywhere in published research. Today this is trivially satisfied, because there are no customers. It is written now precisely because it is easy to keep now.
What this costs us. Case studies are the most persuasive asset a services business has, and this forecloses the shortcut version of them.
When we are wrong
The procedure, so that following it is not a decision made under pressure.
- Correct the page that carried the error, dated, with the original claim still readable.
- State what was wrong, not merely what is now right. A correction that does not name the error teaches nobody anything.
- Say how it was found. If a reader found it, say so.
- If the error came from a method rather than a typo, fix the method and say which other findings used it.
- Where a guard could have caught it, add the guard and prove it fires by breaking it deliberately.
What this is not
- A claim of independence. BvLogic sells into the market it researches, and no charter changes that.
- A certification, an accreditation, or a standard anyone has audited us against.
- A promise that everything published is correct. It is a commitment about what happens when it is not.
- A description of an existing research institute. There is a research programme; the separate institutional identity does not exist yet, and calling it one before it is would be the first violation of the charter.
Everything here is free and stays free. There is no form in front of any document. If you want to know when new guides and templates go up, leave an email.
Roughly monthly. Unsubscribe in one click. We do not share your address, and we will not call you.