Editorial policy

We sell a product that plans, researches, drafts, and reviews content. It would be strange to use something else to write this blog, so we do not. Most posts here are produced by myHERALD, and this page says exactly what that means.

The short version: the machine does the production work, a person decides what runs. No post publishes without a named human approving it.

What the product does

For a typical post, myHERALD plans the topic against what we have already covered, researches it, writes the draft against our brand voice, scores it against a fixed set of review criteria, and queues it. That is the same pipeline our customers run.

Posts marked AI-assisted at the bottom of the page went through it. Posts without that note were written by hand.

What a person does

A human sets the direction, edits, and approves. Approval is not a formality and it is not automatic: a draft sits in a review queue until someone accepts it, and drafts get rejected. Editorial responsibility for everything published here rests with a named person, not with a system.

Sourcing rules

Every claim on this blog has to pass three rules.

1. Attribution is visible to you, not just to us. Every statistic, study, and third-party claim names its source in the sentence. Being able to trace a number internally is not enough: if you cannot see where it came from, it is unverifiable to the person who matters.
2. We cite the original publisher, not the blog that repeated it. Most statistics circulate through aggregators. We follow them back to whoever ran the survey or the study and cite that, and we say so when the original is paywalled or unreachable.
3. Anything we cannot trace does not run. A claim with no locatable source gets cut, not hedged into a vaguer version of itself.

We apply these retroactively. When we find a published post carrying a number we can no longer trace, we correct or remove it rather than leaving it up.

First-party evidence

Where we have our own measurements, we prefer them to borrowed statistics. Our own benchmark numbers, test results, and operating data are the one thing on this blog that nobody else can publish, and they are what we would rather be cited for.

Corrections

If something here is wrong, tell us and we will fix the post rather than quietly delete it. Write to our contact page, or see the privacy policy for how we handle AI transparency across the product.