What has to be true before we'll sell you something.
This page exists so that our claims can be checked rather than believed. It is versioned and dated. When it changes, the old version stays up.
The five tests
- It has to beat what we'd otherwise recommend. If someone else already makes the best version of a thing, we will tell you to buy theirs and we will not make a competing one. Being able to make it is not a reason to make it.
- It has to do less than the competition, not more. If the only way to justify a product is a longer feature list, a bigger page count or more SKUs, it fails. Every product should be defensible on what it leaves out.
- The reasoning has to be publishable. If we can't explain in public why this and not that, we don't ship it. Every product page carries the alternatives we considered and why they lost.
- We have to be able to name its flaws. Every product page says who the product is not for. If we can't think of anyone it isn't for, we haven't understood it well enough to sell it.
- We have to use it ourselves. Not once, for a photograph. Regularly, in the way a customer would.
What this rules out
Deliberately, and permanently: manufactured urgency, countdown timers, inflated MRPs with a line through them, fake scarcity, undisclosed affiliate links, sponsored placement, paid reviews, and any claim we can't show the working for.
It also rules out growth by catalogue. We expect the number of things we sell to stay embarrassingly small. That is a decision, not a stage we're passing through.
How we're funded
By people paying for things we made. Not by display advertising — we tried that, it didn't work, and it pointed our incentives at traffic instead of usefulness. There are no ad units on this site.
If we ever use an affiliate link, it will be labelled as one in the same sentence, not in a footer disclosure.
Things that failed these tests
This list will be longer than the list of things we sell. That's the point of having a standard at all.
- The Crux Monthly (e-magazine) Announced October 2026, never built. Failed test 1 — it was a format decision, not a thing anyone had asked for.
- The Crux Markets (daily brief) Announced for early 2027, never built. Failed test 2 — it was the news product again, with a narrower audience and the same economics.
- Branded merchandise Failed test 5. We would not have used it. A logo on a tote bag is not a product.
- A cross-brand membership Failed test 3. We could not write a sentence explaining what problem it solved for anyone other than us.
Corrections
If something on this site is wrong, or a product doesn't do what we said it does, tell us. Corrections get published on the page where the mistake was, not quietly edited out.