How we build.
Our products cover very different ground, from a household’s bank statements to an engineering team’s architecture. These eight principles hold for all of them. If a product breaks one, the product is wrong.
The principles
-
Show the working.
Every number we put in front of you should trace back to something you can check: a statement row, a changelog entry, a line of code. A result you cannot inspect asks for trust, and we would rather earn it.
-
Your data stays yours.
Ask for as little as possible and keep it where you can see it. Ledgerlite never receives your statements, and DepMatrix runs without an account. Everything you put into one of our products can come back out.
-
Narrow on purpose.
A tool that tries to do everything becomes one more thing to work around. Each of ours does one job properly and stays out of the way the rest of the time.
-
Find the problem while it is cheap.
A breaking change costs least in a browser tab, a duplicate charge in this month’s recap, an architecture drift in review. Good software moves the discovery earlier.
-
Complexity is a cost someone pays.
Usually the next person to touch it. We count it like any other cost, and we leave features out until they have earned a place.
-
Boring technology, careful outcomes.
The novelty budget is small and mostly spent. We use it on the part that is the product, and let the database be a database.
-
Be honest about what is unfinished.
Status labels mean what they say. Beta is usable and still moving; in development means not ready yet. Where a product deliberately leaves something out, its page says so.
-
Write for the reader at the worst moment.
Help text, error messages and documentation get read under pressure by someone new. Write for that person and you have written the version that matters.
What we won’t do
A principle only counts once it has cost something. These are the ones we have turned work down for.
- Ask for your bank password, or any password that is not ours to hold
- Add a cookie banner so we can learn which of two buttons you clicked
- Hold your data hostage as a retention strategy
- Charge per seat for something that runs on your own machine
- Need a six-week onboarding programme before a product is useful
- Call a rewrite a strategy