Engineering journal
Written by the people doing the delivery, about decisions we have actually had to make. No thought leadership, no gated downloads, no newsletter wall.
Latest
Teams treat page speed as an engineering concern and then wonder why it never gets prioritised. Reframing it as a number the business owns changes what actually ships.
Read article
A retrieval prototype takes an afternoon. Making it dependable enough that staff stop double-checking every answer takes far longer — and it is rarely the model.
Read article
Row-level, schema-level or database-per-tenant. The decision is easy to make early and expensive to revisit once you have signed your first enterprise customer.
Read article
Deployment frequency, lead time, change failure rate and restore time. Four numbers that tell you more about an engineering organisation than any roadmap review.
Read article
Retrofitting accessibility costs roughly ten times what designing for it costs — the difference is decisions that are free early and structural later.
Read article
Before anyone proposes a rewrite to save money, there is usually 30% sitting in idle capacity, forgotten resources and storage nobody has looked at in two years.
Read article
An API is a promise with a very long expiry date. A few decisions made in the first fortnight determine whether keeping that promise is comfortable or miserable.
Read article
Fixed-price software fails when scope is guessed. Two weeks of structured discovery turns most of that guesswork into evidence — and sometimes stops the wrong build.
Read article
Configurability is how teams avoid making decisions. Every toggle you ship is a decision deferred, a test matrix doubled, and a support conversation guaranteed.
Read articleNo articles in this topic yet.
Why we publish
Every article here started as an internal document — an architecture decision record, a postmortem, or an argument two engineers had in a pull request.
Publishing forces us to defend positions in public, which is a useful discipline. If you think we have something wrong, we would genuinely like to hear it.
Topics we write about
Available for new engagements
Most of these articles came out of client engagements. If you are wrestling with one of these decisions, we are happy to talk it through.