Slow writing about building software that stays. Case studies, essays, occasionally a small confession when we get it wrong. Twice a month, when there's something to say.
What happens when a saturated market meets a brand that actually means something? Fourteen months, a quote calculator, no paid ads, and a #2 Google ranking. Here's how it played out, week by week.
Read the case →Three projects we shipped, three projects no-one used. The post-mortems, the patterns, and the questions we now ask before we touch a keyboard.
The line between assistant and agent. What it changes for product design when the LLM is a colleague, not a function call.
The HelderCristal one, in detail. The data model, the UX choices, the silly conversion improvements we found in the second week.
Fixed-phase pricing, retainers, and the conversation we have on the first call. Why selling hours is the wrong contract for the work we do.
An honest field report after running n8n in production for a dozen clients. The good, the workaroundy, and the now-obsolete.
Why small studios with regular people are quietly outperforming the hero-shop model. A defence of a small team in a small room.