Building in a lab: why ideas need a space of their own

By Simon Torngren · July 29, 2026

Most ideas do not die because they are bad. They die because nobody has anywhere to put them. An observation shows up in the middle of a client project, someone says it is interesting, and then the week ends. The note stays where it is. That is the entire funeral.

This page is about why ideas need a space of their own to survive their first day, what that space does, and why the first things built there are the things we need ourselves.

An agency has deliveries going out, clients waiting and a calendar that is already full. Let half-finished product ideas into that environment and they compete with the work that pays the bills, and they lose every time.

That is the right priority. An agency that lets experiments push out client work is a bad agency. But the two things need opposite conditions: delivery needs nothing to break, testing needs things to be allowed to. The environment is built for the first.

Formise solves the conflict by moving it. The lab is a space of its own with room to test, build and get things wrong without any client delivery being at risk. What that does for the agency matters just as much: it no longer carries experiments and can be entirely an agency. How the two spaces fit together is under About.

The word lab can sound like an excuse to avoid deciding. Here it means the opposite: an idea is tested properly before anyone bets on it.

The difference from simply having ideas is ownership. Formise is a company that builds its own products, not a consultancy drawing for other people. When something turns out well it belongs to us. When something breaks it is ours that breaks, and that is the whole point: whoever owns both the upside and the downside does not need to prettify either.

The testing follows an order that looks the same every time: explore, define, build, launch. Step by step, it is on the process page. What the order buys matters more than the steps themselves: most things fall early and cheaply, and what reaches a build has earned it.

The fastest way to know whether something holds is to use it yourself, for real, where it shows when it does not work.

So the lab's first deliveries are its own. memorise.se was built in Formise: an agency selling technical SEO, performance and measurability can hardly run a slow site with holes in its tracking, and that site is reviewed by exactly the people who know the craft best. The whole build, from rendering to tracking and accessibility, is in the memorise.se case. The site you are reading this on was built the same way, for the same reason.

Building for yourself first is not modesty. It is the harshest review there is, because you cannot invoice away a site you have to live with.

The lab is not a storage facility. A build that shows it holds moves out: its own name, its own site, its own company. Dynamic SEO went that way, and why the model is rigged like that is in the spinout model.

What stays in the lab is what a lab should contain: what is still being tested, what is allowed to break, and the next note that deserves better than being forgotten.