The attic
Posts on what we build, test and learn, written down and dated. That includes what didn't work.
AI-generatedWe now teach, and we take consulting requests
Second Attic has two new doors: live courses for small companies, and consulting inside one company.
Read the post
AI-generatedWhy we publish the tests that fail
A result nobody can check is a claim. A failed test shown with its numbers is the only proof we have.
AI-generatedSecond Attic goes into public build
The parent brand, four scoped products and two research systems are set up, and this site is the first thing built in public.
AI-generatedCrewFinder: the paywall and heatmap pilot
A live pilot testing where the paywall sits, with usage heatmaps logged across leisure and competitive boats.
AI-generatedTerrarium: how a lie spreads through a closed population
Closed environments of agents with their own personalities, watched while information, true or false, moves through them.
AI-generatedThreshold: testing how software gets paid for
Where the paywall sits, when the upsell arrives and how the price is framed, tested across products and kept as numbers.
Nothing in this category yet.
Still running
- LiveCrewFinderThe paywall and heatmap pilot is running on the live product.
- RunningTerrariumClosed populations of agents, studied for how information spreads.
- RunningThresholdPaywall, upsell and price tests across products.
- BuildingScully, Mimir and TelescreenNamed and scoped. They get a post here the moment there is something to show.
What didn't work: nothing has landed yet. When a test fails, it gets a post here, dated, at the same weight as the rest.