The Novel Approaches blog

Everyone who tests Readinity arrives with a half-finished plan from three years ago. The gap isn't knowledge — it's maintenance, and nothing in the category treats it as the main job.

Monitoring services ask you to hand over the very identifiers you're worried about. We spent a month proving we could do the match locally instead — here's the shape of it.

The standard playbook is to show you a scary finding and put the fix behind a tier. It converts extremely well. We're writing down why we're not doing it, before revenue makes it tempting.

Sensible advice says ship one thing. Readinity and Cyberinity share a premise, a data model and a definition of restraint — splitting them would have meant building both halves twice.

We tested the evacuation page with people holding a phone at arm's length in a dark stairwell. Every sentence that needed a second read got cut.

Three of the largest available datasets would have doubled our coverage on day one. We left them out, and this is the reasoning in full.

Bottled water, batteries, medication and insurance documents all 'expire' in four incompatible ways. Modelling that took longer than the app's entire UI.

Engagement is the wrong goal for software you should rarely open. Here is what we track instead, and what we deliberately do not.