Latest
11 SEP 2026ADDEDHousehold plan now derives quantities from people and pets instead of a flat 72-hour default11 SEP 2026FIXEDExpiry dates entered as DD/MM no longer shift by a day across timezones10 SEP 2026ADDEDBucketed indexes by identifier prefix — weekly refresh is now ~340 KB, down from 8.6 MB10 SEP 2026FIXEDMatch run no longer blocks the UI thread on older Android devices8 SEP 2026ADDEDMeeting point can be set without an address — a dropped pin is enough8 SEP 2026REMOVEDDaily nudge notifications. People muted them within a week; reminders are tied to real expiry dates only11 SEP 2026ADDEDHousehold plan now derives quantities from people and pets instead of a flat 72-hour default
Changelog →
Novel ApproachesPre-launch
Join waitlist
← All posts
Company22 Jul 2026metricsproductfocus4 min

The four numbers we let ourselves look at

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

If Readinity is working, you open it a few times a year. If Cyberinity is working, you hear from it almost never. Optimising for time in app would make both products worse.

So we track four things: whether a plan reaches a finished state, whether flagged items get resolved, whether an alert leads to an action within a week, and how many people mute us.

The last one is the honest one. A mute is a message: we spent attention we had not earned. When the daily reminders got muted by a fifth of testers, we deleted the feature rather than retune its timing.

We do not track session length, screens per session, or anything that would make a dashboard look busy at our expense.

Written by the Novel Approaches team · 22 Jul 2026
Share