Five GO/NO-GO criteria for the 2026-05-31 public beta cutover.

Distinct visitors at each engagement step. Steps are nested — every engaged visitor is also a visitor, every guide reader is also engaged.

A user is activated when they fire house_created + project_created + task_created within 24h of signup_success. Strict by design — mirrors the minimum input the product needs to deliver any of its four value props (see lib/sources/posthog.ts).

Distinct app users with any event in the last 7d (WAU) vs. 30d (MAU). WAU/MAU is a stickiness proxy; top events show feature adoption in the last 7d.

Total custom-event firings and average actions per user across the last 7d, plus per-event trends for the four key action events.

Did users finish what they started in the last 7d? Window-aligned counts; cycles spanning the window boundary suppress completion %.

Distinct users at each step in the last 7d. Drop-off between steps highlights friction in core onboarding. Mobile calculator funnel (opened → completed per tool_slug) renders below.

Distinct users at each paywall step in the last 7d. The last two bars (Failed, Dismissed) are alternative terminal outcomes to Completed — sum the three to compare against Attempted.

Project shares + trade contact shares fired in the last 7d. The coefficient is shares ÷ 7d actives — a directional proxy; true K needs share-attribution params on the share links.

Topic creates, replies, moderation flags and blocked users from the mobile Lexicon integration over the last 7d.

Things users do to/for other users in the last 7d. Adding tradesmen to the directory, leaving reviews, exporting data for backup or sharing.

Cohort retention over the last 7d. Loading…

% of installs in the last 7d that hit each value-prop event. Loading…

Drop-off per step over the last 7d. Loading…

% of each segment that used each feature in the last 7d. Loading…