Status: Private staging — approved public evidence snapshot
Derived from the private HomeLog Agent Log for July 25, 2026. This page preserves only the evidence relevant to the adoption-funnel claim. It is not the complete operational record.
A production audit reported that zero out of seven non-founder users had reached first success through the onboarding and notification path being measured.
The number was accurate within that boundary. It did not establish that nobody beyond the founder used HomeLog. It established that the measured path had not carried any of those seven people to its defined success event.
That distinction became important one day later, when field evidence showed that people classified as non-adopters were already operating HomeLog through LINE—outside the surface covered by the original funnel. Six days after the audit, the interpretation was corrected explicitly in the Agent Log.
HomeLog was being treated as a production pilot rather than a research prototype. The audit followed the real user journey on an actual iPhone through LINE and its embedded LIFF (LINE Front-end Framework) web interface.
The goal was not simply to count broken buttons. It was to find every point where the founder had to step in and explain something the product should have made understandable on its own.
The working unit became a founder-intervention class: a distinct mistaken mental model that required the founder to rescue the user.
The audit stabilized at six classes across 19 observed instances:
The audit stopped after another pass produced no new class—not because the product was considered finished.
The measured path centered on app-assisted onboarding, presence, and useful household notifications.