Household jobs rarely fail because nobody cares. They fail because the plan lived in one person's memory: who was supposed to take the bins out, whether the bathroom was cleaned this week, and which child had already done a turn. A memory cannot be checked, shared, or handed over when someone is busy, ill, or away.
Building for that problem in TaskFamily turned into a familiar engineering lesson, just applied to a kitchen instead of a codebase. State that lives in one place, owned by one person, and queried by shouting down a hallway will drift. The fix is not better remembering. The fix is moving the state somewhere everyone can read it.
What worked
One readable source of truth, not a stream of reminders. A reminder is a message about state; a shared list is the state itself. When every recurring job is written down once, with an owner and a day, checking what is true takes a glance instead of a conversation. Arguments move from "you never told me" to a quick look at the board, which is a much shorter argument to have.
Rotation made visible, not negotiated weekly. Fairness is hard to feel when the history is invisible. When the list shows who did what last time, rotation stops being a claim and becomes a record. Swapping a task is then a simple edit, not a renegotiation, because both sides are looking at the same history. In system terms, the audit log does most of the social work.
Small, exact items beat broad intentions. "Keep the kitchen tidy" decays within a day. "Wipe the hob after dinner" survives, because it is checkable: anyone can tell whether it happened, and the person doing it knows exactly when they are done. The same rule that makes a good ticket makes a good chore — specific enough that done has a clear meaning.
Designing for handover, not just assignment. People get busy, ill, or away, and a list held in one head breaks at exactly those moments. A shared list hands over cleanly: the next person reads the same board and picks up the same plan. No briefing, no reconstruction from messages, no single point of failure sitting in one person's memory.
TaskFamily is where this pattern lives as a shared household board, with named owners for each job, due rhythms, and a finished state anyone in the household can check.
What broke
Listing everything on day one. A complete inventory of every possible household job looks thorough and feels oppressive. Households that started with a short list of recurring jobs actually kept using it; the exhaustive version got ignored, which is worse than a partial list that stays current. Coverage mattered less than currency.
Treating done as a private action. Early versions let a person finish a job without that change being obvious to everyone else. The state was technically shared but practically silent, so other people kept asking anyway. Making completion visible — a finished state on the shared view, not a message in a thread — was what finally retired the "has it been done?" question.
Confusing flexibility with vagueness. Letting any item be edited at any time sounded forgiving and produced churn: times moved, owners blurred, and the board stopped feeling settled. What worked better was a stable weekly shape with explicit exceptions. The plan stays still; only the deviation gets edited.
The lesson that carried over
Most coordination problems are not motivation problems. They are visibility problems wearing a motivation costume. When the plan, the owner, and the history are in one shared place, people argue less about what was agreed and spend the saved effort just doing the next item on the list. That holds for a household board, and it holds for most other shared systems that started life in one person's head.
What state in your own projects still lives in one person's memory — and what broke when that person was away?
