Knowledge-Base Cleanup Rotation¶
Cleanup workflow — instantiates Preventive Maintenance Cadence
Rotates through documentation, FAQs, templates, or precedent libraries to remove stale guidance and repair broken references.
A Knowledge-Base Cleanup Rotation keeps an information corpus from rotting by working through it in rotation, a slice at a time, so that every article eventually comes back around for a freshness pass. Its defining move is coverage over a cycle with a named steward and reserved bandwidth: rather than cleaning whichever page someone happens to complain about, it guarantees that the whole library is revisited over a known period, that each slice has an owner responsible for it that round, and that the time to do the cleanup is protected up front instead of scavenged from spare moments that never come. The unit of work is editorial upkeep — delete the obsolete, fix the dead link, reconcile the contradiction, update the screenshot — not detection and not verification. Its whole reason for existing is that a knowledge base decays silently and no single failure is ever urgent enough to fix on its own.
Example¶
A SaaS company's help center has grown to 1,800 articles, and support agents keep quoting steps that no longer match the product. The team institutes a cleanup rotation: the catalog is divided into batches by topic, and each two-week cycle one batch comes up for review. Whoever owns support content that cycle is the batch's steward — they prune articles for retired features, repair links that now 404, fold three overlapping FAQs into one, and refresh outdated screenshots. Critically, four hours per cycle are carved out and defended as cleanup time, so the rotation does not evaporate the moment a launch gets busy. Over roughly a year the entire library has been swept once, and high-traffic batches come around more often. The corpus stops accumulating content decay and link rot — the slow drift where documentation quietly stops describing reality — because every article has a scheduled date with a named owner rather than waiting for a reader to hit the wall.[n1]
How it works¶
- Divide the corpus into batches. Slice the library so the whole thing is covered over one rotation, weighted so high-use or fast-changing sections come around more often.
- Assign a steward per batch, per cycle. Each round's batch has one accountable owner, so no article falls into the gap between "everyone's" and "no one's."
- Run the editorial service scope. The pass prunes the obsolete, repairs broken references, merges duplicates, and refreshes stale detail — a defined set of edits, not a vague "review."
- Protect the bandwidth. A fixed slice of time is reserved and defended so the rotation survives busy periods instead of being the first thing dropped.
Tuning parameters¶
- Rotation period — how long a full sweep of the corpus takes. Shorter keeps content fresher but consumes more standing capacity.
- Batch weighting — how much more often high-traffic or volatile sections recur versus the long tail. Heavier weighting fixes what readers actually hit; flatter coverage catches quiet rot.
- Reserved bandwidth — how many hours per cycle are protected for cleanup. More sustains the sweep but taxes delivery.
- Steward assignment — rotating owners versus fixed subject-matter owners. Rotation spreads knowledge; fixed ownership deepens it but risks blind spots.
- Scope per pass — light (dead links, typos) versus deep (rewrite, consolidation). Deeper restores more but slows the sweep.
When it helps, and when it misleads¶
Its strength is turning "someone should clean up the docs" into a standing, owned, funded rotation — so an information corpus is held current by rhythm rather than by the accident of a reader complaining. Reserved bandwidth and a named steward are exactly what keep the sweep from lapsing.
Its failure mode is sweep theater: the rotation runs, the batch is "reviewed," but the steward only skims — nothing stale is actually removed and no dead link is actually chased down — so the library looks maintained while decay continues underneath. It also tends to under-serve the long tail: batch weighting toward high-traffic pages can leave rarely-read but occasionally critical articles rotting for years. The classic misuse is treating the rotation as a place to add content (writing new articles on the cleanup slot) rather than to prune and repair, so the corpus grows faster than it is groomed. The guarding discipline is to define the pass as removal-and-repair with a concrete edit checklist, and to let a steward mark an article for deletion rather than feeling obliged to keep everything.
How it implements the components¶
This rotation fills the do-the-cleanup-with-owned-capacity side of the cadence:
service_scope— the editorial actions each pass performs: prune obsolete pages, repair broken references, merge duplicates, refresh stale detail.maintenance_capacity_budget— the protected, reserved slice of time each cycle that keeps cleanup from being cannibalized by delivery work.maintenance_owner— the per-batch steward accountable for that slice this round, so no article is orphaned between owners.
It rotates cleanup through a corpus and reserves the bandwidth for it, but it neither models the decay curve nor formally certifies a repair: the trust- or wear-erosion deterioration_model is Relationship Maintenance Ritual, and the condition-standard post_maintenance_verification that signs off a restored asset is Infrastructure Upkeep Cycle.
Related¶
- Instantiates: Preventive Maintenance Cadence — it is the owned, funded rotation that keeps an information corpus from silently going stale.
- Sibling mechanisms: Infrastructure Upkeep Cycle · Preventive Inspection Schedule · Health Checkup Schedule · Policy Review Cycle · Relationship Maintenance Ritual · Refactoring Cadence
Editorial Notes¶
Form Classification¶
Form family: Intervention, Treatment & Transformation
Rationale: The mechanism directly changes documentation collections by removing stale guidance and repairing broken references.
Nearest alternative: Protocol, Workflow & Routine — Cleanup rotates through a workflow, but success is the transformed knowledge-base condition.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Library & Information Science
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Collection maintenance and digital curation developed systematic review, withdrawal, link repair, and corpus-wide coverage.
Related originating lineages:
- Organizational & Management Science — Preventive-maintenance and stewardship practice shaped named ownership, cadence, and protected capacity.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Independent reviewer agreement; high confidence.
Notes¶
Distinct from a Refactoring Cadence, which grooms code structure under a behavior-preserving contract: this rotation grooms an informational corpus, where "behavior" is what the content tells a reader, and the risk is not a broken build but confidently wrong guidance. The two rhyme but maintain different substrates.
[n1] Link rot (and the broader content decay it belongs to) is the gradual process by which references stop resolving and documentation drifts out of sync with the thing it describes — the specific, silent deterioration a cleanup rotation exists to counter. ↩