Team Rotation Against Knowledge Silos

Knowledge silos break under turnover. The 6-month rotation that prevents siloing without slowing teams down.

Structure

Knowledge-silo rotation runs every six months across focus areas within a team. The successor takes over; the outgoing engineer covers questions for 30 days; the handoff forces documentation that would otherwise never get written. Six months is long enough to develop expertise, short enough that no single person becomes irreplaceable.

• Six-month rotation. Focus-area shift cadence per engineer. Not teams; focus areas within a team.
• Old focus area gets a successor. Named successor per rotation. Outgoing engineer covers questions for 30 days.
• Documentation required at handoff. Captured runbook update per rotation. Forces documentation the team would otherwise never produce.
• Named rotation steward. Per-team the responsible scheduler. Catches the “we forgot to rotate this cycle” failure mode.

Benefits

Rotation produces three structural benefits: no irreplaceable engineer, cross-pollination of patterns across focus areas, resilience to turnover when someone does leave. The bus factor stays above one by design.

• No single irreplaceable person. Multi-owner outcome per system. Bus factor stays above one.
• Cross-pollination of patterns. Engineers carry practices from one focus area to another. Good patterns spread.
• Resilience to turnover. Multi-owner-when-someone-leaves outcome. Departure does not orphan a system.
• Documented growth track. Captured rotation history per engineer supports career-growth conversations.

Limits

Three real limits keep rotation honest. Specialised areas (database internals, distributed-systems algorithms) cannot rotate freely; rotation costs first-month productivity per engineer; teams below five engineers carry too much overhead for the benefit.

• Specialised areas cannot rotate freely. Permanent owner with shadow pattern per area. Some expertise needs continuity.
• Rotation cost. First-month productivity dip per rotation. Annual cost is real and worth budgeting.
• Not for tiny teams. Five-engineer floor. Below five, rotation overhead exceeds benefit.
• Documented exceptions. Named “these engineers do not rotate” list per team. Catches scope creep on exceptions.