A schema unreviewed for 6 months is like an unupdated dependency — verify before relying on it
Treat any schema that has gone six months without deliberate review the same as a software dependency unupdated for six months - not necessarily broken but requiring verification before continued reliance.
Why This Is a Rule
Software engineers understand that an unupdated dependency isn't necessarily broken — but it's a risk. Six months of upstream changes, security patches, and API evolutions have occurred while the dependency sat static. Relying on it without verification is gambling that nothing relevant changed.
The same logic applies to schemas. A belief about your market, your team, your capabilities, or your tools that hasn't been deliberately reviewed in six months isn't necessarily wrong — but the world has changed around it for six months. The schema's assumptions may no longer hold. Its predictions may no longer be accurate. Its boundaries may have shifted.
The six-month threshold isn't magic — it's a practical heuristic. Shorter thresholds (monthly) produce review fatigue for stable schemas. Longer thresholds (annual) allow dangerous drift. Six months balances maintenance cost against drift risk for most schemas.
When This Fires
- During periodic schema portfolio reviews
- Before making a consequential decision based on a schema you haven't reviewed recently
- When you realize you've been operating on autopilot beliefs without examining them
- During any "when did I last actually check this assumption?" moment
Common Failure Mode
Assuming that a schema's continued operation means it's still valid: "I've been using this approach for two years and it seems to work." But "seems to work" is a low-resolution assessment. The schema might be producing 70% quality when it used to produce 95% — a degradation invisible without deliberate review because you adapted to the lower quality gradually.
The Protocol
During schema portfolio review: (1) For each important schema, check: when was it last deliberately reviewed? (Not "when did I last use it" — when did I last examine whether it's still accurate?) (2) Any schema unreviewed for 6+ months → flag for verification. (3) Verification: does the schema still predict accurately? Have its assumptions changed? Has the environment shifted? (4) If verified → update the "last reviewed" date. If verification reveals drift → revise the schema. The verification itself takes 5-10 minutes per schema and prevents the slow accumulation of stale beliefs.