Calibrate buffers by consumption rate — above 80% means too small, below 20% means too large
Track weekly buffer consumption rate—if consistently consuming more than 80% of buffer, increase buffer size; if consistently consuming less than 20%, buffer can be tightened.
Why This Is a Rule
Buffers need calibration, not just existence. A buffer that's consistently consumed entirely (>80%) isn't large enough — it's barely absorbing the variance, and the next slightly-above-average disruption will blow through it. A buffer that's consistently untouched (<20%) is too large — it's reserving capacity that could be used for productive work without any protection benefit.
The consumption rate is the calibration signal. Tracking how much of your weekly buffer gets used provides objective data for right-sizing. The ideal consumption rate is 40-60% — the buffer is regularly used (confirming it provides value) but rarely exhausted (confirming it provides enough protection).
This is the same principle as inventory management in operations: too much buffer is waste (capital tied up in unused inventory), too little buffer causes stockouts (disruptions cascade because there's no cushion). The consumption rate tells you which direction to adjust.
When This Fires
- During weekly retrospectives when evaluating schedule performance
- After 3-4 weeks of buffer tracking when you have enough data to see patterns
- When buffers feel "too generous" or "never enough"
- Any ongoing calibration of your capacity planning system
Common Failure Mode
Not tracking buffer consumption at all — just having a buffer and hoping it works. Without consumption data, you can't tell whether the buffer is too large (wasting capacity), too small (providing inadequate protection), or just right. The tracking is trivial: at week's end, note how many buffer hours were consumed vs. available.
The Protocol
Weekly: (1) Record how many of your planned buffer hours were consumed by overruns, unplanned work, or variance absorption. (2) Calculate consumption rate: consumed hours ÷ total buffer hours. (3) If >80% for 3+ consecutive weeks → buffer is too small. Increase by 20-30%. (4) If <20% for 3+ consecutive weeks → buffer is too large. Tighten by 10-20% and reallocate freed capacity to committed work. (5) Target: 40-60% consumption rate indicates well-calibrated buffer.