Substitute a faster noisy signal for a slower precise one — speed compensates for noise in feedback loops

Why This Is a Rule

When the true feedback signal is delayed — you won't know if your marketing strategy works for 3 months, if your exercise habit improves health markers for 6 months, or if your hiring process produces good employees for a year — waiting for the real signal means operating blind for the entire delay period. Any drift during that period compounds uncorrected.

A leading indicator — a faster, noisier signal that correlates with the delayed outcome — provides course-correction opportunities during the gap. Website traffic this week (noisy, fast) approximates marketing effectiveness this quarter (precise, slow). Workout consistency this week (observable, fast) approximates health outcomes this year (definitive, slow). Interview feedback quality this month (subjective, fast) approximates hire performance next year (objective, slow).

The key insight is that speed matters more than precision for feedback loops. A noisy signal that arrives daily and is 60% correlated with the true outcome enables 365 course-corrections per year. A precise signal that arrives annually enables one. Even with significant noise, the daily signal produces better outcomes through cumulative adjustment than the annual signal produces through one definitive reading.

When This Fires

• When the natural feedback signal for an activity takes weeks, months, or years to arrive
• When you're operating blind during a long feedback delay and want earlier warning
• When Measure feedback delay in actual time units before improving — unmeasured delays appear shorter than they are through habituation (measure delay) reveals a delay that can't be shortened structurally
• When designing measurement systems for long-term initiatives (strategy, health, learning)

Common Failure Mode

Waiting for the "real" signal because the leading indicator is "too noisy": "Weekly traffic numbers fluctuate too much — let's wait for quarterly data." By the time quarterly data arrives, three months of potentially misallocated marketing effort have passed. The weekly fluctuation, while noisy, still reveals directional trends within 2-3 weeks — enough for meaningful course-correction that the quarterly signal can't provide.

The Protocol

(1) When feedback delay exceeds one cycle of the activity (Daily activities need weekly feedback; strategic activities need monthly — longer delays let drift compound undetected), identify a leading indicator: what faster, observable signal correlates with the delayed outcome? (2) Validate the correlation: does the leading indicator actually predict the final outcome, at least directionally? Test with historical data if available. (3) Accept the noise: the leading indicator will fluctuate more than the true signal. This is the cost of speed. Smooth the noise by looking at trends across 3-5 readings rather than reacting to individual readings. (4) Use the leading indicator for directional course-correction, not precise calibration. "We're trending in the wrong direction" justifies adjustment; "we're 3.2% below target this week" doesn't justify panic. (5) When the true signal eventually arrives, calibrate your leading indicator against it to improve the correlation over time.