Noise and Signal in the Tempo of Work

The Problem of Self-Observation

One of the harder aspects of improving at knowledge work is that your own work is noisy. On a good day you produce twice what you produce on a bad day. The bad day feels like failure; the good day feels like the real you. Both experiences distort the signal.

Good observers of others' performance know to smooth across many observations before drawing conclusions. We grant this to our assessments of other people and fail to grant it to our assessments of ourselves.

The result is that most self-assessments of work capacity, output level, and pattern are based on too few observations and too much weighting of recent experience.

What Counts as Signal

Signal, in your own work patterns, is a systematic tendency that shows up consistently across different conditions, different times, different contexts. It is the thing that remains when you average out the variation.

Noise is the variation around that signal - the day-to-day fluctuation that reflects minor factors: sleep, mood, ambient energy, interruption density, the particular difficulty of the task at hand.

The challenge is that signal and noise feel identical in the moment. The bad day doesn't feel like noise - it feels like the real pattern. The good day doesn't feel like an upward fluctuation - it feels like what should be normal.

Separating them requires more data than people usually collect, held over longer time periods than people usually remember.

Types of Signal Worth Tracking

Not all patterns in work tempo are equally informative. Some types of signal are worth tracking because they indicate something actionable.

Energy rhythm: What times of day, day of week, and phase of a project do you produce your best work? This pattern, if stable across many instances, is signal. It suggests how to arrange work to match task demand to energy availability.

Depth variation: Some work requires deep, uninterrupted attention. Other work tolerates fragmented attention. Knowing your actual (not ideal) capacity for each type, and how it varies across the week, helps with realistic scheduling.

Recovery pattern: After intensive periods, how long does recovery take? This is often longer than assumed, and the underestimation of recovery time is a common cause of planning failures.

Degradation curve: How does your output change as a project extends beyond your usual project length? Signal here indicates when to expect declining returns on extended effort, and when to build in stops.

The Tracking Minimum

You don't need elaborate systems to separate signal from noise in your own work. The minimum is:

Periodically rating your productive output at the end of work blocks - not quality, just quantity and ease. Over weeks and months, patterns in these ratings reveal the signal.

Noting the conditions around high and low rated blocks. What is different about the good days? What is consistent about the bad ones?

Resisting the narrative you construct in the moment. The story "I am not working well this week" may be noise. "I have not worked well on Thursdays for the last three months" is more likely to be signal. Require the longer time series before accepting the pattern as real.

The result is a more accurate model of your own work capacity - one that is less subject to the distortions of recent experience and more useful for planning work you actually want to do.