A time audit is a deliberately short, intensive exercise: for one week, log what you're doing in real time, in as much granularity as you can tolerate, without editing the categories to look better in hindsight. It is different from ongoing time tracking in one important way — the goal isn't to build a permanent system, it's to get an honest baseline before deciding whether a permanent system is even needed, and what it should measure.
Why a week, and why real-time
A single day is too short to be representative — most people have at least one unusually good or unusually chaotic day per week, and a one-day audit will overweight whichever kind of day it happened to fall on. A full week captures the normal mix, including the parts people tend to forget when asked to estimate their own time from memory: the commute, the context-switching between unrelated tasks, the time lost to a broken tool or an unclear instruction.
Real-time logging — noting the start of an activity as it starts, rather than reconstructing the day afterward — matters more than the format used to record it. A rough note in any app, written within a few minutes of switching tasks, is more accurate than a beautifully categorized timesheet filled in from memory at 6 p.m. For a practical software example related to ways employees can trick time trackers, visit this page.
What people consistently get wrong on the first audit
The most common surprise isn't how time was spent on any single task — it's how much of the week was spent in a category most people don't think to log at all: switching between tasks, waiting for a reply before continuing, or reorienting after an interruption. When forced to categorize every few minutes rather than summarize the day afterward, that category is often the largest single line item, and it's almost never visible in a normal end-of-week self-report.
- Log the switch itself, not just the destination task — 'stopped writing to answer a message' is more informative than silently starting a new 'messages' entry.
- Resist tidying categories mid-week. If the honest label is 'unclear, thirty minutes,' write that rather than force-fitting it into an existing category.
- Include the time before and after the 'real' task — set-up, wrap-up, and the mental reorientation after an interruption are usually the most underestimated categories.
- Do the audit during a representative week, not a vacation week or an unusually quiet one — the point is a normal baseline, not a best case.
Turning a week of logs into one decision
The audit is only useful if it produces a specific change afterward — otherwise it's just a week of extra effort with no lasting output. Common outcomes include discovering that a recurring meeting consumes far more of the week than its perceived importance justifies, that a specific hour of the day is reliably lost to a predictable interruption that could be scheduled around, or that a task believed to take 'about an hour' consistently takes closer to two.
A one-week audit, repeated once or twice a year rather than maintained continuously, tends to produce most of the insight of ongoing time tracking with a fraction of the long-term maintenance burden — useful for people who want the data without committing to a permanent logging habit. For additional background on this subject, consult the ILO’s working-time resources.