Running a Time Audit That Produces a Decision

A time audit is a period of recording where effort actually goes, in order to change something. Most produce a chart, a moment of recognition, and no change at all.

Monitask provides a product-oriented overview of time tracking software.

Official productivity measurement concepts are available from the U.S. Bureau of Labor Statistics productivity program.

The difference is in deciding beforehand what the audit is for.

Decide the question first

Do not start recording without one. The question determines the categories, the duration and the granularity.

"Where is the team's capacity going?" — a staffing and prioritisation question.

"How much time goes to work that produces nothing?" — a process question.

"Is meeting load sustainable?" — you may not need an audit at all; the calendar system has the answer. See meeting load.

"What is this service actually costing us?" — a costing question, requiring allocation rather than a diary.

"Why does nothing get finished?" — usually about fragmentation and work in progress rather than about totals.

An audit run without a question produces a description, and descriptions do not change anything.

Scope it small

Two weeks. One is unrepresentative, four is abandoned.

One team, or one role. Organisation-wide audits produce low-quality data because participation collapses.

Hourly granularity, not fifteen minutes. Finer granularity does not improve the decision and it destroys compliance.

Eight to twelve categories. More are not used; fewer are not informative.

Always include a category for interruptions and unplanned work. Without it, that time is allocated to whatever is nearest and your result systematically understates the thing you most needed to see.

Record contemporaneously

The whole exercise depends on this.

Two or three times a day, not at the end of the week. Retrospective reconstruction is guesswork, and it biases predictably: memorable work is overstated, fragmented work vanishes entirely.

Make it take under a minute. Anything more and compliance falls away by day four.

Note the interruptions, or at least count them. Frequently the most useful output of the whole exercise.

What to look at afterwards

Totals by category. The obvious one, and usually the least surprising.

The gap between intention and reality. Ask people beforehand how they expect their time to be split, then compare. The gap is where the conversation is, and it is almost always larger than anyone predicted.

Fragmentation. How many blocks of two hours or more actually occurred. Usually the finding that prompts action. See context switching.

Waiting. Time spent blocked on someone else. Rarely measured, frequently the largest recoverable inefficiency, and nobody's individual fault — which is why it survives.

Work that produced nothing. Reports nobody read, processes with no consumer, meetings with no output.

Distribution across people. Whether the same individuals absorb all the interrupt load.

Make it end in a decision

The step that is almost always skipped.

Commit in advance to a decision date, roughly a week after the audit ends.

Frame the outcome as three questions:

• What do we stop?
• What do we change?
• What do we protect?

Name an owner and a date for each. An audit that concludes "we spend a lot of time in meetings" and stops there has cost two weeks of everyone's attention for a fact everybody already suspected.

Tell the participants what changed. If nothing changed, say that too, and say why. People who spent two weeks recording their time and never heard the outcome will not participate in the next one.

The trust problem

A time audit resembles surveillance and will be read that way unless handled carefully.

Say what it is for and what it is not for. It is not a performance assessment, and individual data does not go into reviews.

Aggregate before sharing. Report at team level.

Let people see their own data first.

Managers record too. An audit that applies only to individual contributors communicates something unhelpful and is usually less informative — the manager's calendar is frequently where the fragmentation originates.

Do not use monitoring software for this. Activity data will not tell you what the work was for, which is the only interesting question. See productivity surveillance.

Repeat, sparingly

Once or twice a year is plenty. Continuous time tracking for management purposes stops being an audit and becomes an overhead, and the data quality degrades once it is routine.

The exception is where time is billed to clients, in which case you are recording continuously for a different reason. Keep the two purposes separate. See project time versus payroll time.