Every firm reports a utilisation figure. Almost none of them are measuring the same thing.
Utilisation, read as a number that looks back
Fifty notes on utilisation and billable efficiency. Not a guide to raising the number — a working account of what it measures, why six defensible denominators for one person's year give answers thirty-seven points apart, what happens to the week once anybody is appraised on it, and what has to be read beside it for it to mean anything at all.
The measure discussed in “Every firm reports a utilisation figure. Almost none of them are measuring the same thing.” becomes more credible when its source records are timely and reviewable. A professional-services team assessing employee monitoring software for employee monitoring software can connect time with projects and work categories, while documenting the definition separately so the dashboard does not silently decide what utilisation means.
Every figure is an illustration of one firm’s records, not a benchmark or recommendation. Definitions, employment rules and accounting requirements differ and should be reviewed for the organisation concerned.
Additional context for “Every firm reports a utilisation figure. Almost none of them are measuring the same thing.” is available from GitLab remote-work handbook. Current legal, employment, tax and accounting requirements should be confirmed with a qualified adviser where necessary.
The figure one firm reported to its partners last quarter
78.4%
Numerator: Hours booked to client codes, including hours later written off
Denominator: Contracted hours less approved leave
The same quarter read net of write-offs, against attended hours, gives 61.2%. Neither number is wrong. Only one of them was published, and the report did not say which.
One person, one year, 1,250 chargeable hours, six denominators
80.1%
88.2%
60.1%
78.0%
96.9%
92.6%
Six defensible denominators, one unchanged year of work, and answers between 60.1 and 96.9 — a spread of 36.8 points. That gap is larger than any performance difference a firm is likely to be arguing about, and nothing in the arithmetic chooses between them. One firm's own year, not a benchmark.
Eight responses to a utilisation target, with how fast each one works
The client code's hours rose 14% with no change in sold work.
Training bookings fell to nil in one quarter and doubled in the next.
Two improvement projects lapsed for want of volunteers.
Write-offs rose from 6% to 11% of booked hours.
Proposal hours fell by half before anybody noticed.
Review hours vanished from the record; the reviews partly continued.
Raised in exit interviews. Never counted.
24 of 78 people landed in a single one-point band.
Four of the eight were measured, three observed and one only ever claimed. The ratio rose seven points over two quarters, and not one of the eight involved doing more work for a client. One firm's own record, and a statement about a measure rather than about anybody in it.
01 / 07
What the number measures
A ratio of hours to hours. What it is evidence of, and the three things people read into it that it cannot support.
02 / 07
Six ways to compute it
The numerator and the denominator are both choices, and between them the choices are worth thirty-seven points.
03 / 07
The moment it becomes a target
What changes when the number somebody produces is the number they are judged on.
04 / 07
A firm-level ratio on one name
The figure is mostly structural, and it is applied to individuals. What follows for juniors, part-timers, returners and the bench.
05 / 07
What people actually do
Eight responses, named, with the timescale of each and what it does to the number and to the firm.
06 / 07
What the ratio cannot see
Realisation, yield, quality, the unbilled work that wins the next engagement, and the cost of a week with no slack in it.
Software comparisons
Tools behind utilisation and client-work records
Three independently structured comparisons based on direct review, official provider information and practical implementation checks.
07 / 07
What to read beside it
The four figures that make the ratio mean something, the composition that makes it diagnostic, and six rules for running it without the damage.
The short version
Publish the definition in the same line as the number
Utilisation is a quotient whose numerator and denominator are both choices, and six defensible pairs of choices give answers thirty-seven points apart. Published without its definition it cannot be compared with anything, including last year. Used as a target it improves and stops measuring, because every route to a higher figure available to an individual is cheaper than doing more work for a client. Published with its definition, at team level, beside what it cannot see, it is a usable instrument and a dull one.