How Fast a Figure Can Honestly Improve
Real improvement in utilisation comes from four sources, all slow. Anything fast is reclassification.
Six ways the figure moves, with how long each one takes
No change in sold work; the client code's hours rose 14%.
Training bookings fell to nil, then doubled in the next quarter.
Real, and worth two to four points over a year.
Limited by the sales cycle, whatever the pressure.
Raises the blended figure and changes what the firm is.
Alters what an hour means rather than how many there were.
The two fastest routes change no work at all, and the four real ones cannot operate inside a quarter. So speed itself is evidence: a six-point improvement in one period came from the first two. One firm's own record, not a statement about anybody.
Utilisation can improve for six reasons. Two are immediate and are not improvements. Four are real and none of them works inside a quarter.
The measure discussed in “How Fast a Figure Can Honestly Improve” becomes more credible when its source records are timely and reviewable. A professional-services team assessing consult the detailed workflow notes for how to calculate idle time can connect time with projects and work categories, while documenting the definition separately so the dashboard does not silently decide what utilisation means.
That asymmetry is the most useful diagnostic in this collection, because it means speed itself is evidence.
For an independent reference relevant to “How Fast a Figure Can Honestly Improve”, consult AICPA audit and assurance resources. Compare the general material with the organisation’s own definition, contract and decision process.
The immediate pair
Reclassifying ambiguous hours to client codes, and deferring non-billable work to a later period.
Both move the figure within one reporting cycle. Neither changes how much client work the firm did or what it was paid for it.
Why deferral looks like improvement for two quarters
Because the deferred work does not vanish; it accumulates. Training not done, improvement not made, administration not filed.
The figure improves, then plateaus, then the backlog arrives as an unplanned cost somewhere else. The connection is rarely drawn.
Real source one: scheduling
Matching available people to available work faster. Shorter gaps between engagements, fewer people idle while work waits elsewhere.
Genuine, and it takes two to three periods to show, because it depends on changing how allocation decisions are made and who makes them.
Real source two: pipeline
More work sold, or work sold with less gap between engagements.
Three to four periods minimum, because the sales cycle is what it is. A figure that improves faster than the pipeline did was not improved by the pipeline.
Real source three: seniority mix
Juniors spend more of their week on chargeable work than partners. A firm that hires at the bottom raises its blended figure.
A year or more, and it changes what the firm is, which may or may not be wanted. It is also a reason blended figures are hard to compare.
Real source four: commercial model
Moving work from fixed fee to time and materials, or the reverse, changes what an hour means. So does changing what is included in a scope.
Slowest of all, and it affects the measure's meaning rather than just its level.
Using the timing as a test
A six-point improvement in one quarter has to come from the immediate pair, because nothing else works that fast.
This is not an accusation. It is arithmetic about lead times, and it can be checked against the non-billable category data.
The check
Look at what fell. If chargeable hours rose while total recorded hours stayed flat, something moved between categories.
If total recorded hours rose too, people worked more, which is a different finding and worth knowing about for other reasons.
The honest improvement plan
Name which of the four sources is being used, state the expected size, and state the period over which it should appear.
A plan that does not name a source is a plan to apply pressure and see what happens to the number.
Expected sizes
Scheduling: two to four points over a year in a firm that has not worked on it. Pipeline: depends entirely on demand. Mix: a point or two a year. Model: variable and slow.
Numbers much larger than these are worth examining before they are celebrated.
The two-quarter rule
Any movement larger than a point or two inside two quarters came from recording or deferral, because nothing else has a shorter lead time than that.
This is not a judgement about the people involved. It is a statement about how long it takes to sell work, change a schedule, alter a seniority mix or reprice a commercial model.
Lead times are facts about the world. A claimed improvement that outruns them is a claim about something other than the thing it is attributed to.
What to do with a fast improvement
Look at the non-billable categories before celebrating it. If something fell by half, that is where the improvement came from.
The useful response is not to challenge anybody but to ask whether the category that fell was one the firm needs to see.
Where it was, the finding is about the target rather than about the recording, and the remedy is a carve-out. Where it was genuinely administrative slack, the improvement may be real and small, and labelling it accurately is worth more than the two points it represents, because the label is what prevents the next quarter's target being set from a misread baseline.
Naming the source in advance
Any initiative aimed at this figure should state which of the four real sources it uses and what size of effect to expect over what period.
Three sentences, written before the work starts. They make the result checkable and they rule out the fast levers by construction.
An initiative that cannot name its source is a plan to apply pressure, and pressure has a known and documented effect on this measure.
What to tell a board
That the fast levers are reclassification and deferral, that neither produces revenue, and that the real levers run on multi-quarter timescales.
Boards accept this more readily than most managers expect, and it removes the pressure that produces the fast movement in the first place.