What Goes in the Denominator
The bottom of the ratio decides more than the top does. Seven things that may or may not count as available time.
The bottom of the ratio is where most of the argument lives, and it is also where most of the error lives, because the denominator is usually set once and never revisited while the numerator is scrutinised every month.
The measure discussed in “What Goes in the Denominator” becomes more credible when its source records are timely and reviewable. A professional-services team assessing inspect the time-recording approach for how employees cheat time trackers can connect time with projects and work categories, while documenting the definition separately so the dashboard does not silently decide what utilisation means.
Seven items sit on its boundary.
Additional context for “What Goes in the Denominator” is available from Singapore Ministry of Manpower working-hours guidance. Current legal, employment, tax and accounting requirements should be confirmed with a qualified adviser where necessary.
Annual leave
The first question. Leaving holiday in the denominator means a person who takes their entitlement scores lower than one who does not, by roughly eleven points on a standard entitlement.
Removing it is the right answer and it requires the leave system and the timesheet system to agree, which is a data problem rather than a policy one.
Sickness
Same logic, more awkward. Removing approved sick leave is consistent with removing holiday.
Firms that remove holiday and keep sickness are quietly measuring health, which is both unfair and in several jurisdictions a problem of a different kind.
Public holidays
Routinely forgotten. Eight or nine days a year, which is four points on the ratio.
They are almost always meant to be out and are frequently left in by a denominator expressed as a flat 2,080.
Part-time contracts
A flat denominator for everybody turns a three-day contract into a sixty per cent score at full commitment.
The correction is to pro-rate by contracted hours, which is arithmetic and is the single most common omission in this measure. It has its own note here.
Overtime
Hours worked beyond contract. If they are chargeable they go in the numerator, and if the denominator is contracted hours the ratio can exceed a hundred per cent.
A figure above a hundred should be read as a warning about workload rather than as an achievement, and most reporting does the opposite.
Formally committed internal time
A day a week on management, a fixed training commitment, a secondment. If the commitment is real, the time is not available for client work.
Removing it makes the ratio measure the use of genuinely free time. Leaving it in makes the ratio punish people for roles the firm assigned them.
The first weeks of a new joiner
Nobody is available for client work on day three. Including induction in the denominator gives new staff an unreachable figure for a quarter.
Most firms exclude a ramp-up period. Fewer write down how long it is, which leaves it to be negotiated per person.
The seven rulings
- Annual leave: out.
- Approved sickness: out.
- Public holidays: out.
- Part-time: pro-rate, always, in the report.
- Overtime: in the numerator, and flag figures over a hundred.
- Committed internal time: out where the commitment is formal.
- Ramp-up: out, for a stated number of weeks.
Why the denominator gets less attention
Because it moves slowly and looks like a constant. The numerator is built from this month's timesheet and feels like the live part.
But a wrong denominator is wrong for every person in every period, and it is wrong in a direction nobody notices because everybody is wrong together.
The check worth running once
Take the denominator for each person and compare it to the hours they could actually have worked: contract, less leave taken, less public holidays, less formal commitments.
Where the two differ by more than a few days, the figures for those people are not comparable with anybody else's.
The part that cannot be fixed by arithmetic
Whether a commitment is "formal" is a judgement, and a generous judgement lowers the denominator and raises the ratio.
So the denominator is negotiable in a way the numerator is not, which is a quiet route to a better figure and the reason the rule has to be written down and applied by one person rather than per manager.
Why a round number is a warning sign
A denominator of 2,080, or 1,800, or any other figure that looks designed rather than derived, was almost certainly chosen for convenience rather than built from contracts.
Derived denominators are untidy: 1,418 hours, 1,536, 1,284. The untidiness is evidence that leave, holidays and contracts were actually subtracted.
A report whose denominators are all identical across people with different contracts has not been built from the contracts, and every figure in it is wrong for anybody non-standard.
The annual rebuild
Contracts change, entitlements change, people move between roles. A denominator built once and carried forward decays.
Rebuilding it annually from current contracts takes an hour with the HR data and it catches the mid-year changes that nothing else catches.
Firms that do not rebuild typically find, when they eventually look, that several people have been measured against a contract they left two years ago. The resulting figures were wrong in a stable direction the whole time, which is why nobody queried them: a consistently wrong number looks more trustworthy than a correct one that moves.
What to publish
The denominator in hours, per person or per team, beside the ratio. Not the ratio alone.
Published denominators are checked. Hidden ones drift, and drift in the bottom of the ratio is indistinguishable from performance in the top.