Workspace utilisation analysis

Workspace utilisation analysis is the study of how much of a workspace available capacity is actually used over a period, expressed as a percentage of what was on offer. If a team has 40 desks and 24 are occupied on an average day, desk utilisation is 60 per cent. If a consulting team has 1,000 available hours in a week and bills 720 of them, utilisation is 72 per cent. The same logic applies to rooms, equipment and compute.

Utilisation matters because unused capacity is a cost you carry whether or not you use it. Rent, salaries and reserved infrastructure are paid in full regardless of how much gets consumed. Analysis turns a flat percentage into a decision: release space, rebalance work, or invest in more capacity because the slack has run out.

Divide the capacity you actually used by the capacity that was available, then multiply by 100. The art is in defining both terms honestly. Available capacity should reflect realistic availability, not a theoretical maximum that ignores holidays, maintenance or non-working hours. Used capacity should count genuine use, not time merely booked and then left idle.

Worked example: a 12-person delivery team has a 480-hour week at full availability. Two people are on leave, removing 80 hours, so realistic available capacity is 400 hours. The team logs 320 productive hours. Utilisation is 320 divided by 400, which is 80 per cent. Measured against the theoretical 480 it would read 67 per cent, which would understate how stretched the team actually is.

1. 1

   Define available capacity

   Set the realistic maximum for the period, net of leave, maintenance and non-working time.

2. 2

   Measure capacity used

   Capture genuine consumption: occupied desks, billed hours, room minutes in use or compute consumed.

3. 3

   Choose the period

   Fix a consistent window, such as a working week, so peaks and troughs are not hidden by averaging.

4. 4

   Divide and express as a percentage

   Used over available, times 100, computed the same way every period for a comparable trend.

A single utilisation percentage hides why capacity sits idle or runs hot. To act on it you have to break the number into the things that move it: how much capacity exists, how demand arrives, and how well demand is matched to supply. A room at 40 per cent might be unbooked, or booked and abandoned, or simply the wrong size for the meetings people hold. Each cause has a different fix.

KPI Tree decomposes utilisation into these causal drivers so the analysis becomes a set of owned, fixable nodes rather than one figure on a dashboard. When utilisation drifts, the push goes to the accountable owner of the driver that moved, not to a generic facilities or operations inbox. That is the gap between a dashboard that reports a number and a system that turns the number into the next action.

Healthy utilisation depends on what is being used and how costly it is to run short. Billable service teams target high figures because idle hours are lost revenue. Physical space targets lower figures because a buffer protects flexibility. Treat these ranges as starting points and adjust for how expensive a full or empty state is in your context.

Improving utilisation means closing the gap between what you pay for and what you use, in whichever direction the analysis points. Sometimes that is lifting a low number by routing demand better. Sometimes it is relieving a number too high by adding capacity before something breaks. Work the driver, then confirm the headline followed.

1. 1

   Using theoretical capacity

   Measuring against a maximum that ignores leave and downtime understates how stretched a workspace really is.

2. 2

   Counting booked as used

   A reserved desk or room that nobody turns up to is idle capacity, not utilisation.

3. 3

   Chasing 100 per cent

   Pushing every workspace to full removes the buffer that absorbs spikes and protects flexibility.

4. 4

   Averaging away the peaks

   A monthly average can read healthy while every Tuesday afternoon is gridlocked; analyse the distribution, not just the mean.