Utilization

Share of available hours logged to client work. Hours billed to clients divided by the team's available hours (the expected weekly hours on file, over the same period the entries cover). For this team right now, billed hours are divided by available hours taken from the expected weekly hours on file. How this metric is measured changed between the two most recent periods, so the difference between them would not be a real movement. Comparison resumes once two periods share a measurement basis.

Utilization is where payroll turns into income: the share of the hours you pay for that clients actually consume. The healthy level is about 60%. Well below it you are paying for idle capacity; well above it the team has no room for the work that grows the firm. It sits on this board against target ~60%.

Hours that exist and are not earning arrive two ways, and this number cannot tell them apart. Either the work is not there, so capacity freed by a pause or a finished project sits until somebody points it at sold work, or the work IS there and the hours are not being logged against it, in which case the capacity is being consumed and only the record is missing. The first is a redeployment problem and the second is a recording problem. They move this number in the same direction by the same amount and they need opposite responses, so only one of them is fixed by finding more work. The check that separates them is on this board and it has already run: Timesheet completion reads 78%, which is outside its band. On that reading some of these hours are earning and simply unrecorded, so the recording problem is live and reassigning against it would move people who were already busy. A share inside its band is not the same as every hour being logged, so read this as the check pointing rather than as the answer.

Before moving anybody, the check is already done. The check that separates them is on this board and it has already run: Timesheet completion reads 78%, which is outside its band. On that reading some of these hours are earning and simply unrecorded, so the recording problem is live and reassigning against it would move people who were already busy. A share inside its band is not the same as every hour being logged, so read this as the check pointing rather than as the answer. Where the hours really are idle the move is to point them at sold work this week. Where they are already earning and unrecorded, moving people fixes nothing and costs the week it takes to notice.

This is the cheapest possible ordering and it is the one people skip. Reassigning against a recording gap moves people who were already busy, costs the week it takes to notice, and leaves the actual problem in place. The check itself is one number on the same board. Where the hours genuinely are idle, the thing that makes the recovery expensive is the lag rather than the pause, so the same week matters more than the perfect assignment.

This closes no measured driver above, and does not claim to.

Keep enough sold work in front of the team that a pause has somewhere to send its hours the same week, and put the reassignment decision on a standing rhythm rather than leaving it to whoever notices.

A pause is not an unusual event and it should not need an unusual response. What makes it expensive is the lag between the work stopping and the hours being re-pointed, and that lag is a function of two things: whether there is sold work waiting, and whether anyone owns the decision. Coverage handles the first. A standing review handles the second. Neither is something the product decides for you, and both are cheaper than the recovery.

This closes no measured driver above, and does not claim to.

Idle capacity does not announce itself. It shows up later as income per person falling while headcount holds, which reads as a revenue problem and gets answered with a sales push when the hours to deliver that push were already sitting unused. Left long enough it becomes a hiring decision made against the wrong picture: the team looks fully committed on paper while the hours say otherwise. Where this lands is not this tile. It is AGI per FTE reads $192k against target ~$175K, which it is outside. It is already outside its band, so this is not a forecast. Part of it has happened, and that tile is where to go and look at it. Hours that exist and are not earning arrive as income per person, because the roster is the denominator either way. That tile is where idle capacity stops looking like a scheduling detail and starts looking like a revenue problem, which is also where it gets answered wrongly.