The Utilization Number That Nobody Believes
A 400-person consulting firm reports firm-wide utilization of 78%. Delivery leaders say the number is inflated. Investigation shows: billable time is booked against client projects, but internal 'shadow' work for a large client (pre-sales solutioning, unpaid pilot support) is coded to an internal project and excluded from the numerator. Bench staff are counted in the denominator only from their availability date, not from the date they rolled off. Three regions use different definitions of a standard working day (7.5, 8, and 8.5 hours). One region books training as billable if it is client-specific.
1. Reconstruct a defensible utilization definition and explain how each of the four distortions changes the reported figure.
Read the worked answer
Utilization is a ratio, so every distortion either inflates the numerator or shrinks the denominator. Name the direction for each.
Shadow work excluded from the numerator. Unbilled client-facing work is real consumption of capacity. Excluding it inflates utilization by making the firm look busier on billable work than it is. Remedy: either include approved client-facing non-billable work in a separate 'client-facing utilization' measure, or report a chargeability ratio alongside utilization so the gap is visible. Do not silently drop it.
Bench counted from availability date. Starting the denominator at availability rather than roll-off shrinks the denominator and inflates utilization. Remedy: count people from the date they become available, regardless of internal transition time, or explicitly define a transition allowance and apply it uniformly.
Inconsistent standard day (7.5/8/8.5). Hours-based utilization across regions with different denominators is not comparable. A region with a 7.5-hour day shows higher utilization for the same hours worked. Remedy: convert to a common basis - either standardize the standard day or compute utilization in FTE-months rather than hours.
Training billed as billable when client-specific. This is a definitional choice, not an error, but it must be consistent. If one region does it and others do not, cross-region comparison is invalid. Decide the rule centrally and apply it everywhere.
A defensible definition states: numerator (which time categories count), denominator (who counts and from when), time basis (hours versus FTE, standard day), and exclusions. Publish it and version it, because redefinition mid-year invalidates trend.
2. Leadership wants a single headline utilization target of 80% applied to all regions. What is your recommendation and why?
Read the worked answer
A single 80% target applied uniformly will produce the wrong behavior, because the regions do not have the same mix of work or the same denominator. The recommendation is a common definition with differentiated targets, plus a secondary measure that prevents gaming.
Why uniform targets fail here. With standard days of 7.5, 8, and 8.5 hours, an 80% target means different absolute hours in each region. A region with a large share of fixed-price work has different billable-time dynamics than a time-and-materials region. And a region with a big shadow-work load for the anchor client will hit 80% only by under-recording that work - which is exactly the distortion already present.
What to do instead.
- Standardize the definition firm-wide (numerator, denominator, time basis) so the metric is comparable.
- Set targets by region and by role band, anchored to the demand plan and the sold backlog, not to a round number.
- Add a paired measure - for example, client-facing utilization including approved shadow work, and a bench-to-demand ratio - so improving the headline number cannot be achieved by reclassifying time.
- Audit coding quarterly on a sample basis; if shadow work for the anchor client is material, either bill it or formally recognize it as a cost of the relationship in the account plan.
The governance point. The moment a target can be met by moving hours between codes, it will be. Pair every efficiency metric with a quality or integrity metric.