Record-low utilization is a staffing-data problem

A point of utilization is billable time a firm already pays for. In 2025 the average firm billed 66.4% of its consultants’ available hours, down from 68.9% a year earlier.

SPI Research treats 75% as the target. The gap is 8.6 billable hours in every 100 a consultant is available.

The usual reading is that clients are buying less. The same benchmark says otherwise.

The work is there; the staffing view isn’t

SPI’s 2026 benchmark covers more than 500 firms and 2025 data. Project margins rose to 37.7% from 35.9%, and bids won climbed to 48.1% from 47.3%.

Pipeline coverage rose to 175% of the quarterly bookings forecast. Revenue per billable consultant grew about 6%, to $210,000.

A firm with a thin pipeline doesn’t win more bids and raise its margin in the same year. These firms had work to sell and sold it at a better price. Then they billed less of their people’s time.

That points inside the firm. Utilization falls when the right person can’t be found for the work in time, so the work waits and the person sits.

Bench and overtime on the same team is a data signal

We call it the hidden bench: consultants who are free next week, but whom nobody staffing a project can see. It shows up in two places at once.

The COO sees bench next month and overtime this week on the same practice. The CFO sees it later, in write-downs and unbilled work in progress.

Three data gaps create it. First, tentative holds that never became confirmed bookings, or never got released.

Second, skills and certifications recorded once, at hiring, and not since. Third, resource requests that live in email and hallway conversations, so the resource manager sees only part of the demand.

Each one makes a free consultant look busy, or a busy one look free. The practice lead then staffs the person they know, which is rational and expensive.

One definition of utilization comes before any dashboard

Utilization is billable hours divided by available hours, and firms disagree on “available.” Some use standard hours; others subtract holidays, training, and approved leave. Two offices using different denominators can’t be compared, and a firm-wide number becomes an argument.

The CFO and COO should agree on the definition before anyone builds a report. Then write it down, with the reason, so the next acquisition or new office uses the same one.

Bookings people trust are the whole fix

The practical target is one view of supply and demand, by skill and by week. Every request goes in as a request.

Every hold is either confirmed or released by a set day. The practice leads who know each person’s skills keep them current.

On Microsoft Dynamics 365 Project Operations↗, that means resource requests, tentative and confirmed bookings on the schedule board, and skills on each person’s profile. Microsoft has added a staffing timeline view and a team-member staffing view, both in preview since mid-2026 and planned for general availability in October 2026.

The software is the easy part. The harder part is a weekly staffing meeting that works from the view.

Good booking data pays twice

Confirmed bookings show who is free, and they also feed time entry.

Agents at work: the Project Operations Time Entry Agent, in paid public preview, drafts each team member’s time from their assignments and bookings for the person to review and submit. Where no bookings exist, it can only copy the previous week.

So a firm that keeps bookings current gets better staffing decisions and more accurate timesheets from the same data. Who works with which client, and who grows into the next role, stays with the practice leads.

Some firms have a demand problem after all

This argument doesn’t hold everywhere. If your pipeline coverage fell and your win rate dropped, the problem is in sales, and better staffing data won’t fill the bench. Check those two numbers first.

Some bench is also a choice. A firm training people on new methods, or reshaping work as generative AI enters projects, may carry bench on purpose.

SPI found 27.1% of projects used generative AI in 2025. Count that time as an investment in the firm.

Small firms are different too. Under about 50 people, the resource manager usually knows everyone, and a lighter tool is the better answer.

What firm leaders should do this quarter

Agree on one utilization definition before any dashboard. The CFO and COO sign off on the denominator and write down why.

Put every request and every hold in one place. A hold that isn’t confirmed or released by a set day expires.

Give skills to the practice leads. The people who staff the work keep skills and certifications current, every quarter.

Measure bench and overtime together, by practice. When both are high on one team, fix the staffing data.

Check pipeline coverage and win rate before you blame demand. If both are healthy, the fix is inside the firm.

How we know this

This point of view comes from Ludia’s delivery on Microsoft Dynamics 365 Project Operations. It also comes from running a firm whose consultants bill their own time.

(1) SPI Research, 2026 Professional Services Maturity Benchmark (2025 data), as reported by Deltek, 2026 PSO Benchmarks: Insights from SPI Benchmark Maturity Report (July 30, 2026). Source of: utilization 66.4% (68.9% in 2024), the 75% target, project margin 37.7% (35.9%), bids won 48.1% (47.3%), pipeline coverage 175%, and revenue per consultant of $210K (up 6%) across more than 500 firms.

(2) Certinia, Analyzing the 2026 SPI Research Professional Services Maturity Benchmark Report (2026). Source of: the lowest utilization in SPI’s history, 27.1% of projects using generative AI, and the 509-organization sample. The 8.6-hour gap is our arithmetic (75% minus 66.4%).

(3) Microsoft Learn, New and planned features for Dynamics 365 Project Operations, 2026 release wave 1 (updated Sept 3, 2026). Source of the staffing views and their dates.

(4) Microsoft Learn, Responsible AI FAQ for the Time Entry feature of the Time and Expense Agent (updated July 2026). Source of how the agent drafts time.