Headcount and Utilization

The mechanics underneath the numbers

Booz Allen is, mechanically, a labor business: it hires cleared people, deploys them against funded contracts, and bills their hours. Its own filings say so — "substantially all of our operating margin is derived from direct client staff labor," and headcount growth "after attrition" plus backlog are the drivers of earnings growth [1]. In FY2026 that relationship reversed for the first time in years: the workforce fell about 12% while revenue fell 6%. This chapter traces how the three levers — headcount, utilization, and mix — moved through the shock, and why the gap between them matters more than the revenue line alone.

A business that bills its people

The revenue model is close to an identity: revenue is client-staff headcount, multiplied by how fully those people are utilized, multiplied by the rate their labor bills at, deployed against available backlog. Client staff are roughly nine-tenths of the total workforce — about 31,200 of 34,200 employees at the FY2024 year-end [2]. Because margin comes from labor rather than from pass-through subcontractor and materials costs (billable expenses "generally have lower margin"), the number of billable people and how busy they are drives profit [3].

For a decade the model ran one way: hire, deploy, repeat. Headcount climbed every year from FY2022 to a FY2025 peak, and revenue climbed with it. The relationship gives a clean read on operating discipline, because when the two lines diverge, utilization and mix are doing the work.

Sources: employees per FY2022–FY2026 10-K "Our People"/Item 1 disclosures [4] [5] [6]; revenue per FY2026 10-K results of operations [7].

FY2026: the workforce cut harder than the top line

The two lines broke apart in FY2026. The company employed about 35,800 people at the end of FY2025 and about 31,500 a year later — a reduction of roughly 4,300, or 12% [8] [9]. Revenue over the same period fell 6.4%, to $11,217 million from $11,980 million [10]. Headcount fell about twice as fast as the top line — the mirror image of every prior year, when revenue grew faster than the payroll.

Source: derived from reported employee counts and revenue, FY2023–FY2026 10-Ks [11] [12].

This was a decision, not attrition running its course. The FY2026 accounts carry $61 million of severance and related charges — $46 million in cost of revenue, $15 million in G&A — booked "as a result of cost management initiatives and the restructuring of the Civil business," and the company's actions to cut costs against "the slowed procurement and funding environment" [13]. The cut was concentrated where the demand shock landed: Civil, whose revenue fell more than 20% (Federal Demand). Management framed it as a rotation rather than a retreat — "aligning our workforce with our key growth areas, including ramping up hiring in critical mission and technology sectors" even as it shed staff elsewhere [14]. The severance charge and the net headcount decline confirm that the shedding outweighed the hiring in FY2026.

The workforce cut did more than reset the cost base; it also defended the year's reported earnings. Booz Allen's FY2026 net income fell only 9% while operating income fell 25% — a gap bridged by a tax charge that collapsed from $284M (23.3%) to $11M (1.3%) — and the FY2025 base it is measured against was itself inflated by ~$237M of one-time settlement benefits ($122M reserve release + $115M insurance recovery) while the FY2026 result was defended by a 12% workforce cut (~35,800 to ~31,500) and $61M of severance, so the earnings 'resilience' is largely constructed rather than operational (Financials and Estimates).

Sources: FY2026 10-K Note 1 and Results of Operations [15] [16]; severance note [17].

Utilization and revenue per head

Cutting people protects margin only if the people who remain stay busy. Utilization — the share of paid client-staff hours that are billable — is the variable that moves fastest between hiring cycles, and its own commentary tracks it through the cycle. In FY2022, revenue growth of 6.4% was "partially offset by reduced staff utilization" [18]; in FY2023, the 10.7% gain was driven partly by "higher staff utilization compared to the prior year" [19]. Through the FY2026 downturn the firm leaned on it hard: by the second quarter, management reported "customer-facing staff utilization significantly exceeding the prior year's levels" as it "intentional[ly]" tightened the workforce [20].

The output of a smaller, busier workforce is rising revenue per head. On total revenue, revenue per employee climbed every year and jumped again in FY2026 despite the revenue decline, because the denominator fell faster.

Source: derived from reported revenue and year-end employees, FY2022–FY2026 10-Ks [21] [22].

Total revenue overstates the labor content, because roughly a third of it is pass-through billable expenses. Stripping those out gives a cleaner productivity read, and the picture is the same, sharper. Revenue excluding billable expenses per employee rose from about $229,000 in FY2025 to about $247,000 in FY2026, up roughly 8% — and closer to 9% once the FY2025 figure is cleaned of the $113 million claimed-costs reserve reversal that inflated it (Reserves and Recoveries) [23].

Source: FY2026 10-K non-GAAP reconciliation (revenue excluding billable expenses) and year-end employees [24] [25].

This is why the underlying operating margin held near 9% through a falling top line rather than collapsing with it: the firm removed the labor cost as fast as the revenue disappeared, and worked the remaining base harder. The counter-fact sits in the same accounts. Utilization cannot rise indefinitely, and a productivity gain manufactured by cutting 4,300 people is a one-time step, not a compounding trend — a point worth holding against the FY2026 revenue-per-head jump before extrapolating it.

The AI question the model has not yet answered

Everything above describes the old algorithm: to grow revenue, deploy more billable bodies. Management now argues that algorithm is changing. Asked directly why revenue per employee rose 8% in a single FY2026 quarter, the CFO answered that outcome-based and fixed-price contracting, plus "empowering our incredible talent with AI-assisted coding and other tools," "should disconnect our growth algorithm a little bit from headcount growth" and push revenue per employee up [26]. It is the same claim the growth story rests on (The Growth Engine), viewed from the cost side.

The claim cuts both ways. If it holds, Booz Allen can grow revenue and margin without a proportional hiring bill — the productivity gains of FY2026 become structural rather than a downturn artifact, and the trough read gains support. If it does not, the FY2026 revenue-per-head jump was simply a busier workforce during a hiring pause, and growth reverts to needing bodies the moment demand recovers. The evidence to date is a single year of falling headcount and rising productivity, which is consistent with both stories; a full year of the workforce growing again alongside rising revenue per head would be the first genuine test.

There is a moat cost to weigh against the efficiency, and it is specific to this firm. The 4,300 people who left were disproportionately experienced, and a large share of the workforce holds active security clearances — the asset the competitive case is built on (The Cleared Moat). Cleared, deployable staff cannot be rehired on a quarter's notice; clearances lapse and reinstatement takes time. Flexing the workforce down protects this year's margin, but it draws down the inventory of exactly the people the firm competes on, and rebuilding that inventory into a recovery is slower and costlier than the severance line suggests.

What to watch

Three things will show whether FY2026 was disciplined cyclical management or the start of a structurally smaller firm. First, whether headcount stabilizes and turns back up in FY2027 — a continued decline would argue the business is rebasing rather than pausing. Second, whether utilization and revenue per employee hold their FY2026 gains once hiring resumes; a give-back would confirm the productivity step was a one-time squeeze. Third, whether the "disconnect" between revenue and headcount actually appears in the numbers — revenue growth outpacing headcount growth over a full year, on outcome-based and fixed-price work, rather than in a single downturn quarter. Until then, headcount and utilization remain the clearest real-time gauge of whether demand is returning.