Agency Utilization Rate: The Formula, the Benchmarks, and the Trap

Utilization is the most quoted number in professional services and one of the most casually misused. Two agencies can report "78% utilization" and mean genuinely different things, because they divided by different denominators. A third can hit its utilization target every month and still lose money, because the hours were tracked, delivered, and then quietly written off at invoicing.

This guide covers the formulas, which denominator to use, what the number is actually diagnosing, and the trap of treating it as a target.

The formula, and the part everyone gets wrong

At its simplest:

Utilization rate = billable hours ÷ total hours × 100

The dispute is entirely in the denominator, and it is not pedantry - it changes the number by 10 to 15 points.

Billable ÷ total hours worked. Everything a person worked, including internal meetings, admin, and recruitment. Lower number, and the one that most honestly reflects "what fraction of the time we pay for is time we bill for".

Billable ÷ available hours. Total minus holiday, public holidays, sick leave, and sometimes agreed internal commitments. Higher number, and more useful for capacity decisions, because it measures against time that was genuinely available for client work.

Scoro's breakdown of billable utilization and Asana's guide to utilization rate both walk through the variants and where teams confuse them. The specific practical failure is comparing your internally-calculated figure against a published benchmark computed the other way, concluding you are under-performing, and pushing a team that was already at capacity.

Pick one, write down which one, and use it consistently. The definition matters more than the choice.

What good looks like - with a caveat

Benchmarks are worth reading and worth holding loosely. Mosaic's collection of professional-services utilization statistics is a reasonable starting point for where firms actually land rather than where they aspire to.

The caveat is that a benchmark is only comparable if it shares your denominator, your definition of billable, and roughly your role mix. A firm of billable consultants and a studio with a large non-billable strategy and account function will produce different numbers from identical underlying health.

Two things are more reliable than any published figure:

Your own trend. Utilization moving from 62% to 71% over two quarters tells you something real. Utilization of 71% compared to somebody's benchmark of 75% tells you almost nothing.

The spread across people. A team averaging 70% where everyone is between 65% and 75% is healthy. A team averaging 70% where two people are at 95% and two are at 45% is not healthy at all - and the average conceals exactly the problem you need to fix. Always look at the distribution, not the mean.

Why maximizing utilization is a mistake

The intuitive move once you can measure utilization is to raise it. This is a trap, for three reasons.

No slack means no capacity to absorb anything. A team at very high sustained utilization has no room for a scope change, a sick week, or a project that runs long - all of which are certainties, not risks. The first surprise turns into overtime, because there is nowhere else for it to go. This is why capacity planning deliberately plans to something below 100% of available hours: the slack is the shock absorber, not waste.

Non-billable work is not worthless work. Pitching, internal tooling, training, hiring, and the case study that wins the next three clients are all non-billable and all load-bearing. An agency that drives utilization to its maximum has, by construction, stopped doing them.

It corrupts the data. The moment utilization becomes a performance metric that people are judged on, timesheets start reflecting what is expected rather than what happened. You will hit your target and lose the ability to trust any number derived from time tracking - which is most of them.

Read it as a diagnostic:

• Consistently low across the whole team: you have sold too little, or your available-hours figure is wrong.
• Consistently very high: you are one surprise away from a bad month, and probably from someone resigning.
• Wildly uneven between people: your allocation is broken. This is the most fixable of the three and usually the most expensive to leave alone.

The number that matters more: realization

Utilization tells you how much of your team's time was billable. It does not tell you whether that time was billed.

Realization rate = invoiced hours ÷ billable hours tracked

Anything under 100% means work was done, logged as billable, and then written off - through discounting at invoicing, unbilled retainer overage, or scope absorbed without a change order.

High utilization with low realization is the most dangerous combination in a services business, and it is common. Everyone is at capacity, everyone is busy, the team feels stretched, and the revenue does not reflect any of it. Utilization alone will not show you this. It will show you a healthy business.

If you measure one thing after reading this, measure the gap between tracked billable hours and invoiced hours for a single month. That gap is real money, and most firms have never looked at it directly.

How to actually calculate it without lying to yourself

1. Fix the definitions first, in writing. Which denominator. What counts as billable. Whether agreed internal commitments come out of available hours. Ten minutes of agreement here prevents a year of incomparable numbers.

2. Derive available hours from data, not assumption. Nobody has 40 available hours a week. Take a month of tracked time and calculate the real ratio per role - for most agency roles the honest figure is between 25 and 32, and it drops with seniority as management responsibility grows.

3. Get the underlying time data honest. Every number here is derived from tracked time, so timesheets reconstructed on Friday from memory make the whole exercise decorative. Getting time tracking reliable is the prerequisite, not an optimisation.

4. Calculate per role and per person, then aggregate. The aggregate is for reporting. The distribution is where the decisions are.

5. Pair it with realization, always. Reporting utilization without realization is reporting activity and calling it performance.

Working out your own available-hours figure

Every utilization number depends on the denominator, so it is worth deriving yours properly once rather than adopting someone else's assumption.

Take three months of tracked time. Per person, total the hours actually worked and the hours logged against client projects. The ratio between them is your real billable proportion, and it will be lower than you expect.

Subtract the non-negotiables from the calendar first. Holiday entitlement, public holidays, and a realistic sick-day allowance. For a person on 25 days' holiday in a jurisdiction with nine public holidays, that is roughly 310 hours before anything else.

Then subtract structural internal time. Standups, all-hands, one-to-ones, recruitment, training, tooling, and the admin that surrounds client work without being billable to it. For most agency roles this lands between 15% and 25% of what remains.

Worked through, a full-time person typically has 1,400 to 1,500 available hours a year, against the 2,080 the naive calculation assumes. That gap of roughly a third is the single most common source of broken agency arithmetic - it inflates apparent capacity, deflates apparent cost, and makes every rate built on it wrong in the same direction.

Do this per role rather than once. A senior person carrying management responsibility might have 1,100 available hours; a junior specialist 1,550. Applying one figure across a mixed team distorts every project estimate that involves both.

Utilization by role, and why the targets differ

A single agency-wide target is close to meaningless, because different roles have structurally different capacity for billable work.

Delivery specialists - designers, developers, writers - should sit highest, commonly 70-85% of available hours. They have the fewest structural claims on their time.

Project managers run lower, often 50-70%, because coordination is partly billable and partly not depending on how you scope it. The decision about which side of the line project management sits on is one of the biggest single swings in a reported gross margin, and it should be made deliberately rather than by accident.

Account leads lower again, perhaps 40-60%. Much of their value is relationship work that is real and hard to bill.

Founders and directors are wherever the business needs them to be, and the honest version usually declines over time as the business grows - which is correct, and is worth planning for rather than discovering.

The practical consequence: an agency-wide utilization figure moves when your role mix changes, without anything about performance changing at all. Hiring a project manager mechanically lowers the average. That is not a decline; it is a different business. Track it per role, and compare like with like over time.

Setting a target you can actually defend

Most agencies adopt a utilization target from something they read. A defensible target is derived from your own economics, and the derivation takes about twenty minutes.

Work backwards from the margin you need. If your fully-loaded cost per delivery person is £55,000 a year and you want a 50% gross margin, that person needs to generate £110,000 of billable revenue. At £95 an hour, that is roughly 1,160 billable hours. Against 1,450 available hours, your required utilization is about 80%.

That number is now yours. It is defensible in a conversation with the team because it connects to something real, and it changes when your rates or costs change - which is correct, because the required utilization genuinely does change when those move.

Then sanity-check it against sustainability. If the arithmetic demands 92% utilization, the business model is broken rather than the team being insufficiently busy. Nobody sustains 92%, so a plan that requires it is a plan to burn people out. The fix is upstream - rates, cost base, or the mix of work - and it is much better to discover that from a calculation than from a resignation.

Build in the slack deliberately. Whatever target you derive, plan capacity to roughly 80% of it rather than to the number itself. The gap absorbs the scope changes, sick weeks and overruns that are certainties rather than risks. An agency that plans to its exact required utilization has, without saying so, decided that every surprise will be absorbed as overtime.

When utilization is the wrong metric entirely

Three situations where tracking it produces worse decisions than ignoring it.

Fixed-fee work where you are genuinely fast. If you deliver a £20,000 project in 80 hours because you have done it forty times, your utilization on that project looks poor and your margin is excellent. Utilization measures time occupied, not value created, and on productised or highly efficient work those diverge sharply. Judge that work on project margin instead.

Very small teams. At three people, utilization is dominated by whether a project happened to start on the 3rd or the 18th. The noise exceeds the signal, and the useful question is simply whether the pipeline covers the next eight weeks.

Retainers with outcome-based scope. If you sold responsibility for a result rather than an allocation of hours, hours are your cost rather than your product. Efficiency raises margin and lowers utilization simultaneously, which makes the metric actively misleading.

In all three, the better instrument is project or client margin, covered in agency profit margins. Utilization is a proxy for profitability that works well on time-based work and poorly everywhere else.

The conversation to have with the team

Utilization is the metric most likely to be misunderstood internally, and how it is introduced determines whether the data stays honest.

Say what it is for, explicitly and repeatedly. Capacity and pricing decisions. Not performance evaluation, not comparison between people. If anyone believes it is being used to judge them, the timesheets will start describing what people think is expected, and every number downstream becomes unusable.

Show them the whole calculation. People are considerably more willing to record time honestly when they understand that the number feeds hiring decisions and rate-setting rather than disappearing into a management report they never see.

Publish the denominator. A team that knows available hours are calculated as roughly 1,450 rather than 2,080 understands why 75% is a healthy figure rather than a sign of slacking. Without that, "75% utilized" sounds like a quarter of the week is being wasted.

Never set individual targets. Team or role level only. Individual targets create exactly the incentive to misreport that makes the metric worthless, and they punish whoever happens to be between projects through no decision of their own.

Utilization and hiring

The metric's most concrete use, and the one that saves the most money.

Sustained high utilization with pipeline is the signal to add capacity. Three months above 85% across delivery roles, with signed work continuing, means you are turning away revenue or burning people out - frequently both.

Sustained high utilization without pipeline is a different problem. It usually means estimation is wrong: work is taking substantially longer than sold, so the team is fully occupied on a volume of work that should not require it. Hiring against that multiplies the underlying error rather than fixing it.

Low utilization with a full pipeline points at a bottleneck rather than a capacity shortage - a single specialist everyone waits on, or an approval step that stalls work. Adding people to a bottlenecked process increases the queue behind the bottleneck and nothing else.

The check before any hire: is the constraint capacity, or is it process? Our guides to hiring your first employee and capacity planning both come back to this, because hiring against a process problem is the most expensive mistake available at small scale.

Improving it, in order of effort

If your utilization is genuinely below where it needs to be, the causes are limited and the fixes are ordered.

Check the denominator first. A meaningful proportion of "low utilization" is an arithmetic problem rather than a real one - available hours calculated against 2,080 rather than the true figure. Correcting this sometimes resolves the whole concern in an afternoon.

Then look at the distribution. If two people are at 90% and two at 45%, you do not have a utilization problem, you have an allocation problem. That is the fastest fix on this list and it usually improves how the team feels immediately.

Then look at bench time between projects. Gaps between engagements are the largest single drain in most agencies. The fix is scheduling rather than selling - starting the next project's discovery phase before the current one finishes, so the transition is a handover rather than a gap.

Then look at non-billable creep. Internal projects, tooling, meetings and admin expand to fill available time. An audit of where non-billable hours actually go usually finds one or two recurring commitments that nobody would defend if asked directly.

Only then consider selling more. It is the slowest lever and the one agencies reach for first.

The metric alongside it

Utilization on its own has misled more agency owners than almost any other number, because it is easy to measure and easy to misread. Three pairings make it trustworthy.

With realization, so you know whether occupied time became revenue. Covered above and the single most important pairing.

With project margin, so you know whether the work being done at high utilization is worth doing. A team fully occupied on unprofitable projects is worse off than one at 65% on good ones.

With the distribution, always, rather than the mean.

Together those three answer the question utilization alone only appears to answer: is the capacity we are paying for producing money? Our guide to agency financial metrics covers reading them as a set.

The measurement traps

Four ways utilization data becomes misleading, all common.

Time entered late. Reconstructed timesheets are systematically understated, and the understatement is worst on the projects that ran hardest - so your busiest work looks like your most efficient work. Entry within two days is the threshold where the data becomes trustworthy.

Non-billable work logged as billable to protect a number. Happens the moment utilization becomes a performance metric. Once it starts you have lost every downstream calculation - margin, realisation, project profitability - because they all derive from the same records.

Rounding up. Fifteen minutes rounded to thirty, consistently, across a team, materially inflates the figure.

Excluding people who should be counted. Reporting utilization only for delivery staff produces a healthy number that says nothing about whether the business as a whole converts payroll into billable output.

The defence against all four is the same and it is cultural rather than technical: utilization is used to make decisions about capacity and pricing, never to evaluate individuals. Say that explicitly and mean it, because people will test whether it is true.

What to do with the answer

Utilization is not an end in itself. It feeds three decisions:

• Hiring. Sustained high utilization across roles, with a pipeline, is the signal to add capacity. Sustained high utilization without a pipeline is a signal to fix estimation instead.
• Pricing. If utilization is healthy and margin is not, the problem is your rate or your scope, not your team's effort. That is a commercial conversation and no amount of delivery efficiency will substitute for it.
• Allocation. An uneven spread is the fastest thing on this list to fix and produces the largest immediate improvement in how the team actually feels.

None of those decisions requires a sophisticated model. They require one number you trust, calculated the same way every month, looked at alongside its distribution and alongside realization.

That is genuinely the whole discipline. The firms that do this well are not running more elaborate analytics than everyone else - they are running the same simple calculation consistently, and acting on what it says.