Agency utilisation and recovery rate: how to measure whether your team is actually paying for itself

Most agency owners can tell you how busy the team feels. Far fewer can tell you what that busyness is worth. There’s a gap between a full calendar and a healthy business, and the two numbers that close it are utilisation and recovery rate.

I’ve run an agency for about twenty years, and for a long time I tracked neither properly. We were busy, everyone was stretched, and I assumed busy meant profitable. It doesn’t, not on its own. Busy tells you the hours got used. It says nothing about whether those hours turned into money. This piece walks through both numbers, how to calculate them without a finance degree, what a healthy figure looks like, and why recovery rate is the one most agencies never measure even though it’s the one that tells the truth.

Start with the hours: billable versus non-billable

Everything here rests on one honest split. In any given week, a person’s time falls into two buckets. Billable time is work that a client is paying for, or should be. Non-billable time is everything else: internal meetings, admin, new business pitches, training, the Monday status call, and the twenty minutes lost to finding the right file.

Non-billable time is not wasted time. An agency needs new business, needs to train people, needs to run itself. The point is to see the split clearly, because the ratio between the two buckets is where your economics live.

Most agencies guess at this ratio. They have a feeling that people are “mostly on client work,” and the feeling is usually generous. When you actually track it, the non-billable share is often larger than anyone expected, and a good chunk of it hides inside meetings nobody counted.

Utilisation rate: how much of the available time went to billable work

Utilisation rate is the simplest of the two. It answers one question: of the hours a person had available, how many went to billable work?

Utilisation rate = Billable hours ÷ Available hours × 100

If someone is contracted for 160 hours in a month and logs 112 billable hours, their utilisation is 70%. The other 48 hours went somewhere, and that somewhere is worth understanding.

A few things trip people up here, so they’re worth stating plainly.

Available hours should be realistic, not theoretical. If you divide by the full eight-hour day, you’re assuming nobody ever attends a meeting, answers an email, or takes a break. Six productive hours a day is a more honest denominator for most creative and delivery roles. Divide by eight and every utilisation figure you produce will flatter you.

Utilisation is measured per person, then rolled up. An agency-wide average of 60% can hide a senior sitting at 85% and a junior at 35%, and those are two completely different problems. One is a burnout risk, the other is a resourcing or skills gap. The average tells you neither.

What is a good utilisation rate for an agency?

There’s no single right answer, and anyone who quotes you one number is selling something. As a working guide, client-facing delivery roles in a healthy agency tend to land somewhere between 70% and 85% billable utilisation. Below that and you’re likely carrying more capacity than the work supports. Push much above it consistently and you’re running people with no slack for the non-billable work that keeps an agency alive, which is its own kind of expensive.

Roles change the target. A designer or developer should sit high. An account director who spends real time on client relationships and new business will sit lower by design, and that’s correct rather than a problem to fix.

The honest version of this metric measures each person against their own contracted hours, month by month, because the working days in a month move and a fixed target lies to you in February.

The problem with stopping at utilisation

Here’s where most agencies stop, and here’s why stopping there is a trap.

Utilisation tells you the hours were used on billable work. It does not tell you that you got paid for them, or paid at the rate you intended. A designer can be 90% utilised and still lose the agency money if half those hours overran the estimate and were written off, or were billed at a rate that doesn’t cover what that person costs to employ.

Full utilisation on unprofitable work is just efficient loss-making. That’s the sentence I wish someone had said to me fifteen years ago.

To see whether the used hours actually earned, you need the second number.

Recovery rate: how much of the work you did actually converted into money

Recovery rate, sometimes called realisation rate, closes the gap between hours worked and money earned. It answers the question utilisation can’t: of the billable work your team performed, how much did you actually recover in billing?

Recovery rate = Value invoiced ÷ Value of billable time worked × 100

Say a job was estimated at 40 hours. Your team tracked 52 hours of billable work against it, worth R78,000 at their rates. You invoiced the client R60,000, because the extra 12 hours were an overrun you chose to absorb rather than pass on. Your recovery rate on that job is roughly 77%. Twenty-three percent of the work you performed never made it to an invoice.

That is the number that explains why a busy agency can still be tight on cash. The hours were used, utilisation looked fine, and a quarter of the value walked out the door through overruns, scope creep, goodwill write-offs, and rates that were set once and never revisited.

Why recovery rate is the one most agencies never measure

Utilisation is easy to calculate from a timesheet. Recovery rate needs three things joined together that usually live in separate systems: what was estimated, what was actually tracked, and what was actually invoiced. When your project management sits in one tool, time tracking in another, and billing in your accounting package, joining those three up is a manual, month-end reconciliation job that most agencies never get around to doing per job. So they fly on utilisation alone and wonder why the bank balance doesn’t match the calendar.

A worked example: two people, same utilisation, very different value

Take two designers, both 80% utilised in the same month. On paper, identical:

• Person A works inside their estimates, hits the scoped hours, and their work invoices cleanly. Recovery rate around 100%. Every billable hour they logged turned into billed value.
• Person B is just as busy, but consistently runs 30% over estimate, and those overruns get absorbed rather than billed. Same 80% utilisation, but a recovery rate closer to 70%. The agency is paying for the same seat, getting the same “busy,” and recovering far less.

Utilisation alone would rate them equally. Recovery rate shows you which one is actually carrying the agency, and it turns a vague sense that “something’s off with margins” into a specific, fixable conversation about scoping and estimates.

The cost side: you can’t judge either number without a real resource cost

Both metrics assume you know what an hour of each person’s time actually costs the agency. Not their salary divided by hours, the real figure, including the overheads that person’s work has to help recover: their share of rent, software, tools, and the salaries of everyone who can’t bill.

If you leave that number out, every job looks more profitable than it is, and a strong-looking recovery rate can still hide a loss. A rough model that works: take a person’s salary, add their department’s share of tools and software, add a share of general agency overheads, and divide by their realistic monthly billable hours. Build the rate that way and both utilisation and recovery start telling you the truth instead of a flattering version of it.

How to actually track this without a month-end spreadsheet marathon

You can absolutely do all of this in spreadsheets. I did for years. The problem is that by the time the spreadsheet is built, the month is over, the overrun is spent, and the moment to have the conversation with the client has passed. Utilisation and recovery are most useful while the work is still moving, when you can still add time to a scope or catch a job before it runs away.

This is the specific problem Agencydesk was built to solve, because it was built only for agencies rather than adapted from a general tool. Because job costing, time tracking, and invoicing all run on one structure, the three numbers recovery rate depends on are already joined together. Every job carries a live reconciliation as the team logs hours: estimated against tracked, tracked against billable, billable against invoiced. Recovery rate is measured per job as the work happens, not rebuilt weeks later.

Timesheets carry a workload score and a productivity score, so utilisation is visible per person and per department against each person’s own contracted hours. The Job Recon view costs tracked time at each person’s real resource cost, subtracts supplier spend, and prices in the meeting time that usually escapes the calculation entirely, so what’s left is the true nett profit on the job rather than a hopeful estimate. And Multi-Job Recon rolls it up to answer the bigger question: is a high-turnover client actually building the agency, or absorbing it?

For an account or project manager, this changes the daily job. Instead of reading numbers off a report someone built, they can see a job drifting past its estimate while there’s still time to act, and have the resourcing conversation with the client before the write-off rather than after it.

The short version

Utilisation tells you the hours were used. Recovery rate tells you the hours earned. You need both, and most agencies measure only the first, which is why so many busy agencies stay tight on money.

Track them per person and per job, against a realistic view of available hours and a real resource cost, and you move from a feeling about whether the agency is healthy to a number you can act on. That’s the whole game: fewer surprises at month-end, and enough visibility during the work to fix a job while fixing it still matters.

If you want to see what these numbers look like on your own jobs, you can book a demo and we’ll walk through it against the way your agency actually runs. For the detail on how recovery rate and resource cost are calculated, the Job Recon guide in the help centre goes through it step by step.