Service business profitability: the five levers that actually move margin

In a services firm, your people are both the product and the cost, so service business profitability gets decided in five places at once. It's what you charge, what you collect, how much capacity you sell, who does the work and what the firm costs to run. This guide runs all five through one worked model. Then a Lever Sensitivity Test shows which lever to pull first, and it's rarely the one leaders reach for.

What does service business profitability actually measure?

Service business profitability is how efficiently a firm turns its people's available hours into revenue it keeps after delivery and overhead. Profit is a single number at year-end. Profitability is the rate at which billable hours become kept revenue, and that's the part you can manage.

For example, a firm that bills $10M, spends $4M on delivery and $5M on overhead keeps $1M, a 10% margin. The mechanics of that calculation are already covered in our explainer on what an agency profit margin is, so I won't re-teach them here.

Three traits make profit harder to hold in services than in a product business:

• No inventory: You can't stockpile finished delivery, so every quiet week is stock you never get back.

• Expiring capacity: An hour nobody bills on Tuesday can't be sold again on Wednesday.

• Labour-heavy costs: Direct labour averaged 40.6% of revenue in 2024, according to SPI Research's 2025 Professional Services Maturity Benchmark.

That last trait is why every lever in this guide traces back to how your people spend their hours. If an hour is sold at the wrong price, billed at a discount or worked by the wrong person, its margin is gone.

Why fully booked firms still miss their margin

Now picture a firm that watches only one of those variables: busyness. Every calendar is full, the resource plan glows green and the quarter's margin still misses. Leadership asks for more sales, because surely the problem is volume.

That firm is closer to normal than it looks. SPI Research found professional services EBITDA fell to 9.8% in 2024, its lowest in five years and down from 15.4% in 2023. That is more than a third of the average firm's EBITDA margin gone in a year. Billable utilisation slipped to 68.9%, under the 75% threshold the benchmark treats as optimal.

Revenue leakage, which SPI defines as revenue earned but lost before it's realised, rose to 5.3% of revenue in 2024, up from 4.7%. That's the gap between busy and profitable: hours get worked, then drain away through billing errors, misquotes and scope nobody billed.

A word of caution on those figures. SPI's sample spans IT consulting, management consulting, software services and agencies, and 2024 was an unusually weak year. Treat them as a sector-wide signal rather than a verdict on your firm.

Pricing pressure makes the squeeze harder. In Teamwork.com's 6 Strategic Shifts for 2026 research, 66% of senior leaders say clients are now more demanding but less willing to pay.

Before I joined Teamwork.com, I spent years in roles reconciling delivery reports against the P&L. The hours report said the team was flat out, while the finance report showed margin sliding. Neither was wrong; they just never met.

If that sounds familiar, you're not bad at your job. The numbers you need are scattered across timesheets, rate cards, payroll and invoices, so no single report can show the whole picture. Our guide to margin erosion in professional services maps where those leaks usually start, including scope creep and unbilled overruns. Here, the focus is on the variables sitting behind them.

The Five-Lever Profit Chain

Five variables decide whether all that busyness turns into profit, and they multiply rather than add. A small slip in one shrinks the value of every other. I think of them as a chain, because profit only survives if every link holds.

$$\text{Profit} = (\text{Available hours} \times \text{Utilisation} \times \text{Rate} \times \text{Realisation}) - \text{Delivery cost} - \text{Overhead}$$

To make that concrete, take a hypothetical firm built purely for illustration, with no link to any real customer. It has 60 billable staff with 1,800 available hours each, giving 108,000 available hours. At 70% utilisation, it sells 75,600 billable hours.

Its standard rate is $150 an hour, but it realises 88%, so the realised rate is $132. Revenue is 75,600 × $132, or $9,979,200, which works out at $166,320 per billable person.

Delivery cost, at a loaded $38 per available hour, comes to $4,104,000. That's 41% of revenue, close to the 40.6% direct labour ratio SPI Research reports. Overhead is $4,900,000, leaving a profit of $975,200, or a 9.8% margin.

That 9.8% happens to match SPI's 2024 EBITDA average, which makes the model a fair stand-in for a typical firm. The margin is thin enough that small moves in any lever matter.

The chain also explains why spreadsheets struggle with profitability. Rate lives in the proposal, time in the timesheet, cost in payroll and collections in the ledger. Nobody sees the multiplication happen until the month closes.

You can build your own version of this model in an afternoon. You need five numbers from the last 12 months:

1. Available hours: Billable headcount multiplied by the hours each person could work, after leave and holidays.

2. Utilisation: Billable hours logged, divided by available hours.

3. Standard rate: Your rate card value per billable hour, before any discounts.

4. Realisation: Invoiced revenue, divided by billable hours valued at standard rate.

5. Delivery cost and overhead: The loaded cost of billable staff, then everything else on the P&L.

Plug them into the formula above to see which link in your own chain is weakest. If realisation comes out lower than your gut feel, trust the number, because write-offs are easy to forget.

Lever 1: rate (what you charge)

Rate is the price you put on an hour or a deliverable, and it's the lever most firms review least often. Blended rates hide the problem, because one average price means a senior strategist and a junior producer earn identical revenue per hour. In the model firm, a 1% rise across the board is worth $99,792 a year.

Role-based rates let you price the actual mix of people on a job, and they make discounting visible. Our guide to forecasting project profitability before work starts walks through pricing a job from its expected staffing mix.

Lever 2: realisation (what you actually collect)

Realisation is the share of your standard rate you actually bill and collect. For example, at 88% realisation, the model firm keeps $132 of every $150 hour it delivers. The other $18 disappears into write-offs, discounts and unbilled overruns.

Most of that leakage happens between the timesheet and the invoice, so billing accuracy belongs on the profit agenda. The SugarCRM customer story shows what tight billing looks like at scale. After unifying projects, time tracking and billing in one platform, SugarCRM credited less than $20K on more than $10M in annual invoicing.

The formulas for realisation, write-off rate and billable ratio sit in our breakdown of project profitability metrics every agency should track. For the chain, the point is simpler: recovered realisation flows almost straight to profit, because the delivery cost is already spent.

Lever 3: utilisation (how much capacity you sell)

Utilisation is the share of available hours you sell to clients, and it's the lever every firm already watches. In the model, each point of utilisation is worth 1,080 billable hours, or $142,560 of revenue at the $132 realised rate. SPI Research treats 75% as its optimal threshold, but your right target depends on staff mix, since senior people carry more non-billable load.

Formulas, target ranges and ways to lift the number are covered in our guide to resource utilisation rates and formulas. To price a few points of recovered capacity for your own firm, try Teamwork.com's revenue gain calculator.

Lever 4: delivery cost (who does the work)

Delivery cost is the loaded cost of the people doing the work, and staff mix moves it more than salary negotiations do. When a senior consultant spends a day on junior tasks, the job carries senior cost for junior output that the client never pays for. I'd check every project's staffing mix before touching headcount, because a mismatched team quietly inflates the cost of every hour.

Lever 5: overhead (what it costs to exist)

Overhead is everything the firm pays for before anyone bills an hour: non-billable roles, tools, office space and leadership time. In the model firm, it's $4,900,000, almost half of revenue, so even small percentage moves matter.

Overhead tends to creep as firms grow, because each new operations hire, software seat or office upgrade looks sensible on its own. The test I use is simple: does this cost lift rate, realisation or utilisation enough to pay for itself? If nobody can name the lever it moves, it's creep.

Which lever moves profit most? The Lever Sensitivity Test

The chain gives you five levers, and the standard answer to which one to pull first is "cut costs and push utilisation". That answer is incomplete. The Lever Sensitivity Test moves each lever by one point in the model firm, holds everything else constant and measures the profit change.

Add the two price-side levers together and you get +$213,192. The two cost-side levers add +$90,040, so pricing beats cost-cutting by about 2.4 to 1 per point.

The price side wins for a simple reason. A 1% rate rise is worth 1% of revenue, but a 1% cost cut only saves 1% of that cost line. Since revenue is always the bigger base, a 1% rate rise alone beats a 1% cut to both cost lines combined in any profitable firm.

Utilisation tops the table, but read that result carefully. A point of utilisation is a bigger relative move than a point of realisation or a 1% rate rise. It's also the lever most likely to drag the others down with it.

I'd pull realisation first, because the delivery is already done and recovering it costs almost nothing. Rate comes next, starting with new quotes rather than existing contracts.

Here's the Utilisation Trap in numbers. Say the model firm pushes utilisation up to 73%, but the extra hours come from rushed scoping and goodwill fixes, so realisation slips to 84%.

$$\text{Revenue} = 78{,}840 \text{ hours} \times 126 = 9{,}933{,}840$$

Billable hours rise to 78,840, but the realised rate falls to $126. Revenue lands at $9,933,840, which is $45,360 less than the baseline, and the team worked 3,240 more hours to get there. Delivery cost and overhead don't change, so profit falls by the same $45,360.

That's why a utilisation push can backfire so quietly: people feel busier, the capacity dashboard looks healthier and the P&L goes the other way. In my experience, a margin squeeze almost always triggers a utilisation target first, while realisation rarely gets one at all. The warning sign is a rising utilisation line next to a flat revenue line. When you see that shape, look at write-offs and discounts before you celebrate the hours.

What a good profit margin looks like for a service business

Knowing which lever to pull matters little without a target, and most published margin targets compare different layers of the P&L. One source quotes gross margin, another EBITDA and a third net profit, then the numbers get treated as rivals. I'd benchmark each layer separately, using the Margin Layer Benchmark Map below.

Delivery margin has no reliable external benchmark, because it depends entirely on your rates and cost rates. Set a floor for each role from your rate card, then track every project against it. Anything that dips below the floor gets a conversation before the next invoice.

In the model firm, delivery margin is 58.9%: revenue minus the $4,104,000 delivery cost, divided by revenue. That looks healthy until overhead takes another 49.1 points and leaves 9.8%.

EBITDA is the layer SPI Research tracks across professional services firms. The 9.8% average for 2024 is a blended figure across verticals and firm sizes, so treat it as a reference point.

If your EBITDA sits below that average, work back through the chain before cutting anything. Check realisation first, then rate, then whether overhead has crept past what your utilisation can carry. Headcount comes last.

Net margin is the bottom line after tax, and it's where agency data is richest. Promethean Research's Digital Agency Industry Report puts the average agency's net margin at 13% in 2025, against about 15% since 2015. It also finds agencies under 10 FTE earn roughly double the margin of agencies with 50 or more.

Treat all three as directional. Both studies are US-weighted, so UK and Irish firms should read them as a guide rather than a local standard.

So which service business is most profitable? The economics point to firms with pricing power, high realisation and a lean overhead ratio, whatever their sector. In the model, those are exactly the levers worth the most per point.

The AI problem hiding inside your hourly rate

Those benchmarks describe last year, so picture the model firm's next one. AI tools arrive, and the same deliverables now take 10% fewer hours. Nobody works less hard; the tools simply do some of the execution.

On hourly billing, the firm now sells 68,040 hours instead of 75,600. At the same $132 realised rate, revenue falls to $8,981,280.

$$\text{Revenue} = 68{,}040 \text{ hours} \times 132 = 8{,}981{,}280$$

That's a drop of $997,920, which is more than the firm's entire $975,200 profit. Unless pricing changes, the firm posts a loss of $22,720 for delivering the same output faster. I find it the most sobering number in the model, because the firm did nothing wrong.

Promethean Research's report makes the same point for agencies: AI shrinks the time needed for writing, analysis, coding and design exploration. Because so much of the market prices on time and materials, that puts hourly billers directly at risk. The data covers digital agencies, but the arithmetic applies to any firm that sells hours, and the risk is highest where execution makes up most of what you sell.

Clients are already pushing on price, as the Teamwork.com research earlier showed, and budget volatility adds to it: 27% of leaders name mid-project budget moves as their top frustration. The answer is to move the rate lever before the hours fall. Keep hourly billing for genuinely open-ended work, and move repeatable delivery onto structures that don't shrink with the hours:

• Fixed-fee deliverables: Price the output, so faster delivery widens your margin instead of shrinking your invoice.

• Retainers: Sell a defined scope each month, then review it quarterly against actual effort.

• Outcome-based fees: Tie part of the fee to a measurable client result that you can genuinely influence.

Each one needs a quote built from what similar jobs actually cost, and that history comes from logging time against roles with Teamwork.com's time tracking. Without that history, a fixed fee is just a guess with a nicer label. That shift is where service business profitability will be won or lost over the next few years.

Five profitability mistakes that look like good management

The AI squeeze exposes habits that were already costing margin. Each of these five looks like sound management from the inside, which is exactly why they last so long.

Reporting margin at month-end

Month-end margin reports tell you what happened after it's too late to change it. By the time a project shows red in the monthly pack, the overrun is spent and the client conversation is harder. I'd rather see a rough live margin every week than a perfect one 30 days late.

Treating every client as equally valuable

Revenue per client is the wrong ranking, because a big account with heavy write-offs can earn less than a small one paying full rate. Run each client through the chain: rate, realisation, the hours they consume and the senior time they demand.

Our guide to client profitability analysis sets out how to sort accounts by what they actually return. The uncomfortable result is often one or two large clients that the firm is quietly subsidising.

Chasing utilisation without watching realisation

A utilisation target on its own rewards hours worked, whether or not anyone pays for them. Teams hit the number by absorbing rework and small favours, and realisation quietly pays for it. Set a realisation floor next to every utilisation goal and report the two side by side.

Blending rates so nobody sees the mix

A single blended rate is easy to quote and easy to explain, which is why it survives. It also hides which roles make money and which quietly lose it.

Say a blended rate works on a typical team. Put a senior-heavy team on the same rate and the job's cost rises while its price stays flat. Margin drops, and nobody ever decided it should.

Role rates fix the visibility problem. Price each role separately and track cost rates per role too, so you can see margin by person as well as by project. I'd keep the blended figure for a client summary if they prefer it, but never for internal decisions.

Running delivery and finance in separate systems

When projects, time and invoices each live in a different tool, every lever gets measured in a different place on a different day. Realisation leaks in the handovers, and SPI Research's leakage figure is made of them: billing errors, misquotes and scope nobody billed.

A PSA (professional services automation) platform that runs quote to cash in one place closes those gaps. A closed deal becomes a project, its budget tracks against logged time, and that time flows into invoicing without a spreadsheet in between. Nobody rebuilds the numbers at month-end, because they were never apart.

Pro tip: Start every new project from a profitability tracking template with role rates, a budget and a realisation target built in. You'll find ready-made starting points in the Teamwork.com templates library.

How Teamwork.com connects all five levers

Connecting the five levers means rate, time, cost and invoices share one record, so each mistake above surfaces while you can still fix it. As the agentic PSA, Teamwork.com connects projects, resources, financials and AI agents in one platform. Generic project tools track the work but cannot manage the money, while traditional PSAs manage the money but get resisted by the teams who have to log into them. Teamwork.com is the best of both: financials and margin visibility captured as delivery happens, in a platform teams actually want to use.

Rate and realisation only help if you can see them together, project by project. Catch margin slips before the invoice goes out: Budgeting and profitability shows live margin on every project. It tracks fixed-fee, time and materials, and retainer budgets against actual cost as time is logged.

Seeing margin is one thing; knowing how much runway is left is another. Re-plan while there's still money to move: Budget insights compares spend against what remains on each budget.

The rate lever starts at the quote, long before delivery begins. Price from what comparable jobs actually cost: Quoting and costing with role rates lets you set billable and cost rates per role. A senior-heavy staffing plan shows its delivery cost and margin before the proposal goes out.

Once a deal closes, that quote carries through to the project budget and on to invoicing. Quote to cash runs natively in Teamwork.com, so there's no separate billing tool to reconcile.

Utilisation needs a live view of capacity to be worth managing. See who's overbooked or sitting idle at a glance: the Workload Planner shows each person's scheduled hours against their capacity. Team utilisation reporting sits alongside it, so you can spot the Utilisation Trap before it reaches the P&L.

AI is fast becoming a resource you plan, manage and pay for, just like people. Get a quick read on capacity without building a report: the AI Utilization Summary summarises who's over or under capacity across the team. AI Teammates then take on costed, supervised tasks with a named owner, so AI effort sits on the same books as human effort.

FAQ