Agencies typically target a 50% or better gross margin on delivery and 15-25% net profit, with stronger performers running above that. But benchmarks are only comparable if the other firm counts costs the same way you do, and most do not - the biggest variable is whether owner salaries and non-billable staff sit above or below the line. Before comparing yourself to anything, calculate gross margin per project. Agency profitability is decided at the project level and averaged at the company level, which means a healthy-looking overall margin routinely conceals two or three accounts that are losing money.
Most agency owners can tell you their revenue instantly and their margin approximately.
The approximation is the problem. Not because the number is hard to calculate, but because "margin" describes at least three different things, agencies compare across them without noticing, and the aggregate figure hides the project-level reality where profitability is actually decided.
This guide covers the three margins that matter, what the benchmarks say and how much to trust them, why project-level margin is the only view that changes decisions, the five things that erode margin, and what to do about each.
Why "we're profitable" is not an answer
Ask most agency owners whether the business is profitable and the answer is yes. Ask which projects made money and the answer is a pause.
That gap is the subject of this guide. Profitability at the company level is an average of decisions made at the project level, and averages are precisely the wrong instrument for finding out which decisions were bad ones.
The three margins
Getting these separated is most of the work.
Gross margin (delivery margin)
Revenue minus the direct cost of delivering it, divided by revenue.
Direct cost means the people doing billable work - their salaries, employment costs, and any contractors on the project. Not rent, not software, not the operations manager.
This is the most important number in an agency, because it tells you whether the core activity is economically viable before any overhead. 50% or better is the usual target, and Swydo's profitability guide treats it as the primary health indicator for good reason: an agency with a 30% gross margin cannot overhead its way to profitability, because there is not enough left.
Operating margin
Gross profit minus overhead, divided by revenue. Overhead being rent, software, admin salaries, marketing, insurance - everything that is not direct delivery.
This tells you whether your cost base is proportionate to your delivery capacity. A healthy gross margin with a poor operating margin means you are over-structured for your size.
Net margin
What is left after everything, including tax and interest. The number people quote and the least comparable, because it depends heavily on the legal structure and how the owners pay themselves.
What the benchmarks say, and how much to trust them
Published figures for agencies and professional services firms cluster around 15-25% net profit, with stronger performers above that. Mosaic's consulting profitability benchmarks put operating margins for consulting firms in a broadly similar 15-30% band, and Haus Advisors' breakdown of agency margins notes the pattern most sources agree on - that margin tends to improve with scale, as fixed overhead spreads across more revenue.
Treat all of it as orientation rather than a target, for one specific reason: the biggest variable in any agency's net margin is how the owners pay themselves.
An owner-operator taking a modest salary and leaving profit in the business reports a spectacular net margin. The same business with a market-rate salary for the same person reports a mediocre one. Nothing about the underlying economics differs.
So before comparing yourself to any published figure, apply one correction: is the owner's time costed at market rate? If you are doing billable work and paying yourself below what you would pay someone to replace you, your margin is flattered by exactly that difference. That is a legitimate way to fund early growth; it is not a sustainable margin.
Two more comparability traps:
Where non-billable staff sit. Project managers, account leads and QA are direct cost in some firms and overhead in others. That choice alone moves gross margin by ten points or more.
Whether pass-through spend is in revenue. An agency that books client media spend as revenue reports enormous revenue and a terrible margin. One that nets it off reports the opposite. Neither is wrong; they are not comparable.
Project-level margin is where the answer is
The aggregate margin is a company-level average. Averages hide the thing you need to see.
A typical agency with a healthy 22% net margin frequently looks like this underneath: a handful of projects at 60% gross margin, a majority around 45%, and two or three at or below zero. The good work is subsidising the bad, and because nobody calculates it per project, the bad work continues indefinitely - often for the clients everyone finds most demanding.
Calculating project margin:
- Project revenue - what you invoiced, or the fee.
- Direct cost - actual hours logged by each person × their fully-loaded hourly cost. Fully-loaded means salary plus employment costs plus benefits, divided by their genuine available hours. Our guide to calculating a billable rate covers deriving that figure properly, and the usual error is dividing by 2,080 hours rather than real capacity.
- Gross margin = (revenue − direct cost) ÷ revenue.
Run it on your last ten completed projects. Two things almost always emerge: at least one project you believed was fine was not, and the pattern of which projects lose money is more consistent than you expected - usually a client type, a service line, or a project size.
That pattern is the actionable output. Company-level margin tells you there is a problem; project-level margin tells you where.
The five things that erode margin
In rough order of how much damage they do.
1. Scope absorbed without a change order
The largest and quietest. Each individual request is too small to raise, and collectively they are a third of the project. The work is delivered, the invoice is unchanged, and the margin absorbs the difference.
The fix is mechanical rather than a matter of firmness: a scope with explicit exclusions, and a change order process small enough that using it is easier than absorbing the work. Our guide to preventing scope creep covers the structure.
2. Underpricing
Common, and usually invisible because it looks like being competitive. The diagnostic is straightforward: if you win nearly everything you quote for, and nobody ever questions the price, you are priced below what the market would bear. Our guide to raising rates covers the sequencing.
3. Poor realisation
Tracked billable hours that never reach an invoice - discounted at billing, absorbed as goodwill, or written off because nobody could reconstruct what they were for.
Realisation rate = invoiced hours ÷ billable hours tracked. A rate of 85% means 15% of your delivered work was free, which is equivalent to a 15% price cut applied without deciding to. Recovering it requires no client conversation at all, which makes it usually the easiest available margin improvement.
4. Low utilization
If your team is billable 45% of available time and your rates assume 70%, the arithmetic does not work regardless of how well you deliver. Asana's guide to utilization rate and Scoro's breakdown of billable utilization cover the formulas; our post on agency utilization rate covers why the denominator you choose changes the answer substantially.
Low utilization has two causes with opposite fixes: not enough sold work (a sales problem) or too much non-billable overhead per billable hour (a structure problem). Diagnose which before acting.
5. Seniority drift
The project was priced assuming a mid-weight designer and delivered by a senior one. Nothing in the invoice reveals it, the client is delighted, and the margin quietly halves.
This is the specific risk of a blended rate, and it is invisible unless you compare planned staffing to actual on completed projects. Sampling a few projects a quarter is enough to spot it.
Improving margin, in order of effort
Cheapest: fix realisation. Stop discounting at invoice, bill the overage you already agreed, and make sure tracked time actually reaches invoices. No client conversation, no delivery change.
Next: stop doing the unprofitable work. Once you have project-level margin, the loss-makers are visible. Re-price them, re-scope them, or let them go. A client at negative margin is one you are paying to serve.
Next: raise rates. Covered in its own guide. New clients first, existing at renewal.
Next: improve utilization. Either sell more or restructure. Slower and involves people.
Slowest and largest: change what you sell. Specialisation, productisation, or moving up-market. Real margin transformation usually lives here, and it takes a year or more.
Most agencies attempt these in reverse, starting with the strategic repositioning and never doing the realisation work - which is the one that would have produced results this quarter.
The margin levers, quantified
Worth putting rough numbers against each lever, because it changes what you do first.
Realisation. Moving from 85% to 95% is a 10-point improvement in effective revenue with no client conversation and no delivery change. On a £500,000 agency that is £50,000, and it requires only that tracked billable time reaches invoices. Fastest and largest available win for most agencies.
Scope discipline. If you absorb 15% of project scope on average, recovering half of that through change orders is worth roughly 7% of revenue. Requires process rather than negotiation.
Rate increase. 10% on new clients flows almost entirely to margin, since costs do not move. Slower to take effect because it applies only to new work, and it compounds.
Utilization. Moving from 65% to 72% is meaningful and it is the slowest lever, because it requires either more sold work or a restructure. Frequently attempted first and rarely the right starting point.
Cost reduction. Usually the smallest, because in a services business the costs are people and people are the product. Worth reviewing annually and rarely worth leading with.
The ordering is consistently the reverse of what agencies attempt. Realisation and scope discipline are internal, fast and uncomfortable in a small way. Rate increases and restructures are external, slow and uncomfortable in a large way - which is why they get discussed more and done less.
Why margin varies by service line
Aggregate margin also hides variation between the things you sell, and the pattern is consistent enough to be worth checking directly.
Retainers usually carry better margin than projects - lower sales cost, predictable capacity, familiar work. Unless they carry silent overage, in which case they are the worst thing in the business, because the loss repeats monthly.
Strategy and advisory carry high margin and poor realisation. The rate is good; the write-offs at invoicing are worse than anywhere else, because advisory time is the easiest to feel awkward about billing.
Production and implementation carry lower margin and better realisation. The work is concrete, so it is easier to bill in full.
Anything novel loses money the first two or three times. That is a legitimate investment in a capability, provided it is a decision. It becomes a problem when the third and fourth instances are priced as though the learning happened, and nobody checks whether it did.
The action is not to stop selling low-margin lines - a low-margin service that wins high-margin follow-on work earns its place. It is to know which is which, so the cross-subsidy is deliberate rather than accidental.
Margin and agency size
The relationship is not linear, and knowing the shape prevents some avoidable panic.
Very small agencies (1-5) often report strong margins because overhead is minimal and owner time is undercosted. The figure is real in cash terms and overstated as a business margin.
The 6-15 range is where margin typically compresses. You have added non-billable roles - operations, account management, a project manager - before the revenue base is large enough to absorb them. This is the most common point at which an owner concludes something has gone wrong. Usually nothing has; the overhead arrived before the scale to carry it.
Above 15-20, margin tends to recover as fixed overhead spreads and specialisation raises rates.
If you are in the compression zone, the useful response is patience plus attention to utilization - not cutting the structure you just built, which is the instinct and usually the wrong move.
Running a project margin review
The single most useful hour an agency owner can spend, and almost nobody does it.
Pick your last ten completed projects. Not a sample of the memorable ones - the last ten, in order, including the small ones.
For each, calculate three numbers. Revenue invoiced. Actual hours logged, multiplied by each person's fully-loaded hourly cost. The gross margin between them.
Then add two columns that are not financial. How many out-of-scope requests were absorbed, and how many revision rounds beyond the agreed number. Those two frequently explain the margin better than anything in the accounts.
Look for the pattern rather than the outliers. One bad project is noise. Three bad projects sharing a client type, a service line, a size or a salesperson is a finding, and it is almost always more consistent than people expect.
What typically emerges is uncomfortable and actionable: a specific kind of work that is systematically unprofitable, usually the kind that felt easy to sell. Our guide to niche positioning covers what to do when the pattern points at a segment rather than a project.
The margin conversation with the team
Margin improves faster when the people delivering understand it, and most agencies keep it from them entirely.
Share the commercial shape of each project. Fixed-fee or time-based, roughly what margin it carries, what absorbing an extra request costs. A delivery team told nothing has no reason to treat scope as finite, and then gets blamed for over-servicing.
Do not share it as pressure. The framing is context, not a target. "This is a fixed-fee project at a tight margin, so flag anything out of scope rather than absorbing it" is useful. "We need to protect margin on this one" without the mechanism is just anxiety.
Show them where the leaks are. Realisation and absorbed scope are things delivery people can directly affect, and most would happily flag out-of-scope requests if anyone had explained why it mattered.
The agencies with the healthiest margins tend not to have better financial controls. They have delivery teams who know what a change order is for and use it without being asked.
What to measure monthly
Five numbers. Fifteen minutes if the data is clean.
| Metric | What it tells you | Rough target |
|---|---|---|
| Gross margin | Whether delivery is viable | 50%+ |
| Net margin | Whether the business is viable | 15-25% |
| Utilization | Whether capacity is sold | Per your own model |
| Realisation | Whether sold work is billed | 95%+ |
| Project margin spread | Where the losses are | No project below 30% |
The last row matters most and is measured least. A company-level margin is a summary; the spread is the diagnosis.
When margin is fine and the business still feels wrong
A situation worth naming, because the numbers can look healthy while something real is wrong.
Good margin, exhausted team. Usually means the margin is being produced by unsustainable utilization rather than by good pricing. Check the utilization distribution - the margin is real and it is being borrowed from people, which is a loan that eventually gets called.
Good margin, no growth. Frequently a positioning problem rather than a delivery one. The work is profitable and there is not enough of it, which points at specialisation and pipeline rather than at anything in this guide.
Good margin, high churn. Points at over-servicing being absent rather than present - clients getting exactly what was scoped and no more, which is commercially correct and can feel transactional. Worth checking whether the difference between profitable and generous has been drawn slightly too far.
Good margin, founder cannot step back. The margin depends on the founder's own billable output, which is not a margin, it is a job. Costing owner time at market rate reveals this immediately and usually reduces the reported figure substantially.
In all four the finance review says the business is healthy and something else is telling you otherwise. The numbers are necessary and they are not sufficient, and the useful discipline is to look at margin alongside utilization distribution, churn and what the founder actually spends their week doing.
The one-hour diagnostic
If you do nothing else after reading this, spend an hour on the following and you will know more about your business than most agency owners do.
Take your last ten completed projects. For each: revenue invoiced, actual hours logged, each person's fully-loaded hourly cost. Compute gross margin per project. Sort them.
Then answer three questions. What is the spread? What do the bottom three have in common? Would you take that work again at that price?
Almost everything worth changing about an agency's profitability is visible in those three answers, and none of it is visible in the revenue line.
The honest summary
Agency profitability is rarely fixed by cost-cutting, because in a services business the costs are people and people are the product.
It is fixed by four things: knowing your real delivery cost, pricing above it deliberately, billing all the work you actually do, and identifying which projects lose money so you can stop repeating them.
All four depend on the same foundation - time data honest enough to trust. Which is why an agency with a margin problem almost always has a measurement problem first, and why the fix usually starts somewhere less strategic than it feels like it should.