Why your agency has the information but still misses the warning signs
Why margin problems reach founders as a surprise, and how to see which projects are losing money while there's still time to act

Michael Gold, CEO and Co-founder, Aria
12 min read
One source of truth tells you where to look. It doesn't tell you when to look, and for most agency founders that is the part that matters. The information that would have warned you usually exists somewhere, but it arrives late and in pieces, and someone still has to put it together before anyone can act on it. In a lot of agencies, that someone is the founder. That is why margin problems so often reach founders as a surprise rather than a warning.
I've worked in delivery for 13 years, and I've heard "we've got everything in one system now" from plenty of businesses that were still finding out about their margin at the invoice. Getting the data into one place is worth doing. It just answers a different question from the one founders are really asking: which projects are losing money, and can I still do something about it?
Why do margin problems reach founders as a surprise?
Because by the time a problem is big enough to reach the founder, it has usually been building for weeks in places the founder doesn't look every day. Nobody is hiding it. It just hasn't been joined up yet.
In most agencies I've worked with, the truth lives in three places. The plan and the deadlines are in a project tool. The budget and the margin are in a spreadsheet, often one that finance updates once a month. Everything else is in somebody's head: the client who asked for a bit extra, the designer who has been working late, the account that feels shakier than the numbers say. Quite often that head belongs to the founder.
None of this is anyone's fault. It is how most agencies grow. You build the business on instinct and a spreadsheet, and it works, until there are too many projects for one person to keep in view.
The Wow Company's BenchPress research asked owners of UK agencies with £1m+ fee income what would happen to the business over three years if they left. Nearly half, 48%, said it would shrink or go out of business (The Wow Company, 2026). BenchPress looks at that as a risk for anyone buying an agency, and it is one. I read it as something more everyday. In many agencies, the founder is still the place where the whole picture finally comes together.
What owners expect if they leave
48% expect contraction or closure
- The business would grow19%
- Stay roughly the same size33%
- Contract33%
- Go out of business15%
The trouble is that being that place means you hear about things last. You're blindsided in front of a client or your own team, or you find out when the invoice goes out, after the moment you could have changed it.
Why doesn't one place to look tell you when to act?
Because the problem is rarely that the information is in the wrong place. It is that it arrives late, and the small differences between each view of the business build up quietly while nobody is checking them against each other.
Picture a fairly ordinary month. A client asks for an extra round of changes. The account lead agrees, because it seems small and the relationship matters. The plan isn't updated, because nothing on it has moved yet. The budget looks fine, because most of the extra time hasn't been logged. Each view is accurate as far as it goes. It is only when you put all three side by side that you can see the project has started to drift, and nobody is putting them side by side until the month-end numbers do it for them.
Finance teams feel a version of this too. In a 2024 survey across seven countries, including the UK, nearly 40% of CFOs said they don't completely trust their own organisation's financial data (BlackLine/Censuswide, 2024). If the people whose job is the numbers aren't fully confident in them, the founder piecing it together from three sources is in a harder spot still.
I've said before that a source of truth that is two weeks behind can still leave you flying blind. The cost isn't really the discrepancy itself. It is the time between when it started and when somebody finally saw it.
I've managed projects where the margin looked fine until a week or two of missing timesheets finally landed. Then 40 or 50 hours appeared and the project changed in front of me. The frustrating part wasn't simply finding out we had spent more than expected. It was finding out after the work had already happened. If I had seen those hours building at the time, I could have challenged scope, changed the plan, moved people or had a commercial conversation with the client. Once the time finally hit the numbers, most of those options had gone.
If you run operations rather than the whole business, you'll recognise this from the other side. You are often the one asked to explain the gap once it has already turned into a problem.
What are the signs your agency has an operating problem, not a workload problem?
If the same kinds of problem keep coming back whoever is working on them, it is usually the way the agency runs rather than the amount of work. A workload problem eases when you add people. An operating problem tends to follow the new people in.
This is why I push back when a growing agency tells me the answer to its delivery problems is its first junior PM. By the time a business decides it needs someone, things have usually started going wrong already. Projects are under pressure, clients need more attention, the founder is stretched and nobody really owns the process.
The new person, often on around £30k, is then asked to do two jobs at once. They have to manage everything that is currently on fire and also design the systems that would stop the next fire starting. The reality is that a £120k Head of Delivery could struggle with that. When the junior PM burns out, it is very easy to decide the hire was the problem and put the next person into exactly the same system.
This isn't one client's story. It's a pattern I've seen more than once, and if the same problems survive changes in personnel, that is normally the giveaway that the system is the issue, not simply the workload or the individual.
These are the signs I look for that it is the operating model, not the workload:
- You find out whether a project made money at the invoice or at month-end, rather than while it is still running.
- The same project sounds different depending on whether you ask the PM, the account lead or finance.
- Projects are green, but people are working late to keep them that way.
- Status gets rebuilt by hand before every meeting, and another meeting gets booked to work out what has changed.
- Problems only get joined up when they reach you.
- You've already hired to fix it, and the same problems came back.
If several of those sound familiar, adding people may help for a while, but it won't change when you find out.
What does finding out late actually cost?
Mostly it costs you options. The earlier you see a project drifting, the more ways you have to deal with it, and in a UK agency there often isn't much margin to absorb the ones you miss.
Overruns themselves are normal. Across industries, more than a quarter of projects miss their budget and more than a third miss their schedule, according to PMI's latest Pulse of the Profession (PMI, 2025). Scope moves, clients change their minds, and sometimes the extra work is the right call. What matters to a founder is whether you hear about it while you can still talk to the client about it, or only once it is in the numbers.
The margin you have to play with is thinner than most people outside agencies would guess. I've written about what a good profit margin looks like for a UK agency elsewhere, but the short version is that most are below the 50% gross profit benchmark The Wow Company uses. In 2025 the average for agencies with £1m+ fee income was 43%, and only 36% reached 50% (The Wow Company, 2026).
After overheads and owners' dividends, the room gets smaller still. Average operating profit was 14%, against the 20% BenchPress recommends, and 8% of £1m+ agencies made no profit at all (The Wow Company, 2026).
UK agency operating profit
- 2025 average14%
BenchPress recommends above 20%
BenchPress measure, after owners' dividends. This differs from operating profit in statutory accounts.
When the whole business is working on that kind of margin, a project that quietly runs over for a few weeks doesn't only dent its own numbers. It takes a bite out of the profit the agency was counting on, and by the time it shows up at month-end there is very little left to do except explain it.
How can a founder see which projects are losing money while there's still time?
By having something join up the three views of the business continuously, so you are not the one doing it at month-end. The signals that a project is losing money usually show up weeks before the invoice. They just need to be put side by side while there is still time to act on them.
If I were running an agency today, these are the questions I would want answered every week, without having to ask anyone:
Which projects have used more of their budget than the work they have actually delivered?
Where is time missing, so that the margin I am looking at is probably better than the truth?
What has been agreed with a client that isn't in the plan or the budget yet?
Who is stretched to keep a project on track?
Underneath all four sits one question: what changed this week that affects delivery, margin, capacity or client confidence? You can answer all of that by hand. Most founders do, in their heads, on the way home. The problem is that it relies on you having the time and the full picture, which is exactly what stops being true as the agency grows.
That is what we built Aria to help with. Aria is agency management and PSA software. It brings projects, people, time and money together, with connected activity added through background updates. Margin to date uses recorded revenue, time and approved expenses. Forecast costs also account for remaining estimates and bookings, so you can look ahead rather than only add up what has happened. Aria can highlight delivery risks and suggest a next move.
If you have been looking at agency profitability software or agency management software in the UK, that is the part I would test hardest: whether it tells you in time, rather than whether it stores the numbers.
What happens next is up to you. Aria gives you warnings and options to consider, and the founder decides which fit the business. Changes to plans and client conversations still need your judgement. You can see how that works for founders and CEOs.
Should you hire someone to hold the picture instead?
Sometimes that's the right call, but hiring someone to rebuild the picture by hand moves the job to another person rather than taking it away. The picture is still put together after the fact. It is just someone else doing it.
There are really three ways agencies handle this today. The first is to find out at the invoice and deal with it then, which is where most of us start. The second is the project tool, the spreadsheet and the one person who holds it all in their head, which works until that person is stretched or away. The third is to hire someone to hold it for you: a PM, an operations lead, or a fractional COO or delivery consultant.
I do that third kind of work myself through Gold Strategic Delivery, so I'm not going to tell you it has no value. A good consultant brings judgement and experience that software doesn't. But if the picture they rely on still has to be rebuilt by hand every week, it fades when they step back, and you are still the person stitching all the systems together. A new tool won't change that either if someone still has to work out what the numbers mean. That's the difference worth testing between Aria and other PSA software: whether it tells you what needs attention in time to act.
What I would want for any founder is simpler than a new hire or a new dashboard. It is not spending your evenings piecing the business together in your head, or asking five people for five versions of the truth. It is knowing which exceptions need your attention, having the client or commercial conversation before it becomes an escalation, and being able to step away for an evening or a holiday without worrying that the business only works because you are personally holding the picture together. That is when you get to stop being the person who joins everything up, and go back to being the founder.
If you want to see it, book a walkthrough or start a free 14-day trial.
Common questions
Sources
- The Wow Company, BenchPress 2026: Buying & Selling Agencies (2026), £1m+ edition. Free, registration required.
- BlackLine, CFO survey conducted by Censuswide (2024). Commissioned by BlackLine, a finance software company.
- PMI, Pulse of the Profession 2025 (2025)
- The Wow Company, BenchPress 2026: State of the Agency Nation (2026), £1m+ edition. Free, registration required.
Written by

Michael Gold
CEO and Co-founder, Aria
Michael spent 13 years in delivery across agencies and consultancy firms and founded Gold Strategic Delivery before co-founding Aria.