The service line growing fastest in your business is frequently the one losing money. And it is usually growing fastest for exactly the same reason it is losing money.
That connection is the part almost nobody makes. Fast growth reads as validation. It shows up in the board deck as the bright spot, the thing to lean into, the line that deserves another two hires and a bigger marketing budget. So it gets them. Volume increases, the loss per unit stays constant, and the total loss scales in lockstep with the growth everyone is celebrating.
By the time it becomes obvious, the business has hired around it, committed capacity to it, and built a client base that depends on it. Unwinding is now expensive in a way it wasn’t three years ago.
This is how it happens, why the reporting cooperates, and what the analysis actually looks like when you go find it.
The Line Grows Fast Because It Is Mispriced
Start with the mechanism, because it explains everything downstream.
A service line wins deals when it is easy to say yes to. Low price relative to the alternative, short sales cycle, few objections, minimal procurement friction. Those are the conditions that produce a high win rate, and they are also, quite often, the conditions that produce a bad margin. The line is winning because it is cheap, and it is cheap because somebody priced it wrong.
So the growth rate is not evidence that the offering is strong. It is evidence that the price is below what the market would bear, or below what delivery actually costs, or both. Read correctly, an unusually high win rate on one line is a pricing signal, not a demand signal.
Two things reinforce it. Sales compensation is almost always tied to revenue rather than margin, so the easiest thing to sell is the thing that gets sold, regardless of what it earns. And the low-margin line is frequently the one that started as an accommodation, the thing a good client asked for once, that got said yes to, that then got offered to everyone without anyone ever going back to price it properly.
The origin story matters because it explains the absence of pricing discipline. Nobody built that line. It accumulated.
The Loop That Scales the Loss
Once the line exists and is growing, a specific feedback loop takes over.
Leadership watches a small set of numbers: revenue, growth rate, utilization, sometimes gross margin. The unprofitable line performs well on every one of them, because the costs that make it unprofitable are not the costs that appear in those metrics. Resources follow performance, so the line gets hiring approval, marketing spend, and sales attention.
More resources produce more volume. More volume produces more revenue, which strengthens the signal that justified the resources. It also produces proportionally more loss, but the loss is distributed across shared overhead, senior time, and support functions where nobody is measuring it by line.
The loop does not self-correct because nothing in it is designed to detect the problem. It runs until something forces the question, usually a cash crunch, a lender conversation, or diligence.
And it entrenches while it runs. Three years of scaling means people hired specifically for that work, capacity committed to it, clients who now depend on it, and often a senior champion whose credibility is attached to it. The cost of fixing the problem rises with every quarter it goes undetected, which is the real reason this matters more than a straightforward pricing error.
Where the Money Actually Leaks
The costs that make a line unprofitable share a common property: they are real, they are significant, and they land somewhere other than that line’s cost of sales.
Senior time nobody costs
This is the largest leak in most professional and technical services businesses, and it is almost universally invisible.
The low-margin line is frequently the one that requires rescuing. A principal steps in when the client escalates. A senior engineer gets pulled onto the job that ran into trouble. The owner takes the difficult call. None of that time is billed, and in most firms none of it is tracked against the engagement either, because senior people rarely fill out timesheets with any discipline.
So the line consumes the most expensive labor in the company and the accounting shows none of it. Meanwhile the high-margin line, which runs smoothly and never needs rescuing, absorbs its full share of overhead and looks worse by comparison.
Cost to serve per dollar of revenue
Different lines generate wildly different support loads per revenue dollar. Tickets, calls, revisions, account management hours, billing disputes. A line producing 20% of revenue can easily generate 50% of support volume, and support sits in operating expense where it is invisible by line.
The same asymmetry applies to acquisition. A $15,000 engagement and a $150,000 engagement often require similar sales effort: the same discovery calls, the same proposal, the same procurement cycle. Cost of sale per revenue dollar can be ten times higher on the smaller line, and commission structures rarely reflect it.
The costs that arrive late
Some costs land in a later period than the revenue that caused them. Rework and warranty claims. Support volume that builds after delivery. Churn that shows up at renewal. Collections effort on clients who were price-sensitive going in and turn out to be payment-sensitive coming out.
During growth, this makes the line look better than it is, because current-period revenue is being compared against costs generated by a smaller volume of prior-period work. The true economics only surface when growth flattens and the lagging costs catch up to the revenue base. Which is precisely the worst moment to discover them.
SPI Research puts average revenue leakage at roughly 4.5% of revenue across professional services firms, the gap between work earned and revenue actually realized. The causes are missed billing, scope creep, and write-offs, and none of it appears in your margin reporting.
Source:
SPI Research, Professional Services Maturity Benchmark
Why the Numbers Say It Is Fine
Three specific reporting behaviors keep the problem hidden, and all three are standard practice.
Gross margin is the wrong measure
A line can carry a perfectly healthy gross margin and still destroy value, because gross margin only captures direct cost of delivery. Everything discussed above sits below that line: senior intervention, support load, sales cost, account management.
Contribution margin is the measure that matters, meaning revenue less all costs that vary with the line’s existence, whether or not accounting classifies them as cost of sales. That reclassification is the entire exercise. A line showing 45% gross margin and 4% contribution margin is not a healthy line with high overhead. It is a line that consumes nearly everything it produces.
SPI Research’s 2026 benchmark found average project margins rose to 37.7% and 88.8% of firms hit their annual margin targets. Firm-wide EBITDA still sat at 9.9%. Project-level discipline does not automatically survive the trip to the bottom line.
Source:
SPI Research, 2026 Professional Services Maturity Benchmark
Allocating overhead by revenue
If shared overhead is spread across lines in proportion to revenue, the high-revenue low-margin line absorbs overhead proportional to what it bills rather than to what it consumes. Since it consumes disproportionately, the allocation systematically flatters it and penalizes the efficient line.
This is the core critique behind activity-based costing, and the fix does not require implementing a full ABC system. It requires picking a driver that reflects actual consumption for the few overhead categories that matter most. Allocate support cost by ticket volume. Allocate sales cost by engagement count rather than revenue. Allocate senior oversight by rough time estimate. Three sensible drivers on the three largest shared costs will move the picture more than a precise system nobody maintains.
Averages conceal the distribution
Reporting margin as a single figure per line is the most common analytical failure here, because it hides the shape of the thing.
Rank engagements or clients from most to least profitable and plot cumulative profit, and what emerges is rarely a gentle slope. It climbs well past total company profit, then a loss-making tail drags it back down. Researchers in management accounting have documented this pattern consistently enough that it has a name and a characteristic shape.
The implication for service lines is direct. An average tells you nothing about whether a line is uniformly marginal or contains a healthy core and a destructive tail. Those two situations look identical in a line-level report and call for completely different responses. One needs repricing across the board. The other needs a qualification rule.
The Loss Leader Argument, and How to Test It
At this point in the conversation somebody makes the strategic case: the line does not need to be profitable, because it brings in clients who buy the profitable work.
Sometimes that is true. It is also the single most-asserted and least-tested claim in service business management, and it is entirely testable with data you already have.
Pull every client whose first purchase was the suspect line, over a window of at least twenty-four months. Then measure three things. What share of them subsequently bought the profitable line. How much revenue that produced, and how long it took. And how that compares to clients who entered through the profitable line directly.
The claim survives only if the cross-sell rate is materially higher than your baseline and the resulting margin exceeds the accumulated loss on the entry line. Run it in a spreadsheet with client-level revenue by line and a first-purchase date. It is an afternoon of work.
What usually turns up is one of three answers. The cross-sell is real and concentrated, meaning it works for a specific client profile and not broadly, which converts a blanket strategy into a qualification rule. The cross-sell is real but slow, arriving in year three, which makes it genuine but requires the business to fund the gap deliberately rather than accidentally. Or the cross-sell barely happens, and the strategic story has been doing the work of an analysis nobody ran.
Finding It: The Analysis That Actually Works
Four steps, and the first one is where most attempts go wrong.
Fix your labor rate before anything else
Most firms cost labor by dividing annual salary by 2,080 hours. That number is wrong in both directions and the error compounds through every downstream calculation.
The numerator should include payroll taxes, benefits, insurance, and any other employment cost, not just salary. The denominator should be actually available productive hours, not calendar hours. Subtract paid time off, holidays, training, internal meetings, and administrative time.
Run it on a $90,000 salary. The conventional calculation gives $43 an hour. Add roughly $7,000 in payroll taxes and $12,000 in benefits, then reduce 2,080 hours by 120 for PTO, 80 for holidays, and around 300 for internal and non-productive time, leaving about 1,580. The loaded rate is closer to $69 an hour.
That is sixty percent higher, and every engagement margin calculated with the wrong rate is overstated by roughly the same proportion. A line showing 12% margin at $43 an hour is underwater at $69.
Go to engagement level, not line level
Build a table with one row per job, project, or client engagement over the last twelve to twenty-four months. Columns for revenue, direct labor hours by role, materials or subcontractor cost, and any identifiable direct expenses. Apply your corrected loaded rates.
Then sort by margin and look at the distribution rather than the average. Where does the bottom quartile sit. Is the tail concentrated in one client type, one engagement size, one delivery team, one industry. That pattern is the actionable finding, and it is invisible in any line-level summary.
Harvard’s Robert Kaplan and V.G. Narayanan found the most profitable 20% of customers typically generate 150 to 300 percent of total profits, while the least profitable segment destroys the difference. Line averages tell you nothing about that shape.
Source:
Kaplan and Narayanan, “Measuring and Managing Customer Profitability,” Journal of Cost Management, 2001
If your accounting system cannot produce revenue and direct cost at engagement level, that is a prerequisite problem to solve first, and it is a real project rather than a report you can run this week.
Add the shared costs that vary with the line
Support hours, account management time, sales cost per engagement, and senior intervention. Estimate these if you have to. A rough attribution based on a manager’s honest recollection beats a precise attribution of zero, which is what you have now.
Check the tail against the rest
Compare your worst quartile of engagements against the rest on the dimensions you can see: average size, client industry, how the client was acquired, who delivered the work, whether the scope changed mid-engagement. The differences that appear are your qualification criteria.
What to Do About It
The instinct once the analysis lands is to kill the line. That is usually the wrong first move, and the sequence matters.
Reprice first, because it is the fastest lever and it frequently solves the whole problem. A line losing money at current pricing is often perfectly viable at a correct price. You will lose some volume, and the volume you lose will be disproportionately from the clients who were destroying the margin, which improves the mix twice over.
Requalify second. If the analysis showed the losses concentrated in a specific client profile, engagement size, or delivery pattern, you do not have an unprofitable line. You have an unprofitable segment inside a viable line. Set a minimum engagement size, decline the profile that loses money, and keep the rest.
Redesign delivery third. Productize what has been custom. Move work down to the appropriate seniority level. Template the deliverables that get rebuilt every time. Most low-margin service work is low margin because it is delivered as bespoke work at bespoke cost while being priced as a commodity.
Kill last, and with the fixed cost math in front of you. This is where owners get hurt. Eliminating a line removes its revenue immediately and its associated costs only partially, because the shared overhead it was absorbing does not disappear with it. If a line contributes 20% of revenue and covers 15% of fixed cost above its direct cost, cutting it makes the remaining business carry that 15% alone. That can be worth doing. It is not automatically worth doing, and the calculation should be explicit rather than assumed.
The Bottom Line
Unprofitable service lines do not scale because management is careless. They scale because every signal management watches points the same direction, and the costs that would point the other way are sitting in shared overhead where nothing separates them by line.
The line grows because it is underpriced. It gets resources because growth is the metric. The loss compounds with volume while the reporting stays flat. And the whole thing entrenches, so the fix gets more expensive every quarter it goes unexamined.
Finding it takes correcting your loaded labor rate, moving from line averages to engagement-level distribution, and attributing the shared costs that actually vary with the line. That is a week of work in a spreadsheet, and it is the difference between scaling something that compounds and scaling something that drains.
Where to Go From Here
The analysis is straightforward once the data is structured for it. The harder part is usually the structuring, and after that, deciding what the distribution is telling you to do about pricing, qualification, and delivery.
At Frak Finance, margin analysis by service line, engagement, and client is core to the work, along with the reporting infrastructure that keeps the answer current instead of a one-time exercise.
Schedule a free consultation and we’ll find out which part of your business is funding the rest of it.
