The six places margin disappears, and the signal that catches each one.
Firms rarely lose margin in one visible event. They lose it in six recurring places, most of which look fine on a dashboard right up until the quarter closes. Here is each one, what causes it, and the signal that catches it early.
What happens. The work turns out bigger than the estimate. Rarely from one bad assumption, usually from an accumulation of small accommodations that individually never justified a change order conversation.
Why it hides. Time-based progress reporting. We are in week six of ten describes the calendar, not the deliverable. Calendar progress is always exactly on schedule.
The signal. Budget burn against scope completion. If you have consumed 60 percent of budget and delivered 40 percent of scope, the engagement is already losing and will keep losing at the same rate. Check weekly, not monthly.
What happens. People are assigned to a paying engagement and are not producing billable hours. Waiting on client credentials, waiting on kickoff, waiting on repository access, waiting on a decision.
Why it hides. This is the most expensive blind spot in the category, because the resource plan looks perfect. Everyone is allocated. Utilization by allocation reads 100 percent. The gap only exists between allocated hours and billed hours, and most firms never put those two numbers next to each other.
The signal. Allocated hours minus billed hours, per person, per week.
What happens. An engagement ends, the next one starts three weeks later, and five people are paid for three weeks of nothing.
Why it hides. Bench cost lands in overhead rather than against any engagement, so every individual engagement can show healthy margin while the firm makes no money. This is the classic case of every project looking profitable and the business not being.
The signal. Forward capacity against contracted work. Not current utilization, which tells you about the past. You need to be looking six to twelve weeks ahead.
This is why headcount forecasting is the largest single margin lever available to a professional services firm. Every other item on this list is worth a few points. Getting hiring and ramp-down timing right against your pipeline is worth more than all of them together, because it operates on your entire cost base rather than on one engagement.
What happens. Cost rates rise every year. Bill rates on long-running engagements do not, unless somebody raises them. A client you have served for four years is often being billed at year-one rates against year-four salaries.
Why it hides. It is gradual and no single month looks wrong. The engagement was profitable when it started, and nothing dramatic happened since.
The signal. Margin percentage by engagement, plotted over time rather than viewed as a snapshot. A slow downward slope on a long-running account is rate erosion, every time.
What happens. Rework, unbilled revisions, extra meetings, hand-holding, quick favors. Real hours at real cost, no revenue.
Why it hides. Only if your margin calculation counts non-billable hours as cost. If it counts billable hours on both sides, this leak is mathematically invisible, and it is invisible specifically on the engagements where it is largest.
The signal. Ratio of non-billable to billable hours per engagement. Normal at 5 percent, alarming at 25 percent.
Once you see it, find out whether it is scope, meaning the client is getting more than they bought, or quality, meaning you are redoing your own work. Those are completely different problems with completely different fixes, and the ratio alone will not tell you which.
What happens. You priced the engagement for two mid-level engineers and staffed it with two seniors, because the seniors were available. You bill the same and pay more.
Why it hides. The engagement delivers well. Clients are happy. Nothing looks wrong except the margin, and the gap is small enough per engagement to attribute to noise.
The signal. Planned role mix against actual role mix.
This is usually an allocation habit rather than a pricing problem. Seniors get allocated because they are the safe choice, and the cost of that safety never appears in the conversation.
If you only track one of these, track the second one. Allocated people not billing is the most common, the most expensive per incident, and the most completely invisible on a standard resource plan.
If you can track two, add forward capacity against contracted work. Headcount forecasting is the biggest lever you have.
Tell us what you need. We will point you to the answer or write the article.
