Where Engagement Margin Leaks, and How to Find It
Margin is almost never lost in one decision. It is lost in a series of small, individually defensible concessions that nobody adds up until the engagement closes.
Ask a partner why an engagement made less than planned and the answer is usually a single cause: the client was difficult, the estimate was optimistic, the team was too senior. When the numbers are examined, there are almost always five or six causes, each worth a few percentage points, none large enough to have triggered a conversation on its own.
That is what makes margin leakage hard to manage. It is invisible at the level of any individual decision and obvious only in aggregate, by which time the work is done. The remedy is to know the specific places it hides and to look at them on a fixed rhythm rather than when something feels wrong.
The seven leaks
- Unpriced changes. Work agreed in a conversation, delivered in good faith, and never reflected in the fee. This is consistently the largest single leak and the easiest to prevent.
- Extra review rounds. A deliverable priced for two rounds that takes four has consumed the margin of the whole deliverable, and nobody noticed because each round felt like finishing.
- Grade drift. Work performed by someone more senior than the plan assumed, usually because they were available or because it was faster to do it than to explain it. The effect on recovery is immediate and it never appears in utilisation.
- Unbilled time. Hours recorded, never invoiced, and eventually written off in a batch nobody discusses. Where staff stop recording once the budget is spent, the same loss occurs invisibly.
- Client-caused delay absorbed silently. Late data, unavailable interviewees, decisions that take three weeks. Each one holds the team, and unless the dependency schedule says what happens, the firm pays for it.
- Rate concessions granted at the proposal and never revisited. A discount agreed for a first engagement that quietly becomes the standing rate for a five-year relationship.
- Expenses and disbursements absorbed rather than recharged, usually because recharging feels petty next to the fee. In travel-heavy work this alone can be several points.
The monthly review that finds them
The review is short and it must happen while the engagement is live, because every one of these leaks is recoverable early and none is recoverable at the end.
Compare effort consumed against progress, not against elapsed time. Read the change log and ask whether each approved change reached the fee. Look at effort by grade against the plan. Check unbilled work in progress and its age. Then look at the dependency schedule for anything the client owes that is late, and decide whether the consequence stated in the contract is going to be applied or waived, because waiving it should be a decision rather than an omission.
The measure to watch between reviews
Effort consumed as a proportion of effort planned, compared against percentage complete, is the earliest reliable warning. It moves weeks before margin does, and it is available from data the engagement already produces.
A simple rule works better than a sophisticated one: when effort consumed exceeds progress by more than ten percentage points, the engagement leader owes an explanation and a recovery action. The threshold matters less than the automatic nature of the trigger, because the value comes from removing the judgment call about whether it is worth raising.
Structural fixes, in order of return
Two changes return more than any amount of monitoring. The first is a deliverables schedule that states the number of review rounds, because it converts the most invisible leak into a visible change conversation. The second is a dependency schedule with stated consequences, because it moves the cost of client delay to the party that controls it.
The third, less often adopted, is to price the grade mix explicitly and report against it. A firm that plans an engagement as sixty percent mid-level work and delivers it as sixty percent senior work has changed its cost base by a large fraction, and no other measure will show it.
What not to do about it
Two responses are common and both make things worse. Cutting quality to recover margin creates rework and acceptance disputes, which cost more than the saving. And squeezing the team by expecting unrecorded hours destroys the data the firm needs to price the next engagement correctly, which converts a one-off loss into a permanent one.
The honest response to a structurally underpriced engagement is to finish it properly, record what it actually cost, and price the next one on that evidence. Firms that do this consistently develop pricing that beats their competitors' guesswork within a couple of years.