The Hidden Margin Killer in Mid-Sized AEC Firms: Unbilled Control Time
Why busy firms lose margin in hours nobody invoices, and six ways to get them back.
Ask a principal what a project's fee is, and you'll get a number in seconds. Ask how many hours the firm spent redoing work it had already done once, and the room goes quiet.
In DD, I've corrected door, window, and wall schedules that no longer matched the model. I've also reviewed late sets and had to go back to something that was supposed to be resolved phases earlier. That work is real, and it protects the firm. It's also time nobody planned and nobody bills.
I call it unbilled control time: the hours a firm spends checking, correcting, and re-coordinating work after the fact. It rarely gets its own line item. It hides inside "project management," inside overtime, and inside the fee the firm quietly writes down at closeout.
Busy is not the same as profitable
The industry numbers suggest firms are feeling it. In Deltek's 47th Clarity study of nearly 900 architecture and engineering firms in the U.S. and Canada, median operating profit on net revenue fell to 16.7%, down from a ten-year high of 21.4% the year before. Utilization of labor cost on billable work slipped to just under 60%. According to Deltek, the cause isn't spare capacity; more time is going into coordination and administration that never gets invoiced.
Firms also sent out 32% more proposals while win rates fell. More pursuits, same senior people.
To be fair, 16.7% is not a crisis. Plenty of industries would take it. It is a squeeze, though, and it lands on the few senior people who run projects, review sets, write proposals, and keep clients calm.
Busy, but squeezed
-
Operating profit on net revenue (median)
21.4% down to 16.7%
Down from a 10-year high
-
Labor cost on billable work
Less than 60%
More time in coordination and admin that isn't invoiced
-
Proposals submitted
+32%
While win rates declined
Source: Deltek Clarity Architecture & Engineering Industry Study, 46th and 47th editions (2025, 2026).
What one point of margin is worth
Now put that at the scale of a typical firm. Zweig Group's 2025 valuation report describes its median surveyed firm as 45 full-time staff, with net service revenue of $7.7 million and pre-bonus profit of about $1.2 million.
At that size, each 1% of net revenue lost to unbilled correction is about $77,000. That's roughly 6% of the firm's annual profit, spent on work the client already paid for once. Three points is close to a fifth of the year's profit.
This is my arithmetic, not a finding from the report, and your numbers will differ. What matters is the proportion. For a firm this size, small leaks in production time add up to a large share of what the owners actually keep.
Where the hours go
Control time leaks out in a few predictable places:
• Senior review that happens late, all at once, and only by the principal.
• Redrawing, because the project was set up from scratch instead of from a firm standard.
• Owner changes absorbed "just this once" because nobody logged them the day they came in.
• Coordination rounds repeated because consultant backgrounds went out before anyone checked them.
• Pursuits won without asking who would actually deliver them.
None of this is dramatic, which is why it survives. Each item looks like diligence, good client service or a busy month.
What the timesheet shows, and what it doesn't
What gets invoiced
DesignProductionConstruction administration
Control time
- Late senior review
- Redrawing from a blank start
- Owner changes absorbed
- Repeated coordination rounds
- Work won without delivery capacity
FlowBIM.ai
Six ways to recover control of time
The firms that protect margin don't work less. They decide where checking happens, who does it, and how they'll know it's working.
1. Make correction time visible
You can't manage hours you can't see. Add a timesheet code for internal correction and re-coordination, separate from design and production, and ask the team to use it honestly. Tell them why it exists: to find process problems, not to find people to blame.
After two or three months you'll know which projects… A question I ask principals: how many times did your team review this set before it went to a reviewer or a consultant, and what were the most common issues? The answer shows you where to look.
2. Set review gates with an owner
A review that happens "when there's time" happens the night before the deadline. Define two or three gates per project, for example end of DD, pre-permit, and pre-bid, and name who owns each one. That person can hold a set that isn't ready.
Gates move review earlier and spread it out. An error found at the end of DD costs an afternoon. The same error found in permit comments costs a resubmittal, paid for in staff hours the fee never covered.
3. Take the founder out of the bottleneck
In firms of 10 to 50 people, the principal is often the firm's real QA system. That works until the firm grows. Then every set waits in one inbox.
Split review into two layers. A project architect runs a written checklist first: dimensions, clearances, schedules against plans, sheet coordination. The principal reviews what needs judgment: design intent, risk, what this particular client cares about. The checklist catches what any trained person can catch, and the principal's hours go to the calls only they can make.
4. Stop starting from zero
A lot of redrawing is set-up debt. When every project begins with a new template, new families and a new detail set, the team rebuilds decisions the firm already made.
A firm standard doesn't have to be big. A project template, a vetted library of doors, walls and assemblies, the details you actually reuse, and a short start-up checklist cover most of it. In Revit, that means your template and families already carry the parameters your schedules and checks rely on. Every project that starts from the standard skips a round of correction later.
5. Log scope changes the day they happen
Many firms don't lose margin on the work they priced. They lose it on work they never priced. An owner asks for one more option, a consultant changes a system, the program shifts after DD, and the team absorbs it.
Keep a simple change log on every project: date, request, who asked, estimated hours, in scope or not. Look at it in the monthly project review. You won't bill every item, and some you shouldn't. But you'll decide on purpose, and when the additional services conversation comes, you'll have a record instead of a recollection.
6. Put delivery capacity into go/no-go
Winning work you can't staff is one of the fastest ways to create control time. The team rushes, review gets compressed, and correction moves downstream.
Add one question to every go/no-go: who will run this project, and what comes off their plate if we win? If nobody can answer, the pursuit carries a cost the proposal doesn't show.
A monthly project review
Six questions, one page, every active project.
- Work
- What's done, what's left?
- Budget
- Hours spent vs. planned?
- Schedule
- Next milestone on track?
- Fee
- Billed vs. earned?
- Correction hours
- How much went to redoing work?
- Open changes
- What's logged, what's decided?
FlowBIM.ai
Where to start
You can begin this month with the tools you already have:
1. Add a correction and re-coordination code to timesheets, and explain to the team what it's for.
2. Pick two active projects and name a gate owner for the next milestone on each.
3. Write a one-page checklist for your project architects, built from the issues your principal keeps catching.
4. Open a change log on both projects and review it in a monthly meeting on work, budget, schedule and fee.
By the end of the quarter, you'll have your own number for control time and a clear idea of where it comes from.
What changes
Recovering control time doesn't mean checking less. It means checking earlier, in a set order, by people with the authority to stop a set. The senior hours you get back go to design, clients, and the next pursuit instead of late-night redlines.
The knowledge to catch these issues is already inside your firm. The work is placing that control inside the production process, phase by phase, so it no longer depends on one person's evenings.
If you suspect your margin is leaking through hours nobody invoices, the first step is finding out where. Start with a Production Control Diagnostic™.
Frequently asked questions
How much control time is normal? I don't know of a reliable public benchmark for correction hours in design firms, and I'd be careful with anyone who quotes one. The number that helps you is your own, tracked for a few months and compared across projects and phases.
Won't a correction code make my team defensive? It can, if it's used to assign blame. Present it as a way to find process problems, share results by project rather than by person, and act on what you learn. Once people see a new standard or checklist come out of it, the tone changes.
Our principal likes to review everything. Is that a problem? Not while the firm is small and the principal has the time. It becomes one when sets wait days for review, or when redlines crowd out pursuits and client work. Two-layer review keeps the principal on the decisions that need their experience.
Isn't this just better project management? Partly. Project management tracks budget and schedule. Control time is about how production is organized: where errors start, where they get caught, and who owns the check. The two work best together.
Do we need new software for this? No. Timesheet codes, a checklist, a change log and a solid Revit template get you most of the way. Software helps once you know what you're measuring.
Recover the margin you're already losing.
We find where unbilled control time leaks, and build the system that stops it.