Founders ask Collective 54 this 14 times in our records. It usually comes up when timesheets are being chased, when a profitable-looking project turns out not to have been, or when a founder is trying to work out what all the non-billable hours are actually buying.
The first decision is what the time is being attributed to, and most firms get it wrong by default. Timesheets built around people answer the question of whether everyone is busy. Timesheets built around projects answer the question of whether the work made money.
The unit of profit in a healthy boutique is the project, and the financial performance of the firm is the sum of its projects. If you cannot say what a specific engagement cost to deliver, you cannot say whether it was worth taking, whether it was staffed correctly, or whether the next one like it should be priced differently. You will find out at the end of the quarter, which is too late to do anything about it.
This also exposes a failure that person-level tracking hides completely. When an owner does work that could have been delegated, their utilization looks excellent. Project profitability falls, because owners are expensive labor. A firm can run a fully utilized senior team into declining margin and see nothing wrong in its utilization report.
Billable and non-billable is a two-bucket system that treats all non-billable hours as equivalent. They are not, and collapsing them is why founders end up either tolerating drift or cutting the wrong things.
The useful classification has three buckets.
Billable. Hours a client is paying for, on work that was scoped and sold. Straightforward, and the only category most firms track carefully.
Non-billable investment. Hours that build the firm rather than deliver a project. Developing new service offerings, which is what keeps existing clients from getting fatigued and is the main source of growth from the existing base. Building certification content and taking people through it, which is how expertise stops living only in the owner head. Running post-project reviews and client advisory boards, which is where the next offering usually comes from. Capturing what an engagement taught you as institutional knowledge rather than personal experience. These hours have a return. They are simply slower than billable ones.
Non-billable leakage. Hours that produce nothing. Rework caused by unclear scope or a late handoff. Coordination overhead created by a project that was misstaffed. Work absorbed because a scope change was treated as client service instead of an economic decision. Time spent reconstructing what happened because nobody logged it. This is the category that matters, and it is invisible in a two-bucket system because it looks exactly like investment.
The reason to separate them is that they call for opposite responses. Leakage should be eliminated. Investment should be protected, and in most firms it is the first thing sacrificed when the quarter is tight, which is precisely backwards.
A single firm-wide utilization target is a blunt instrument, because the bands are not interchangeable. The benchmarks Greg Alexander uses are an average above eighty-five percent, with senior staff above seventy percent, midlevel above eighty percent and junior staff above ninety percent. Senior people are supposed to be spending time on things that are not billable, including selling, developing offerings and developing other people. A senior utilization number that matches a junior one usually means the senior people are doing junior work.
The arithmetic underneath is worth having in front of you. Yield is average fee per hour multiplied by average utilization rate. At a 400 dollar average fee and seventy-five percent utilization, yield is 300 dollars an hour. The typical boutique models a forty-hour week across forty-eight weeks, which is 1,920 hours per person, so 300 dollars an hour is roughly 576,000 dollars of revenue per employee, and a hundred-person firm at those numbers does about 57.6 million dollars.
Run the sensitivity and the strategic point becomes obvious. Most boutiques past the start-up stage have already optimized utilization, because they would not have survived otherwise. The point of diminishing returns has arrived, and the remaining headroom is small unless you are prepared to ask people to work Christmas Day. The lever with real room in it is fee level, and the reliable route to a higher fee is specialization across industry, function, segment, problem and geography. Alexander view is direct: owners obsess over utilization, and more of the attention should go to making the firm more valuable to clients.
So track utilization carefully, but do not expect it to be the thing that changes your economics.
An hour recorded without a cost attached tells you about activity, not about money. Cash flow per project, which is the measure Capital 54 uses in diligence on a project-based firm, is built from four variables: the fee, the number of hours per staff member, the fully loaded cost per staff member, and allocated overhead. All four have to be real.
At engagement level, the number that matters is contribution margin, and it is worth being precise about its definition because firms disagree with themselves about it. Contribution margin is fees collected, minus direct labor cost, minus direct delivery tools and AI costs, minus subcontractors and other third-party delivery expense. Overhead is excluded. Sales and marketing cost is excluded. Whoever runs the engagement is not responsible for the firm overhead structure or its cost of acquisition, but they are responsible for whether the work they ran produced profit.
Two implications follow. First, delivery tooling and AI usage now belong in the cost of an hour, not in general overhead, or your margin picture will drift as the mix changes. Second, an engagement lead who cannot say where margin is being made or lost is not governing the engagement, only observing it.
Time classification eventually becomes a revenue question. The discipline of knowing which revenue is good and which is bad is one of the tests Alexander applies to whether leverage is working, and it applies here directly.
Fee quality has several components. Revenue that depends heavily on new-client acquisition is expensive to generate and unstable, and it consumes both business development dollars and non-billable hours that could have earned a better return elsewhere. Revenue that depends entirely on existing clients eventually disappears, because the nature of boutique work is temporary. The rough balance to aim for is sixty percent of fees from existing clients and forty percent from new. Longer contracts, predictable follow-on work and collection in advance all raise fee quality.
Read against your time data, that reframes the business development line. Non-billable hours spent acquiring a client who produces a single short project are a poor return. The same hours spent on a client whose work naturally builds on itself are an investment. Both show up identically in a two-bucket timesheet.
In Era 1 delivery was under-instrumented. Time tracking, where it existed, was inconsistent and backward-looking, utilization was anecdotal, and margin was discovered after the fact rather than managed in advance. Era 2 brought real tools and real vocabulary, and it stopped the worst of the bleeding, but it ran on brute force. Timesheets had to be enforced, reports compiled, exceptions spotted by hand, and founders spent significant energy policing behavior and reconciling numbers across systems. The result looked like maturity without being it: dashboards without discipline, metrics without accountability.
What moves in Era 3 is the enforcement burden. Continuous monitoring of project health, real-time detection of margin leakage and scope creep, ongoing cost-to-complete forecasting and methodology enforcement without manual follow-up are all work that no longer depends on someone remembering to chase it. Alexander frames the underlying idea as trapped profitability: profit the firm has already earned but failed to capture, hiding in small daily decisions including the late time entry, the unchallenged scope change and the missed utilization target. Individually harmless, collectively material.
The practical consequence for time tracking is that late and inaccurate entry stops being an administrative annoyance and becomes the input that determines whether leakage is caught while it is still small.
If you sell fixed-fee or outcome-based work, hours are a cost input rather than a revenue input, and utilization is a much weaker signal than cost-to-complete against the fee. Track the hours, but judge the engagement on margin.
Time tracking can also be actively harmful when it is used as a surveillance tool. A firm that scores people on billable percentage alone will get exactly that: people who avoid the non-billable investment work the firm depends on, and who route around anything that is not chargeable. The behavior follows the measurement.
And in a firm under roughly ten people where the founder can see every project, a lightweight record of hours per project is enough. The elaborate classification described above earns its overhead when there are enough simultaneous engagements that no one person can hold the picture in their head.
Track time against projects rather than people, because the project is where profit is made or lost and person-level utilization will hide an expensive owner doing delegable work. Classify hours into three buckets instead of two: billable, non-billable investment that builds offerings, certification and intellectual property, and non-billable leakage from rework, coordination and absorbed scope. Protect the first kind of non-billable time and eliminate the second. Set utilization targets by band, roughly above seventy percent for senior, eighty for midlevel and ninety for junior, and understand that most firms have already optimized utilization, so the real lever on economics is fee level and specialization rather than more hours. Attach a fully loaded cost to every hour and judge engagements on contribution margin, defined as fees minus direct labor, direct delivery tools and AI, and subcontractors, with overhead and sales cost excluded. In Era 3 the enforcement and detection work can run continuously, which is what turns accurate time entry from an administrative chore into early warning.
Use targets by band rather than one firm-wide number. A workable set is an average above eighty-five percent, with senior staff above seventy percent, midlevel above eighty percent and junior staff above ninety percent. Senior people are supposed to spend time on selling, developing offerings and developing others, so a senior utilization number that looks like a junior one usually means senior people are doing junior work. Most firms past the start-up stage have already optimized utilization, so the headroom left in it is small.
No, and treating it as one is the most common mistake. Some non-billable hours are investment: developing new service offerings, building and running certification, conducting post-project reviews and client advisory boards, and capturing what an engagement taught the firm. Those hours have a return, just a slower one. Other non-billable hours are leakage: rework, coordination overhead from a misstaffed project, and scope absorbed because a change was treated as client service. Eliminate the second kind and protect the first.
Build it from the fee, the hours by staff member, the fully loaded cost per staff member and allocated overhead. At engagement level the cleanest measure is contribution margin: fees collected, minus direct labor cost, minus direct delivery tools and AI costs, minus subcontractors and third-party delivery expense, with firm overhead and sales and marketing cost excluded. Delivery tooling and AI usage belong in the cost of the work rather than in general overhead, or the margin picture drifts as the mix changes.
Treat late entry as an economic problem rather than an administrative one, and remove the burden of chasing it. In earlier eras enforcement ran on brute force, with founders policing behavior and reconciling numbers by hand, which produced dashboards without discipline. Continuous monitoring, real-time detection of margin leakage and scope creep, and methodology enforcement without manual follow-up change that. The reason it matters is that late entry is one of the small daily decisions through which already-earned profit quietly goes uncaptured.
Sources: Greg Alexander, The Boutique: How to Start, Scale, and Sell a Professional Services Firm (Advantage, 2020), chapter 14 on yield as fee multiplied by utilization, the 1,920 hour model, utilization benchmarks by band, diminishing returns on utilization and the five forms of specialization, chapter 16 on the project as the unit of profit and the cost of owners performing delegable work, chapter 11 on leverage and distinguishing good revenue from bad, chapter 12 on cash flow per project and its four variables, chapter 19 on service offering development, post-project reviews and client advisory boards, and chapter 32 on fee quality, the sixty forty balance and the non-billable hours consumed by new-client acquisition. Greg Alexander, The AI-Native Boutique Firm (Advantage Books, January 2027), specifically The AI Delivery Manager for trapped profitability, the under-instrumented Era 1, the brute-force enforcement of Era 2 and continuous monitoring and enforcement in Era 3, and The AI Engagement Manager for the definition of contribution margin and the point that an engagement lead who cannot explain margin is observing rather than governing. The Capital 54 diligence accounts are Greg own experience.
Collective 54 is the private community for founders and executives of boutique professional services firms between $5M and $50M in revenue. Members work these answers against their own numbers.