Founders ask Collective 54 this 9 times in our records, 3 of them in 2026. There are two numbers rather than one, and they are held in different parts of the building.
The published benchmark set for a boutique firm that intends to sell includes both of them: gross margins above 75 percent and EBITDA margins of 40 percent. They sit alongside revenue growth above 30 percent, more than twelve months of forward visibility, a year of payroll in cash on the balance sheet, no debt, and a five to ten year track record of consistency across all of it.
Those figures travel together and a single one in isolation proves very little, which is why the benchmark set is worth reading whole rather than quoting selectively. For the purposes of this question, though, the split between the two margin numbers is what matters.
Gross margin is a delivery and pricing number. It is what remains after you deliver the work. If it is short, the problem is in the cost of producing the service, the price you charged for it, or both. No amount of overhead discipline fixes it.
EBITDA is an overhead and sales number. It is what remains after gross margin funds overhead, which covers operations, marketing, finance, IT, legal and HR, and then funds sales, which covers sales management, account executives, account management and retention.
The arithmetic between them is the useful part. A firm at 75 percent gross margin and 40 percent EBITDA is running roughly 35 points of revenue across overhead and sales combined. If your gross margin is healthy and your EBITDA is not, you do not have a margin problem in the usual sense. You have a cost structure that is consuming more than 35 points, and the question becomes which of those functions should have been bought rather than built.
The first discipline, then, is simply this: do not set a margin target you cannot decompose. If you cannot say which of the two numbers you are missing, you are not ready to set either.
Forty percent is a sellability threshold. It is not an ambition, and treating it as one is now a way to undershoot.
The newer Collective 54 material draws the comparison plainly. Take two firms at twenty million dollars of revenue. A tech-enabled firm at 30 percent EBITDA produces six million. An AI-enabled firm at 60 percent produces twelve million. At the same multiple, that is double the exit price on identical revenue. The claim is not that multiples are expanding. It is that margins are.
That changes how to read the benchmark. The field you are measured against is moving, and a firm setting 40 percent as a destination in 2026 is setting a target that its peers may pass while it is being achieved. Set the floor at the published benchmark. Set the ambition against what the production model can actually support.
Now the second half of the question, which is the harder one.
Margin is created and destroyed at the project level, and a firm-level target laid on top of project-level variance is not a target. It is an average, and averages hide the two things you most need to see: which work is subsidizing which, and how much spread sits underneath a stable-looking blended figure.
Capital 54 has seen how much this matters in diligence. On one commercial photography firm the unit of measure was cash flow per project, built from fee, hours per staff member, fully loaded cost per staff member and allocated overhead. Some projects produced strong cash flow and others produced negative cash flow. The volatility was the disqualifying finding, because it meant the delivery model was not standardized and therefore not scalable. A blended margin would have looked acceptable.
So the target has to be set and monitored per engagement type before it means anything at firm level. Which is also the point at which a target stops being a statement of intent and becomes something you can miss on a Tuesday and know about on a Wednesday.
A margin target set after the work is sold is a report. A margin target set before is a constraint.
The Era 3 practice is to model economics before commitment rather than discover them after. Select the value metric the price is tied to, choose the pricing model, model expected and acceptable margins, forecast cost to serve across human, AI and tooling components, and test sensitivity to discounting and scope creep. Two engagements that model identically at signature can land very differently once a discount and a scope extension are applied, and knowing the size of that gap in advance is what lets you refuse the version that breaks the target.
Margin targets are not missed in one decision. They erode.
Treat pricing decisions as hypotheses and test them continuously by linking the quoted price to delivered effort, realized margin and client satisfaction. That loop is what turns a target into a control. Alongside it, four things need to be visible without relying on anyone to self-report: deviations from approved pricing, discounting as it happens, every exception recorded so that one-time decisions do not quietly become norms, and the moment when delivery effort begins to diverge from what the price assumed.
None of that is policing. It exists to protect margin integrity from erosion that nobody intended and nobody noticed.
Two structural levers sit behind the governance. Leverage, where a ratio of at least ten non-partners to partners is the floor for a firm that intends to scale, because owners performing delegable work is expensive labor charged to every project. And fee quality, because not all revenue supports the target equally: roughly 60 percent of fees from existing clients and 40 percent from new, contracts longer than twelve months, predictable follow-on work, and collection in advance rather than aging receivables.
There is one honest check on the whole exercise. A margin target set without a change to how the work is produced is a wish.
The milestone that actually indicates a firm can hold a higher margin is revenue growth exceeding headcount growth, sustained long enough to lift both gross and EBITDA margins. Top-line growth without profit growth proves only that you can sell. If your plan for reaching 40 percent consists of the same production model running harder, the number will not arrive, and the effort will show up as utilization pressure on people who are already fully utilized.
If you are deliberately building capacity ahead of revenue, expect margin to compress and do not correct it. That is margin you are investing rather than margin you are losing, and the project-level view is what lets you tell the two apart.
If you are genuinely an intellect firm, hired for never-before-seen problems, structurally low leverage may put 75 percent gross margin out of reach, and chasing it will push you toward work you should not take. Set the target from the model you actually run and make your case at exit on fee level, client ROI and cycle resilience instead.
And if you are inside a sale process, do not reset the target now. Buyers underwrite demonstrated performance, and a margin improvement with two quarters behind it reads as an experiment rather than a durable gain.
Target both numbers, not one. Gross margin above 75 percent and EBITDA at 40 percent are the published benchmarks for a sellable firm, they sit inside a wider set that includes revenue growth above 30 percent, twelve months of forward visibility, a year of payroll in cash and no debt, and they are held in different places: gross margin in pricing and delivery, EBITDA in what overhead and sales consume. The gap between them, roughly 35 points of revenue, is your combined overhead and sales budget, so a healthy gross margin with weak EBITDA is a cost structure problem rather than a delivery one. Treat 40 percent as a floor rather than an ambition, because the newer material puts a well-run AI-enabled firm near 60 percent, which on twenty million dollars of revenue is twelve million of EBITDA against six and double the exit price at the same multiple. To hold the number, set it per engagement type rather than firm-wide, since blended margins hide the variance that disqualifies firms in diligence. Model expected and acceptable margins before commitment and test sensitivity to discounting and scope creep. Then govern it: treat pricing as a hypothesis tested against delivered effort and realized margin, keep discounting and exceptions visible, hold leverage at ten to one or better, and protect fee quality. And apply the honest test, which is that a margin target without a change in the production model is a wish.
Two numbers, not one. The published benchmark set for a sellable boutique professional services firm includes gross margins above 75 percent and EBITDA margins of 40 percent, alongside revenue growth above 30 percent, more than twelve months of forward visibility, a year of payroll in cash, no debt, and five to ten years of consistency across all of it. The two margin figures are held in different parts of the business. Gross margin is a pricing and delivery number. EBITDA is a number about what overhead and sales consume. A firm can hit one and miss the other badly, and the correction sits in a different place each time.
It is the floor rather than the ambition. The newer Collective 54 material compares two firms at twenty million dollars of revenue: a tech-enabled firm at 30 percent EBITDA produces six million, an AI-enabled firm at 60 percent produces twelve million, which at the same multiple is double the exit price on identical revenue. The claim is that margins are expanding rather than multiples. So 40 percent remains the level at which a firm becomes sellable, but a firm setting it as a destination in 2026 may find its peers passing it while the target is being achieved.
Because margin is created and destroyed at the project level, and a firm-level target laid over project-level variance is an average rather than a target. Capital 54 saw this in diligence on a commercial photography firm, where cash flow per project was built from fee, hours per staff member, fully loaded cost per staff member and allocated overhead. Some projects produced strong cash flow and others produced negative cash flow, and that volatility was the disqualifying finding, because it showed the delivery model was not standardized and therefore not scalable. The blended figure would have looked acceptable.
Model it before commitment and govern it afterwards. Before: select the value metric, model expected and acceptable margins, forecast cost to serve across human, AI and tooling components, and test sensitivity to discounting and scope creep. After: treat pricing decisions as hypotheses tested by linking quoted price to delivered effort, realized margin and client satisfaction, and keep four things visible without self-reporting, namely deviations from approved pricing, discounting as it happens, every exception so one-time decisions do not become norms, and the point where delivery effort diverges from what the price assumed. Behind that sit leverage at ten to one or better and fee quality.
Sources: Greg Alexander, The Boutique: How to Start, Scale, and Sell a Professional Services Firm (Advantage, 2020), chapter 30 for the benchmark set covering a five to ten year track record of consistent growth, greater than 30 percent top-line revenue growth, more than 75 percent gross margins, 40 percent EBITDA margins, more than twelve months of forward visibility, one year of payroll in cash on the balance sheet and no debt, and for the finding that top-line growth without profit growth is a deal killer because the firm has not decoupled revenue growth from headcount growth; chapter 12 for cash flow as distinct from net income and EBITDA, for cash flow per partner and cash flow per project, and for the Capital 54 diligence on a commercial photography firm in which cash flow per project was built from fee, hours per staff member, fully loaded cost per staff member and allocated overhead, and the project-level volatility revealed a delivery model that was not standardized and therefore not scalable; chapter 11 for leverage as the ratio of non-partners to partners with ten to one as the floor for a firm that intends to scale; chapter 32 for fee quality, including the rough sixty-forty split between fees from existing and new clients, contract length beyond twelve months, fee predictability and collection in advance rather than aging receivables; chapter 21 for decoupling the rate of revenue growth from the rate of employee growth; chapter 13 for the intellect, wisdom and method firm types. Collective 54, The AI-Native Firm Map front matter, for the sequence in which gross margin funds overhead covering operations, marketing, finance, IT, legal and HR, the remainder funds the sales block, and what survives is EBITDA. Greg Alexander, Why Some Boutique Firms Exit Cleanly and Others Never Really Do (Collective 54), for the comparison of a tech-enabled firm at 30 percent EBITDA margins against an AI-enabled firm at 60 percent on twenty million dollars of revenue, producing six million against twelve million and double the exit price at an identical multiple, and for the position that the difference is margin expansion rather than multiple expansion. Greg Alexander, The AI-Native Boutique Firm (Advantage Books, January 2027), specifically The AI Pricing Manager for modeling expected and acceptable margins before commitment, forecasting cost to serve across human, AI and tooling components, testing sensitivity to discounting and scope creep, treating pricing decisions as hypotheses tested by linking quoted price to delivered effort, realized margin and client satisfaction, and for price integrity enforcement, discount visibility, exception tracking so that one-time decisions do not become invisible norms, and delivery alignment checks that detect when effort diverges from pricing assumptions; and The AI Finance Manager for the benchmarking blindness in which a 50 percent gross margin is believed strong and a 25 percent EBITDA margin impressive with no external reference point. Note on scope: the arithmetic identifying roughly 35 points of revenue between the gross margin and EBITDA benchmarks as the combined overhead and sales budget is derived here from those two published figures.
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.