Capacity versus continuity.

There's a category error baked into how most service businesses think about growth.

Capacity and continuity look like the same thing — both mean I can take on more work. They are not the same thing. They are economically opposite, and conflating them is why most consulting firms reach a certain size and then quietly stop being good at the thing that made them worth hiring in the first place.

This piece is about why.

Two definitions

Capacity is a quantity of work-hours that an organization can deliver. The unit is fungible — one engineer-hour is roughly substitutable for another engineer-hour. The scaling vector is hires. More people, more capacity. The math is linear: revenue per hour times hours available.

Continuity is one mind held on a problem from brief to ship. The unit is non-fungible — this specific person's specific judgment, applied to your specific problem, over a sustained period. The scaling vector is not hires; it's tools. More work, more demand on the operator. The math is fundamentally bounded: one person can only hold so many active engagements before the thing they're selling — sustained context — degrades.

What each model sells

The agency model and the operator model are not different sizes of the same business. They are different products.

The agency sells bodies-of-work. You hand them a brief, they deliver an output, the engagement ends or rolls into the next one. The deliverable is the unit. The relationship is project-shaped, with a beginning and an end.

The operator sells access to a specific judgment. You don't hand them a brief once; you call them whenever the problem changes. The relationship is open-ended. What you're paying for is the right to ping a specific person about your specific situation and get a useful answer back, week after week, year after year.

These two products look similar from a distance. They sound similar in marketing copy — we deliver excellence on engagements. The buyer thinks they are choosing between two firms in the same category. They are not. They are choosing between two fundamentally different services.

Most relationship breakage in consulting comes from this: the buyer thought they were buying continuity, the seller thought they were delivering capacity, and three months in the gap shows up. The senior who pitched is suddenly only on quarterly check-ins. The delivery team isn't the one that was promised. The context the buyer thought they had bought is being relearned, badly, by a different person. (See also why I unhired five people — the personal version of this exact failure mode.)

The pricing structures that emerge

You can predict which model a service business runs based on its pricing structure, because pricing structures emerge from what's being sold.

A business selling capacity tends to price one of three ways:

• Time and materials. Hourly billing. The buyer pays for hours; the more hours, the more revenue. The seller's incentive is for the work to take longer, not less time.
• Fixed-price project. A scope is agreed upfront with a fixed price. The seller's incentive is to deliver the minimum that meets the spec, because anything more eats margin. The buyer's incentive is to scope-creep within the spec.
• Retainer. A monthly fee in advance for an unspecified amount of work. The seller's incentive is to not do the work, because every hour spent reduces the retainer's effective margin.

A business selling continuity prices one way: subscription. The fee is for access — to the operator, to the platform, to a specific response time. The work that gets done within the access window is whatever needs doing. The seller's incentive is to make the work small and right, because they have a fixed ceiling on how much they can deliver and a strong interest in not eating their own time. The buyer's incentive is the same: make the work small and right, because they're paying for access whether they use a lot of hours or a few.

Subscription is the only pricing structure where buyer and seller are pulling the same direction.

Why investors prefer capacity

Capacity is investable. Continuity is not.

Capacity is a unit that fits in a spreadsheet. Hire 10 more engineers, deliver 10 more units of work, revenue scales by some multiple. The growth model is mechanical. Investors can model it, board members can question it, founders can present it. There is a flywheel.

Continuity has no flywheel. The product is bounded by one human's ceiling — typically 60 to 100 hours of focused work per month, depending on how much oxygen the person needs. You cannot 10x continuity by hiring. You can build tools that absorb the toil that surrounds the work. You can raise prices. But you cannot scale the unit itself, because the unit is the operator.

This is why agency businesses get funded and operator practices don't. It's also why the operator practice has lower overhead, no board, no growth pressure, and stays the size that works. The constraint that prevents investment is the same constraint that lets the work stay good.

The agency is a business that runs on growth. The operator practice is a business that runs on margin and time.

Defensibility

The interesting question is which model is defensible.

A new agency can be founded next year. The same playbook works: hire some experienced people, get a few clients, market the partners' brand, scale headcount. The competitive advantage of an agency is mostly in: brand recognition, partner network, client list, recruiting pipeline. Each of these is replicable. The agency model is hard to scale and hard to defend.

A new operator with my exact 20 years of context cannot be founded. They can be raised over 20 years, but not next year. The competitive advantage of an operator is the operator: their specific stack of experience, their specific network, their specific way of seeing problems. None of this is replicable in less time than it took to build.

Continuity scales poorly but defends well. Capacity scales well but defends poorly. They are inverse properties.

This is why most consulting firms try to have both, and why none do.

The math

Run the numbers.

An agency of 100 people, $20M annual revenue, 30% net margin. Sounds great. Now subtract: salaries (largest line, 50%+ of revenue), real estate, benefits, software licenses for 100 people, recruiting, HR, sales team, partner draws. The 30% margin is hard-fought and depends on keeping utilization above 70% across the team.

An operator practice. One person, $400K annual revenue, 80% net margin. Sounds smaller. Now subtract: tools and infrastructure ($10K/year), software for the operator ($10K), insurance and legal ($15K), maybe a part-time contractor ($30K). What's left is the better part of $300K in margin, going to one person.

The agency does $6M of margin spread across 100 people, hundreds of clients, an enormous operations footprint. The operator does $300K of margin against zero overhead, hands-on with five to ten clients.

Per hour of work invested, the operator wins. Per dollar of revenue, the agency wins. Per hour of the founder's attention to the work itself, it's not close — the founder of the agency probably hasn't touched code or shipped an engagement in years, while the operator does both every week.

These are different businesses, with different definitions of success. Neither is wrong. They are answers to different questions.

When each model is correct

Capacity is the right answer for:

• Standardized work — feature implementations of well-defined specifications, content production at volume, QA testing, design system component output.
• Well-understood problems — work that can be specified and delivered without sustained context.
• Fungible deliverables — outputs where the value is in the artifact, not in the relationship.
• Low-context engagements — short timelines, transactional, no need to learn the client's specific operating environment.

Continuity is the right answer for:

• Specific judgment problems — should we build this? which fraud detection approach fits our risk profile? what should the AI strategy be next year? — the answers are unique to the client and depend on understanding them deeply.
• Unfamiliar domains — work where the client's domain is new to the consultant, and the consultant's value comes from holding it long enough to develop a useful mental model.
• Multi-year relationships — engagements where today's decisions are downstream of last year's decisions, and where carrying the context is half the value.
• Work where context compounds — the same engineer in year three is more valuable than the same engineer in year one to the same client, because they know the system.

Most clients have both kinds of work. The mistake is not picking one model. The mistake is buying one and getting the other.

The hybrid lie

Many agencies advertise that they offer both — senior consultant oversight on top of delivery team execution. The pitch is: you get the senior partner's continuity, plus the firm's capacity to actually deliver.

This is structurally impossible at scale.

If a senior partner is overseeing 50 accounts, they cannot give continuity-level attention to any of them. The math doesn't work. Continuity requires sustained focus on a specific account; 50 accounts means at most one hour per account per week, which is a status meeting, not judgment-applied-to-the-problem.

What actually happens is that the senior partner becomes a sales role. They show up at kickoffs, at quarterly business reviews, and on escalations. The day-to-day delivery is done by associates and managers who don't have the senior partner's experience or context. The "oversight" is a quality-control review of work the senior didn't do.

The marketing is continuity. The reality is capacity. The gap is what eventually frustrates the buyer.

This is not a moral failing of agencies; the math doesn't allow it to work otherwise. If you charge agency rates, you must deliver agency-shaped work. If you want senior-partner continuity, you must hire a senior partner directly, in a structure where they can actually give it to you.

How to know what you're buying

Four diagnostic questions, when evaluating any service business:

1. Will the person who pitches the work also do the work? Or will the engagement be staffed by a different team after the contract is signed? The honest answer is sometimes "no, our delivery team will own it" — that's fine if it's stated clearly. The dishonest version is implying the senior partner will be involved when they won't.
2. Six months in, who answers your "small question" email? A senior who knows your context, or an account manager triaging it to a delivery team? This is the strongest signal of which model you're actually inside.
3. Is the pricing published? Or does it depend on a discovery call, a custom quote, or your willingness to negotiate? Capacity pricing is custom by necessity. Continuity pricing is published by necessity. The choice is structural.
4. Can you see the work in real time, or is reporting batched? Continuity practices tend toward radical transparency — same numbers I see, you see — because there's nothing to hide and showing the work is the work. Capacity practices tend toward managed reporting — quarterly business reviews, dashboards curated for the executive sponsor — because the underlying work is being done by people the buyer never meets.

The answers won't always match the marketing. They will always match what you actually get.

The closing argument

Capacity is what's easy to scale. Continuity is what's hard. Most service businesses scale capacity, because that's where the venture capital, the team-building playbooks, and the conventional wisdom point.

A few service businesses scale continuity, by refusing to scale headcount, by absorbing toil with tools, by pricing for access rather than output, and by accepting that they will be smaller and more focused than they could otherwise be. They serve a different market — buyers who need specific judgment more than they need predictable throughput.

The wrong choice is not should I run capacity or continuity. The wrong choice is selling one and delivering the other.

— Mikkel, Bangkok