The boutique consultancy math: why I take four clients a year, not forty

My calendar app says I had 47 "introductory call" slots last year. I took 38 of them. I signed contracts with four. This isn't a complaint — it's just the math. Here's how I think about it.

People look at four-clients-a-year and assume it's positioning. A choice you make to look exclusive. It isn't. It's the upper bound of what one person plus a small contracted bench can deliver, if the work itself is supposed to be the reason the client called.

Let me show you where the number comes from.

The supply side

A serious engagement at my end of the market is not a three-week sprint. It's a multi-month build with a discovery period in front of it and an aftercare tail behind it. The standard shape, almost regardless of client, looks like this:

• Pre-contract phase: roughly 3 months. Scoping calls, written discovery brief, security and procurement review on the client's side, contract negotiation, signature.
• Contracted build: 6 months. The core engagement. Two days a week of my direct time on average, with a contracted senior engineer for execution depth.
• Aftercare: 1 month. Stabilization, paging coverage, internal team enablement.

That's ten months of calendar per engagement. There are twelve months in a year. The math is what the math is.

Four parallel engagements is the maximum. Some of those overlap — one is finishing aftercare while another starts discovery — which is the only reason the number is four and not two.

The demand side

47 leads in a year is not a small number for a solo operation. About one a week. They come from:

• Referrals from previous clients (the largest source — maybe 60%).
• Referrals from people who were almost clients (the second largest — maybe 20%).
• Public writing. This blog, talks, the occasional open-source contribution.
• A small amount of cold inbound from companies that found me through someone who once worked with me at Stork.

I took 38 of the 47. The 9 I didn't take were either obviously bad fits — "we need a developer for $40/hour" — or the kind of pitch where the lead can't be bothered to describe the problem in an email and instead asks me to "jump on a quick call."

Of those 38 calls, I signed four. That's an 11% sign rate, or roughly 1 in 9. It's high for an industry where the average sits closer to 1 in 30. The reason it's high is that the calls themselves are already filtered. I don't take calls unless I already have a reasonable belief the work is real.

The rejection side, which nobody talks about

34 of those conversations ended in a "no." Sometimes my no, sometimes theirs, occasionally mutual.

The ones that hurt are not the budget mismatches. Those are clean. Someone has $40K, the work is a $400K problem, we shake hands and move on.

The ones that hurt are the ones where the fit is right, the work is real, and the answer is still no — because I don't have the capacity. There was a startup founder I spent two hours with last September. The problem was interesting, the team was strong, the budget was there, the timeline was reasonable. And I had to write him an email that started, "I can't take this on. Here's who you should talk to instead." I gave him three names.

Two of those three names took the engagement. He sent me a thank-you note three months later. That note is taped, metaphorically, above the desk where I write these.

But the mental cost of writing the no isn't zero. This is the part of the business nobody warns you about. You spend two hours building a relationship with a person you will then disappoint. Doing that 30+ times a year is its own kind of weather.

The economics

Let's put the numbers on the table.

Scenario A — the actual one
4 clients × $200K average engagement = $800K revenue
Gross margin ~40% (contracted engineers, tooling, dead-deal time, taxes)
Net: ~$320K
 
Scenario B — the volume one
40 clients × $20K average engagement = $800K revenue
Gross margin ~60% (less subcontracting, less overhead)
Net: ~$480K

On a spreadsheet, Scenario B looks better. $480K beats $320K. If you stop there, you'd conclude that boutique consultancy is irrational and high-volume is the better business.

Here's why I don't do Scenario B.

Forty clients a year is roughly one new engagement every nine days. The sales overhead alone — discovery calls, proposals, contracts, kickoffs — eats a third of the calendar before you write a line of code. The context-switching cost is real and invisible on the spreadsheet. You cannot hold the architecture of forty different systems in your head at the resolution that produces good work.

At forty clients, I'm running a small agency. That's a fine business. It's not the business I'm running.

At four clients, I'm doing the work. I know the system. I'm in their Slack. I'm the person in their architecture review. What I sell is my actual attention, not a project manager's calendar.

Pricing psychology

A four-client-a-year shop charges per project, not per hour. There's no hourly rate on my proposals. There's a number for the whole thing.

Discovery is free for the first week. Not the first call — the first week of structured work. The output is a written document that says, "here is what I think the problem is, here is what I would do about it, here is what it would cost, here is what could go wrong." If the client says no after that, we both walk away. They keep the document. I keep the time.

This sounds expensive. It isn't. The free week of discovery filters out clients who aren't serious before they cost me a month of build time. A client who refuses to spend a week defining the problem will refuse to spend six months solving it.

After the discovery document, the commitment is contracted. Fixed scope, fixed price, fixed timeline. Change orders exist and are priced separately. There is no hourly billing.

The honest receipt

Four is the average, not the rule.

In 2023 I signed five. There was a quarter where two engagements overlapped more than I had planned, and the calendar got ugly for about three months.

In 2024 I signed two. That was the year of the Stork equity transition. I deliberately took less work because I knew the equity conversation would eat a hundred hours I hadn't budgeted.

In 2025 I signed four. Boring by design.

Average it over five years and the number lands at roughly 3.8. That's why I say four when someone asks.

The counter-intuitive part

Scarcity is a moat.

Saying no to 90% of leads isn't a market-positioning trick. It's the actual reason the 10% pay what they pay. They aren't paying for hours. They're paying for the fact that you said no to nine other people that month. The implicit message is: "If I took your money, I wouldn't have time to do anything else. So if I am taking your money, I am giving you everything."

This is impossible to fake. You can't say you're exclusive while signing every contract that lands in your inbox. The price drops to match.

The four-client number isn't a target. It's an outcome. I've never sat down and decided to take exactly four. The fifth always shows up. I just don't have the time — and when I say that, I mean I look up at the calendar and there is genuinely no slot left.