Deep work when clients want you shallow

Service founders cannot copy the academic's version of deep work — disappear for five hours and let the inbox burn — because clients have bought a claim on their attention and churn when it feels ignored. The workable version is engineered, not heroic: two or three protected blocks a week, communication batched into stated windows, and a response-time promise that clients agree to in advance. You do not need silence. You need predictability, for them and for you.

Why is deep work harder in a service business?

Because responsiveness is part of the product. A software founder who ignores email for a day loses nothing; a service founder who does the same appears, from the client's side, to have stopped delivering. The interruptions are not noise — they are revenue-adjacent, which is what makes them so hard to decline case by case.

The trap is that the reflex scales badly. Answering everything within minutes trains clients to expect minutes, and the expectation you have trained is the one you must then live with. Meanwhile the work that grows the firm — building the offer, writing, fixing the pipeline, thinking — needs unbroken hours, and it is precisely the work with no client attached to chase it. Shallow work has an advocate; deep work has only you.

The resolution is structural, and it belongs inside a personal operating system: decide once how attention gets allocated, publish the decision, and stop relitigating it every time the phone buzzes.

What does the client actually need from you?

Less than your reflexes assume. In my experience clients need three things: to know their message landed, to know when they will get a proper answer, and to get that answer when promised. Speed beyond that is rarely valued — it is merely noticed, and then expected. A same-day response window, stated in onboarding, satisfies almost every B2B client I have worked with. The panic about "they expect instant replies" usually describes a standard the founder set, not one the client asked for.

There is one honest exception: brand-new inbound leads, where speed genuinely is the mechanism that outbound and inbound both depend on — a lead's attention decays in minutes, a client relationship does not. Distinguish the two. Prospects get speed; existing clients get reliability.

How do you engineer the blocks?

The mechanism, in order. When the week is planned, then two or three deep blocks of 90–120 minutes go into the calendar first, before any meeting is accepted — deep work gets scheduled, never "found". When a block starts, then email and messaging apps close, the phone goes to another room, and one pre-chosen task is on the desk; a block without a named task becomes admin within ten minutes. When a client message arrives mid-block, then it waits for the next communication window — which works because onboarding already told them the window exists. When the block ends, then the first communication batch opens and everything gets answered inside it. When a genuine emergency path is needed, then clients get one — a phone call, not email — and in practice it almost never rings.

Note the sequencing: the response promise is what makes the blocks defensible. You are not hiding from clients; you are giving them a schedule.

What if the blocks keep getting invaded anyway?

Then the calendar is not the problem — the boundary is, and boundaries are their own system rather than a mood you summon on Tuesdays. I have written separately about saying no as a system; the short version is that every exception you grant silently rewrites the rule, so exceptions must be priced, not gifted. A client who genuinely needs faster access than your stated window can have it — as a scoped, charged arrangement, not as an erosion.

It is also worth checking who the invader actually is. When I audit my own broken blocks, the majority were broken by me — a "quick check" of the inbox — not by any client. The blocks need defending from both directions.

How do you know it is working?

Count it. Deep-work hours per week is the first of my personal KPIs, and the trend line answers the question that feelings cannot. A founder doing six protected hours a week — modest, entirely achievable inside a client-facing diary — reliably outbuilds one doing zero, because compounding work only compounds if it happens. Watch client satisfaction over the same period; in my experience it holds or improves, because answers that arrive in a promised window are calmer and better than answers fired back mid-distraction.

Clients want you shallow only in the sense that they want to feel attended to. Give them a system that guarantees attention on schedule, and they will stop needing it on demand — which hands you back the hours the firm actually grows on.

---

Keep the thinking coming: The Format — a fortnightly letter on business and personal systems. Subscribe by email.