The delay your client buys

Why this matters to you now

On March 31, 2026, Roelof Botha and Jack Dorsey published an essay on hierarchy that opens with an investor's observation: at Sequoia, speed is the best predictor of start-up success. What follows is more interesting than the line. They take the org chart apart and show that it is not a management philosophy at all, but an information routing protocol two thousand years old. The Roman army grouped eight soldiers under a single leader, ten of those groups made a century of eighty men, and so on up to a legion of roughly five thousand. The constraint that produces the shape is human and has not moved since: one leader can effectively manage between three and eight people. Tighten the span and you add layers. Every layer slows information down.

Nobody has broken that tradeoff in two thousand years. It has simply been paid, and it is still being paid, by every organization your clients run and every firm you bid against. That is where your situation changes. What the large firm sold was two things at once: access to scarce knowledge, and the capacity to process it. Artificial intelligence just levelled the first. A consultant working alone now has analytical depth that required a team three years ago. What is left in the client's comparison, once knowledge has become common, is the time between their question and an answer they can defend to their boss. And that time is precisely what hierarchy taxes.

Clayton Christensen spent his career repeating that clients do not buy a product, they hire something to get a job done. In Competing Against Luck, he and his co-authors show that the job is almost always narrower than the supplier imagines. Look at yours honestly. Your client is not hiring you to understand their market. They are hiring you to make a decision before the board meets on the 14th. Knowledge is the means. The deadline is the job.

That is why nobody buys your speed by name. Clients say "you get it fast," "you're easy to work with," "you answered the same day." They describe a feeling because they have no budget line for a turnaround time. This carries forward edition #27 on the consultant you can clone, moved one notch over: where #27 asked how to duplicate your judgment, this one asks how long that judgment takes to leave your head.

If a client has ever told you "we went with the big firm, but we called you back," it was not your credibility that won. It was your turnaround time. And you have probably never billed for it.

Pricing: turnaround is a version, not a surcharge

The move: Take speed out of your single price. One offer, two versions, two named turnaround times, two prices. Not a discount for clients willing to wait: two distinct products, one of which is called "your written answer is in your hands Monday."

Why it works: Rafi Mohammed, in The 1% Windfall, states the single-price problem in a sentence every consultant should reread each quarter: one price leaves on the table every client who would have paid more for something else. His answer is versioning. Of all the variables you could version on, turnaround is the only one the client can price themselves. They know what a week of waiting costs inside their organization; you do not. Post two turnaround times and you stop guessing at their urgency and let them declare it.

The trap: The rush fee. It looks like the same move. It is the opposite one. A surcharge billed after the fact punishes the client in a hurry, and worse, it teaches them something you do not want them to learn: that your normal speed is slow, and that they have to pay to correct a defect. A version posted in advance sells a capability. An improvised surcharge confesses a problem. The distinction runs back to edition #1 on why you are not too expensive: price is a positioning decision, never a reaction.

This week: Pull up your last proposal. Keep everything, add one line under the price: "written answer within five business days, plus 40%." Do not explain it, do not justify it, do not discount the other version. Send it as is to your next client and watch which box gets ticked. You will learn more about your market from one proposal than from three months of thinking about your rate card.

Sales and business development: clients compare turnaround, not resumes

The move: On your next sales call, swap one demonstration of competence for a demonstration of turnaround. Instead of listing three similar engagements, give the date. "You send me the documents Thursday, you have my written position next Wednesday." A date can be verified. Expertise cannot.

Why now: Matthew Dixon and Ted McKenna analyzed more than two million recorded sales conversations for The JOLT Effect, and their central finding moves the battlefield. Most lost deals are not lost to a competitor. They are lost to inaction. The client does not pick someone else, they pick nobody. What paralyzes them is not doubt about you, it is the fear of getting it wrong. A short turnaround acts directly on that calculation: the sooner the answer lands, the less irreversible the decision is at the moment they make it. You do not fight indecision with more proof. You dissolve it with time.

For years I blamed competitors for the calls I lost. One day I went back through the last twelve and checked what each of those prospects had actually done. Half of them had bought nothing at all, from me or from anyone. I had spent months benchmarking myself against competitors who had not won either.

The trap: Promising a short turnaround to one client and financing it with everybody else's work. A date you hold by working three evenings and pushing back two live engagements is not speed, it is borrowing. You will repay it with interest, in the form of delays elsewhere. Sellable speed comes from preparation done in advance, not from sacrifice made after the fact. That is the mechanic behind edition #24 on the offer that bleeds you: whatever you promise without tooling for it comes out of your margin.

This week: On your next proposal, delete "timeline to be determined" and write a calendar date. Do it before you know exactly how you will hold it. The constraint will force you to build tooling for what you improvise today, and that is the real work.

Collaboration networks: a network with no layer

The move: Stop keeping your network as a list of capable people and start keeping it as a reserve of reachable capacity. The question that qualifies a contact is not "are they good," or even "do they want to collaborate." It is: how many hours until they can answer me?

The mechanism: Frederic Laloux documented in Reinventing Organizations companies of several thousand people that operate without middle managers, and the protocol that makes it possible fits in one rule: the advice process. Before a decision, you must seek advice from the people affected and the people who know. You do not need permission, and nobody holds a veto. Whoever asked the question decides. It is the only known mechanism that adds competence to a decision without adding a layer above it. It is also, at bottom, what a well-connected solo practitioner already does without naming it.

The trap: The collaboration that starts with a scoping call. Sixty minutes to line up in two full calendars is five to eight days in practice, and you have just rebuilt, at your own expense, the layer you did not have. A fast network runs in writing, on closed questions, with the turnaround you want stated inside the request itself. This extends edition #2 on the network you neglect: a network gets cultivated, but it also gets measured, and the measure that counts here is a duration.

This week: Three messages, one question each. "If I write to you on a Tuesday with a specific question in your field, how long until you can honestly answer?" Those three replies will sort your network more usefully than any skills spreadsheet. And the one who answers "it depends" has just told you their turnaround time.

Value creation: the understanding that deepens

The move: Stop measuring your value by what you know. Knowledge is now the part of your work whose cost is collapsing. Measure it by what you understand that the machine does not understand yet, then check one thing and one thing only: is that understanding deeper this month than it was last month?

Why this changes everything: Botha and Dorsey end their essay on a question I find more useful to a consulting practice than to a technology company. What does your organization understand that is genuinely hard to understand, and is that understanding getting deeper every day? If the answer is nothing, they write, artificial intelligence is only a cost-cutting story. Marty Cagan says the same thing differently in EMPOWERED: you hand a team a problem, not a roadmap, because the roadmap is the limiting factor. For a solo practitioner the roadmap is called the service catalogue, and it limits you in exactly the same way. Understanding that deepens eventually produces offers. A catalogue of offers never produces understanding.

My engagement notes sat for years in folders filed by client. Tidy, professional, and completely unusable: I could find what I had told a given client, never what I had come to understand about a given problem. The day I started filing by question instead of by client, my answers got days shorter without my having learned anything new.

The trap: Confusing accumulating data with deepening understanding. A folder of three hundred documents is not understanding, it is a warehouse. The difference is reusability: understanding answers a question it has not been asked yet. That is what edition #6 on your methodology being your product meant by a methodology. What does not get reused is not one. It is a memory.

This week: Write one sentence and put it where you can see it. "My practice understands ______ better than anyone my client has access to." Fill the blank with something narrow enough to be true. If it takes you more than two minutes, you have just named the real work of your fall, and it is worth more than three engagements.

AI: shorten the delay, do not lengthen the deliverable

The reflex, when a tool makes production nearly free, is to produce more of it. That reflex is what will cost you your edge. Your gap against the large firm does not widen on volume: they will always produce more than you, and now they will produce far more. It widens on the interval between the client's question and an answer you are prepared to defend.

The move: Reserve the machine for the work that shortens that interval, and for nothing else. Not producing more, not producing faster what nobody was waiting for: cutting the time between the question and the defensible position. Five uses, all aimed at that one number.

1. 1.The position builder. On your first pass through a file, submit what you have and ask for three things: the most probable position, the two or three facts that would flip it, and what is still missing to decide. You do not get an answer, you get a map of the work remaining. That is what turns three weeks into four days.

1. 2.The challenger. Once your position is written, have it attacked. Ask for the other side's case using its strongest points, not its weakest. A position that has already survived its own rebuttal goes out sooner, because the second-guessing that was holding you back has already happened.

1. 3.The real question. Paste the client's email as is and ask: what decision does this person have to make, and by when? You will often answer a narrower question than the one you were asked, which is both faster and more useful.

1. 4.The position template. Take your last five short notes and ask it to extract the shared structure: which sections recur, in what order, at what standard of proof. The template that comes out is what makes a short turnaround sustainable without working evenings.

1. 5.The understanding log. At the end of every engagement, submit your notes and ask for a single page: what did this engagement teach me that is reusable at another client? File that page by question, never by client. It is the only one of the five that compounds. The other four win you days; this one wins you years.

The shift to make this year: Stop measuring what artificial intelligence lets you produce, and start measuring how much it shortens the interval between the question and your answer. The first number goes up for everyone, including the two-hundred-person firm. The second only goes up for those with no layers to cross, which is you.

The warning: A fast wrong answer destroys faster than a slow answer loses. Defensible speed comes from understanding already accumulated, never from analysis cut short. If you cannot name what makes your position solid, you are not fast, you are rushed. The machine shortens production; it does not shorten learning.

This week: Time one thing. On your next client request, note the exact hour the question arrives and the exact hour your defensible answer goes out. That is your real turnaround time, and you have probably never measured it. Everything else in this edition exists to bring that number down.