The Wrong Instrument

Someone asks me what a website costs. I give them a number.

Then comes the pause, and then the same three words I have been hearing for thirty years. Wow, that's a lot.

Lately there is a second half to it. I could probably do that myself with AI in a couple of hours.

They are right about the two hours. The visible part of that job really has collapsed to about that. And the two hours have close to nothing to do with what they would be paying for.

A few weeks ago I finished Alan Weiss's Value-Based Fees. It did not change my mind, because my mind was already there. He has written that billing by any unit of time is inherently unethical, since the client is best served when the problem gets solved quickly and the provider is best served when it drags. Weiss is absolute on the question, and I agree with him.

For most of my career I assumed the number was the problem.

The part you can see

For every hour of my work a client can watch, there are three to ten they cannot. The number they are calling insane already understates what producing it cost me.

On a website, the part they watch is the page appearing. What they do not watch is the DNS records that decide whether the thing resolves at all, the certificate that expires in thirteen months, the CSS cascade that determines whether the layout survives a phone turned sideways, the decision about where the content actually lives that costs nothing to make today and a rebuild to reverse in year three, and the registrar transfer that will take an afternoon or take a month depending entirely on choices nobody wrote down.

I have spent a career on the principle that the best work leaves no fingerprints. I call it the invisible orchestra: a room where nobody thinks about the sound, and a system where nobody thinks about the system.

I still believe it. It is also the reason I keep losing on price.

A buyer prices what is in front of them. I have spent thirty years removing every trace that any work occurred, and then handed over a number and watched them do arithmetic on what was left.

The second table

A steak at the diner runs eighteen dollars. The same cut, three miles away, cooked by someone who has spent twenty years learning what to do with it, runs a hundred and eighty. Nobody at the second table asks how long it took. Nobody objects to the number either. The card goes down without a flicker, for a piece of beef that left the packing house identical to the one down the road.

The difference is preparation, which is another word for judgment. Somebody chose the ranch. Somebody learned, across twenty years, exactly how hot the pan has to be, which sounds like nothing at all until you have eaten one cooked in a pan that was not hot enough.

Everyone who has ever paid that bill already agrees with me. They agree completely at that table, and then refuse the same argument in my inbox on Monday morning.

And the second table did not arrive in that condition by accident. The room, the reputation, and the years somebody spent deciding what the place would be, all of it happened before the guest saw a price. By the time the bill lands, the argument about whether it is worth it has been settled somewhere else, long before anyone sat down.

Somebody taught them this

The habit of pricing labour by the clock has a birthday, or something close to one.

In 1748 Benjamin Franklin wrote a short letter of advice to a young tradesman. "Remember," it opens, "that time is money." The line outlived everything around it, and the sentences underneath it say something much narrower. Franklin's example is a man who can earn ten shillings a day and instead sits idle for half of it. His point is that the idleness costs five shillings on top of whatever the man spent amusing himself. Franklin was costing out loafing. He was not pricing judgment, and he was reckoning in days.

The instinct runs older than Franklin. Benedict's Rule, written around 530, already assigns labour to numbered hours and appoints a brother to signal them, which puts the metering of work inside a monastery twelve centuries before it reaches a factory floor. Lewis Mumford saw where that led:

The clock, not the steam-engine, is the key-machine of the modern industrial age.

E.P. Thompson called the older way task orientation. The working day stretched and shrank around the job, around harvest and tide and how long a kiln takes. What replaced it was time itself as the thing being counted. Thompson wrote that "time is now currency: it is not passed but spent."

By about 1780 Josiah Wedgwood is writing instructions at Etruria about what to do with latecomers: take an account of the time they are deficient in, and stop that much of their wages.

In 1888 Willard Bundy patented a machine to do it without a human watching. Each worker carried a key cut to a unique shape, and turning it in the recorder printed his number and the time onto a paper tape. Bundy stated the purpose in the patent, which was to compel employees to record the time of their entering. His company became the International Time Recording Company, which folded into Computing-Tabulating-Recording, which renamed itself International Business Machines in 1924.

Taylor arrived in 1911 with a stopwatch and a corps of men he called time-study men, and made every motion in a factory measurable in seconds. He did not make time the unit of payment. He made work legible in units of time, which turned out to be enough. Once labour could be counted that way, the hour became a believable substitute for value, including in places where nobody was counting anything.

So when a client hears my number and says that's a lot, I have stopped hearing cheapness. They are running on nearly three centuries of training, and it was thorough.

Two ways to be hired

The complaint is older than the training, though, and older than the clock.

In Matthew 20 a landowner hires men for his vineyard at dawn and agrees with them on a denarius for the day. He goes out again at the third hour, the sixth, the ninth, and last at the eleventh, and at the end of the day he pays every man the same. The dawn crew objects. "These last have wrought but one hour," they say, "and thou hast made them equal unto us, which have borne the burden and heat of the day."

That is the division, performed out loud, in the first century. They took the payment, divided it by the hours, and objected to the quotient.

The parable is not about wages. It sits between two statements about the first being last and the last first, it opens by comparing itself to the kingdom of heaven, and its subject is the scandal of generosity. The landowner's answer is not an economic argument either. He points at the agreement, asks whether he has wronged anyone, and says he may do as he likes with what is his. What the parable preserves, accidentally, is the instinct. Hours worked feels like the natural measure of what a man is owed, and it felt that way centuries before anyone owned a clock.

One tradition took the two apart and never let them back together.

Jewish law separates a po'el, hired by time, from a kablan, hired to deliver a finished thing. Both are hiring. They are not the same transaction. Working through when a wage becomes payable, the Talmud distinguishes the man paid, in the Koren rendering, "based on the amount of times he performed an action" from the man paid "based on the outcome."

The consequences separate immediately. A po'el who walks off at midday is assessed on what he has already done, and paid for that. A kablan who walks is assessed on what remains to be done, and if finishing has grown expensive in the meantime, the increase comes out of his price. One man sold hours and settles on hours delivered. The other sold a finished thing and answers for the distance between what he handed over and the thing itself.

Rema, glossing the Shulchan Aruch, rules that a teacher hired for a set period is a po'el, while a teacher hired to teach a book is a kablan. Same teacher. Same knowledge in his head. The category is set entirely by how the engagement was written down.

Which is my whole argument, settled in the sixteenth century, four hundred years before I started losing it on the phone.

The professions had it done to them

So the distinction was available, written into law long before anyone needed it for a website.

The professions nearest my own lost it anyway, and they lost it recently. Factory discipline explains the wage worker, and I am not selling wage labour. What explains my problem happened later, and it happened to them more than they chose it.

American lawyers did not bill by the hour. Into the 1960s they ran on flat fees, annual retainers, contingency arrangements, and published minimum fee schedules that priced the task rather than the day. An uncontested adoption cost one amount and a contested one cost more, and the schedule said so in print. Architects had something better than that. They priced at a percentage of construction cost, which is a value-based fee in everything but name, because it scales with the size of the outcome instead of the duration of the effort.

Then it came apart inside three years. In 1972 the Justice Department sued the American Institute of Architects, and the consent decree that followed forced the AIA to stop treating competitive price quotation as an ethics violation. Parallel actions hit the civil engineers and the accountants the same year. In 1975 the Supreme Court decided Goldfarb v. Virginia State Bar and held that minimum fee schedules were price fixing under the Sherman Act, with no exemption for a learned profession. By 1978 hourly billing was the standard across most private legal work.

The drift toward hourly was already underway before any of those filings, pushed by institutional clients who wanted to see what they were buying. Antitrust did not invent the billable hour. It removed the published alternative, and the hour was standing right there.

Which means the thing I have argued against my entire career is roughly fifty years old. Nobody designed it. Nobody sat down and concluded that metered time was the correct instrument for pricing judgment. It filled a vacuum that opened when the professions lost the right to publish anything else in its place.

It is debris. It belongs in a rubbish pile.

Then the visible part got cheap

I have written before that the machine amplifies the craftsman instead of replacing him, and that the judgment is the part that does not transfer. I left out what it does to money.

AI never touched the invisible hours. It collapsed the visible ones. Production that ran forty hours now runs two. Deciding what the thing should actually do did not move. Neither did knowing which choices are expensive to reverse, or what breaks in year three.

So the ratio got worse. It was three to ten. Now the visible portion rounds toward nothing, and the meter reads near zero on the only quantity it was ever capable of reading. All that effort went into making labour countable, and it ends with the countable part worth almost nothing.

The number was never the problem

Which brings me back to the pause on the phone.

When someone hears my number and immediately reaches for how many hours that must be, they have told me something precise. They are auditing inputs. They never entered a conversation about outcomes and they were not going to, and I could have quoted half and heard the identical pause. The objection was never arithmetic.

There is an earlier signal than that, before any number gets spoken. It sits in how the work arrives.

A request phrased as a deliverable sounds like this: can you build me a website. The buyer ran their own diagnosis, reached a conclusion, and is now hiring hands to execute it. Everything that mattered got decided before the call, which leaves price as the only remaining variable. A request phrased as a problem sounds like this: our members stop renewing after the first year. That one arrives early enough for judgment to be worth paying for.

And I could do that with AI in a couple of hours is not an objection to be overcome. It is a person stating their own solution at their own price, out loud, before I have quoted anything. For years I treated those conversations as sales problems and argued with them. They were disqualifications, offered to me at no charge, and I kept declining to accept them. That is where the exhaustion came from. Not the losses. The arguing.

Who does not flinch

Two kinds of people have never once flinched at my number.

The first already bought the cheap version and watched it come apart: the site that went down during their busiest week, the room that sounds wrong and cannot be fixed without opening the ceiling. They are not estimating what the invisible hours are worth, because they have already paid for the absence of them and know the figure exactly.

The second I built myself, slowly. Clients who have watched three or four projects come in and stay standing, who stopped asking how long it takes somewhere around the second one. That conversion takes years, and it only ever succeeds with people who have a reason to listen. A stranger who opens with a price objection is not a candidate for it.

Franklin was warning a young tradesman about wasting an afternoon. We built that into a way to price expertise, then built machines to enforce it, then watched a court remove the only written alternative we had.

I no longer argue about the number. By the time the number arrives, the conversation was already decided by how the work got described in the first place.

They are right about the two hours. Nobody was ever buying the two hours.