Nobody should pay for deep work up front.
Carsten Christiani
If you have ever paid for an assessment of your company and got back a well-made description of things you already knew, the people who wrote it were probably not the problem. You paid for a read of the company before anyone had touched it, and a company shows most of how it really works only when somebody tries to change something in it.
It's a bit like taking your car to a mechanic and describing the noise. A good mechanic learns something from your description, and a lot more once the wheel is off and they've started on the repair. Interviews and workshops are the description of the noise. They give you the org chart and the story people tell about why things are the way they are, all of it honestly meant, and most of it a step away from where the problem lives.
The real arrangement shows itself in the work. To fix something concrete, say a report nobody trusts or a handoff that keeps dropping things, you have to find out who actually decides, and where the knowledge sits in practice as opposed to on paper. You can't fix it without learning those things, and you don't need to interview anybody to learn them. You keep running into them.
Take a website rebuild, which sounds like the least deep piece of work a company can buy. To move a site properly you have to agree what goes on the front page, which means finding out who is allowed to decide that. And you have to find every page and every claim and who stands behind it, which means finding out where the company's knowledge actually lives. I'm not that interested in websites for their own sake. But in that context it's the next logical Lego brick, and by the time the site is done you know a lot about how the place decides.
There's a second reason, which has to do with how deep work behaves. Change at the level of how a company decides runs in ripples. It shows up months later, in places that don't obviously connect to where you started, and it's very hard to credit to anything in particular inside a quarter. So asking someone to fund it before anything has worked means asking them to pay for something they can't see yet, to people they don't trust yet.
I should also admit to a habit. I look at a broken handoff and a decision made two years ago, and the connection between them seems completely obvious to me. Isn't it obvious this fits together? No, it turns out, mostly it isn't, and there's no reason anyone should take my word for it before they've seen one piece of it work.
So the order runs the other way. Start with the problem that is costing the most right now, at a fixed price, small enough that saying yes is easier than thinking about it, and worth having even if you stop there. Read the system while you fix it. Then decide, with something real in your hands, whether the deeper work is worth doing at all. The ambition can be large, and the first commitment still doesn't need to be.
That's how we price our own work, and it's less noble than it sounds. It's simply the only way we know to see a company properly. We get our hands dirty fixing the thing, which is how we come to understand the system, and after that there is something real to talk about. Each step earns the next one, and as trust builds, the conversation can go to places that would have set off everyone's weird-o-meter on the first day.
You can use the same order inside your own company. Next time someone proposes a review or a transformation programme, or you're about to ask for a report on why something keeps going wrong, take the most annoying concrete problem inside its scope and have the person who will live with the fix sort that out first. Give them one instruction: write down everything you had to work around, and everyone you had to ask, to get it done. Read that list before you commission anything bigger. In my experience it tells you more about where the real work is than the review would have, and it costs you one fixed problem, which you wanted fixed anyway.
The deep work will keep. It has been keeping for years.