There's a bloke in every business owner's phone that you should have stopped answering two years ago.
You know the one. The client who rings at 9pm about something that isn't urgent. The one who ignores the onboarding form, then complains the onboarding was unclear. The one who asks for a discount, gets it, and still leaves a three star review because the invoice arrived on a Tuesday.
But you keep him because you think you can manage him. You've told yourself a little story about how he'll come good once the systems bed in.
He won't. And there's a dead Italian economist who can explain why, with a graph.
The four types, on one bit of paper
In 1976 an economic historian called Carlo Cipolla wrote a short essay called The Basic Laws of Human Stupidity. It reads like a joke for about four pages, sort of a wind up, and then you realise he's completely serious and you've been doing business wrong since you started.
So his model is two axes. Along the bottom, what an action gains for the person doing it. Up the side, what that same action gains for everybody else. Four boxes come out of it.
Intelligent. They win, you win. Boring, lovely, rare.
Helpless. They lose, you win. The supplier who underquoted and honours it anyway. You feel a bit guilty and you should.
Bandit. They win, you lose. The one who haggles you to the bone and knows exactly what he's doing.
Stupid. They lose, and so do you. Nobody gains anything. Value just evaporates.
Cipolla's argument, and this is the bit that ruined my week, is that the fourth type is more dangerous than the bandit. Everyone assumes the villain is the worst customer you can have. He isn't. The villain has a plan. You can read a plan. Write it into the contract, take the deposit up front, price the risk in, and get on with your life.
The other one has no plan. He costs you a Thursday and gains nothing from it, which means you can't predict him, can't negotiate with him, and can't build a defence, because there's nothing rational there to defend against.
Here's the controversial part
Your worst client is not the one who beat you down on price.
Your worst client is the nice one. The pleasant, apologetic, enthusiastic one who agrees with everything on the call and then does none of it. He's not trying to take anything off you. That's exactly the problem. There's no motive to work with, no lever to pull, nothing to appeal to. He burns your hours, burns his own money, and at the end of it neither of you has a thing to show.
And you'll defend him to your mates, because he's lovely.
The haggler, meanwhile, pays on time and refers two people. You resent him for the discount and he's the better customer by a distance.
We've got this completely upside down. The whole industry teaches you to spot the "red flag client" by attitude. Rude, pushy, demanding. Wrong signal. Attitude tells you nothing. Cipolla's second law says the trait is spread evenly across every group you can name, so no amount of vibe reading on a discovery call will save you. Polished website, good manners, big follower count, none of it correlates.
What correlates is behaviour. Not what they said on the call, what they did in the fortnight after it.
What this changes on Monday morning
So I've stopped scoring people and started scoring their last three actions. It takes about ninety seconds and it's the cheapest bit of due diligence going.
And grab a bit of paper. Draw the two axes. For the last three things this person actually did, put a dot in a box. Not what they promised. What they did.
Three dots in the top right, they're a keeper, give them more.
Dots in the bandit box, fine, keep working with them, just get it in writing. Deposit up front, scope locked, payment terms that bite. Self interest is readable, and readable is workable.
Dots scattered in the bottom left with no pattern, that's your answer. Not a difficult client. Not a communication issue. Not something a better onboarding sequence fixes. Leave.
But the trap here is the one Cipolla calls the fourth law, and it's the expensive one. Smart people know the type exists and still think they're the exception who can handle it. They can't. You can't outmanoeuvre someone whose next move has no logic behind it, because you're planning and he's just doing things.
I've paid that tuition. I ran two projects last year that were going to be great once the other side got their bit sorted. They never got their bit sorted. Nobody was a villain. Nothing was gained by anyone. Just months of my life converted directly into nothing, which is Cipolla's definition almost word for word.
The bit nobody wants to hear about their own business
The same graph works pointed inward, and it's uglier.
And every business has processes that sit in that bottom left box. The report you generate weekly that nobody opens. The approval step that exists because someone got burned once, years back. The follow up email sequence that annoys people into unsubscribing. Nobody gains. Not you, not the customer. Pure value destruction with a calendar reminder attached.
I made my own AI do this to itself. My assistant now runs one test before it acts on anything: does this leave me better off and the other person better off? If it can't answer yes to both, it doesn't do it, or it asks first. That single rule has killed more pointless output than any amount of clever prompting, because "be helpful" is a much lower bar than "be net positive for everyone in the transaction".
Look, try it on your own week. Go through your recurring tasks and ask who actually gains from each one. Be honest about the answer. Most people find two or three things they've been doing for years that gain nobody anything, and the reason they survived is that nobody ever asked the question out loud.
Do the maths on the nice one
So take that client you're defending. Hours you spent last quarter, times what an hour of yours is worth. Then add the work you didn't take on because he was in the diary.
So now put his fee next to it.
If the number is negative for you and he isn't getting results either, you're both in the bottom left box together, holding hands, going nowhere. That's not a relationship worth saving with a better system. Some things don't need automating. Some things need ending.
Cipolla reckoned the difference between a business going up and one going down isn't how many of these people are about. It's the same proportion everywhere, always has been. The difference is how much room they get given.
Give them less room.
Sources
- Carlo M. Cipolla, The Basic Laws of Human Stupidity, first printed privately in 1976, published in English by Doubleday in 2019. Publisher page
---
Want to know which of your processes are in that bottom left box? That's exactly what the AI Ops Audit does. I go through your operation, find the work that gains nobody anything, and show you what to cut, what to automate, and what to leave alone. You get a plain English report, not a sales deck.
Brewed by Steven, poured by Viktor
About Steven Tann: Steven helps business owners build systems that run themselves using AI. After 10+ years helping 7,000+ businesses and building his own autonomous operations, he's the bloke who actually does it, not just talks about it. Find out more at steventann.com.