Long before I dance under those lights
Muhammad Ali said the fight is won far away from witnesses. So is almost everything else, in the part of a business no customer will ever see.
The fight is won or lost far away from witnesses — behind the lines, in the gym, and out there on the road, long before I dance under those lights. Muhammad Ali
It claims more than it first appears to. He is not saying that preparation helps. He is saying the result already exists before the fight starts. What happens under the lights is the reveal, not the decision.
The lights are the visible part. The launch, the pitch, the day somebody finally looks at you. The road at five in the morning is the part nobody films, and that is where it was settled.
By the time anyone is watching, the outcome is mostly already in.
On the nature of building something
My five year old is screaming.
He is building a Lego set rated seven and up. That was a decision we made together and one I am currently regretting.
I still have not fixed it, so I cannot tell you what went wrong. What I can tell you is that something is not lining up.
Finding even that much took a while. I had to go back through the instruction booklet, page after page, about seven pages, before I could see that the thing in front of me and the thing on the page had stopped agreeing with each other. Somewhere back there a piece is missing or in the wrong place. That is my suspicion. I will find out when I take it apart.
The part worth noticing is the gap. Whatever he did wrong, he did it a long time ago, and it fitted fine at the time. It has been sitting under everything he has built since, quietly passing the error upward, and it only announced itself now.
Anyone who has built Lego with a child knows this experience. It is not an unusual failure. It is the normal one.
That is the whole of this essay, and I could stop here.
Two companies that have been wildly successful because of the body of work customers never see
Amazon is not dominant because of its website. The website is competent. Plenty of retailers have a competent website and none of them are Amazon.
Amazon is dominant because of fulfilment centres, inventory placement, routing, and a delivery network built over decades at enormous cost. A customer experiences exactly one output of all of that. The thing they wanted, at the price they wanted, arriving when they expected. Everything that produced it is invisible to them.
Walmart is the same story a generation earlier. What made Walmart was not the stores. It was distribution centres, a private trucking fleet, and a grip on logistics that let them put the same product on a shelf for less than anyone else could manage. Shoppers experienced a low price. They never saw the machine that made it possible.
In both cases the customer judged the business on the last six inches. The business won it several hundred miles upstream.
What I have been doing with my evenings
This is on my mind, in fact I have become slightly obsessed with it in my own work.
I am building a platform. The fast version of that job is available to me. I could have something out cheaply and quickly, and there are days when that is extremely tempting.
Instead I have spent hours going back and forth with Claude about architecture. Not about features. About how the thing is put together underneath. Where the boundaries sit. What the data model has to allow for. Which decisions are cheap now and expensive to reverse in a year.
None of that is visible in the product. Nobody will ever open the site and think, ah, sensible boundaries. It does not demo. It cannot be screenshotted.
What it buys is the right to build the robust version later without taking the whole thing apart first. It is provisioning for a future I cannot see yet, which is a strange way to spend a Tuesday evening.
I am aware this is exactly the behaviour that can become an excuse not to ship. I do not have a clean test for telling the two apart. I only know which mistake is cheaper to discover late.
What this means when you are not Amazon
The tempting conclusion is that this is a story about scale, and that a business with one person has nothing to learn from a fulfilment network.
It is the opposite.
A service business has its own version of the warehouse, and it is usually small enough to build in an afternoon. The enquiry answered within the hour instead of the next day. The quote that goes out while the customer still remembers calling. The booking that confirms itself at ten at night. The review that gets asked for every single time rather than when somebody remembers.
None of that is visible to a customer. All a customer notices is that dealing with you was easy. Easy is the output. The mechanism is what you built.
And like a warehouse, none of it feels like progress on the day you do it. You will not feel impressive setting up a reminder that fires three days after a quote goes out. You will feel like you are fiddling with settings while the real work waits.
Then a year later the quotes that used to go cold are closing, and you will not be able to point at the day it changed. There was no day. That is the point.
Back to the floor
The set is still in pieces on my living room floor.
I know roughly where to start looking, and I know what the job is. Take apart everything built since, including a lot of perfectly good work that was never the problem, until I reach whatever is missing or wrong. There is no adjustment at the top that fixes it, because the top is not where it is.
He took this badly, and he is right to. It is genuinely unfair. The mistake was invisible for the entire time it was doing its damage, and the price for it is falling due now, on work that was fine.
Businesses get the same deal and usually like it even less. The symptom shows up at the top, in the quiet month or the quote that went cold. The cause is seven pages back, in something that seemed to fit at the time.
A free assessment, five minutes