Thoughts. What I notice while building.
Friday evening, an idea. Saturday, built. Sunday, live. By Tuesday I knew it did not work. Total cost: one weekend, fourteen euros for a domain and a bit of pride.
Ten years ago the same answer would have needed a proposal, six weeks of building and a five-figure sum. At that price nobody tests an idea. At that price you commit to it. And once you have committed, you can no longer afford the result.
That is the part of this moment I love, and it is not the part everyone names. Everyone talks about building having become cheap. True, and it is the duller half. The interesting half is this: being wrong has become cheap.
Before I claim anything, look at this loop. Scroll slowly.
The price of the mistake, not the price of the idea
Ideas were never scarce. Every company I have worked in had a drawer full of them. What was scarce was the answer. Does this work? And that answer had a price, and the price was not in the quote. It was in the cycle time: how long it takes for reality to talk back.
That is what collapsed. Not by twenty per cent. By an order of magnitude.
It does not make my ideas better. It changes how many of them I can push through reality in a year. Twenty-four instead of one. If you get one attempt, you have to be right. If you get twenty-four, you only have to be able to count.
Which is why the twenty-one crosses in the drawing above are not the problem. They are the product. I used to think good people are rarely wrong. That is nonsense. Good people are wrong cheaply.
One warning in the same place, or the rest of this reads too easy. What became cheap is the answer to "does it work", not the answer to "does it hold". A convincing demo now costs an afternoon, and out of that grew an entire market in which the buyer only ever sees the demo. Confuse the low cost of building with a low cost of running and you are building the next disappointment. The mistake is cheap. Operating is not.
Twelve months for a verdict
I am not talking employment down. I learned most of what I sell today inside it, and I had managers I owe something to.
I just never found out whether I had been right.
I found out whether I had been convincing. Those are entirely different things, and it took me years to see the difference. Between my idea and any verdict on my idea stood five people, three rounds and a budget cycle. By the time any of it reached a customer it was no longer my idea. It was the version that fit through every door.
Scroll again. Watch where most ideas die.
Not one of them fails at a customer. They fail at a wall whose opening had nothing to do with them. No villain in the picture, just an apparatus doing what every apparatus does: letting through what suits it.
And on the far right sat the annual review. A measuring instrument with a resolution of twelve months, measuring not whether the idea carried but how I came across in the room, netted against a pot that was fixed beforehand. I never fought the rating. I just noticed at some point that it said almost nothing about my work.
These days my annual review happens on Tuesdays. It takes a second, it says yes or no, and it costs the other person money. It is ruder, more precise and more honest than anything I had before.
Responsibility is the part that gets addictive
I expected responsibility to feel heavy. It feels light, and the reason is deeply unromantic.
When something does not work, there is nobody to blame. So I do not look for anyone. In every company I worked in, looking for blame was the most expensive activity in the building: weeks in which nobody builds because everyone is sorting out the story. Those weeks are simply gone now. What remains is the question of what I do differently on the next attempt. That is not pressure. That is relief.
The second part is less comfortable, which is why it matters more. I profit myself. Not in the sense that money is the point. In the sense that money is the only part of the feedback that is not polite. A client who pays has not agreed with me. They have paid for their agreement. Kind words cost them nothing. I get those for bad work too. A settled invoice I do not.
And yes, it runs both ways. A month without a signature is a statement about me, not about the market. That is the price of the good months also being mine. I have not regretted the trade for a second.
Choosing your clients is the real luxury
When people ask what the biggest difference is, they expect an answer about time or money. It is neither.
It is that I get to choose who I work with.
The wrong client costs more than no client. They cost the month, the head space, and the three right clients I did not find that month because I was busy with them. I only believed that after paying for it twice.
I turn work down now. The first time it felt like arrogance. Today it is the sharpest tool I have, because every no is a decision about what the next six months look like.
The honest footnote: this only starts working above a certain level of utilisation. If the rent is not covered, you do not choose, you take. Which is why the first goal after starting is not revenue. The first goal is the ability to say no. Everything after that gets better because of it, revenue included.
And then the noise
At the start I believed I had to post. So I posted. Regularly, neatly, with the hooks and the line breaks and the small punchline at the end.
The reach was there. What came out of it was nothing.
The reason is the same as with building, one level up: producing content now costs almost nothing. And what costs nothing carries no signal. When anyone can write the same clever sentence in the same formatting, the sentence has stopped being evidence of anything. It is decoration with impressions attached.
Scroll. On the left something is broadcast, on the right something is passed on.
A recommendation is the only signal I cannot manufacture myself. Someone puts their own reputation next to my name, in a room I am not in. That is expensive to fake and therefore cheap to believe. Which is exactly why one sentence in somebody else's meeting beats twelve months of feed maintenance.
Then there is the compounding. A post is back at zero on Monday. A recommendation is still standing in two years and has had offspring in the meantime. The drawing above is unfair at precisely this point, and deliberately so: at the moment the fan looks impressive, the tree has two nodes. Anyone comparing after four weeks picks the fan every time.
I am not claiming visibility is worthless. Reach without referral is just an amplifier with no signal going into it. And I write long pieces here, which is itself a form of visibility. The difference is testable, and the test question is: does it cost me anything? A text that takes three days and risks a thesis that could be wrong is work. A post that breaks a truism across seven lines is attendance.
The fastest route to the second referral, by the way, is not a marketing channel. It is the first job, finished properly, including the awkward last part nobody pays for any more.
When I advise against it
Otherwise this reads like a brochure, and the point matters too much for that.
I am writing this from inside a good phase. That is survivorship bias and I cannot argue my way out of it. Ask me again in a bad quarter.
If your fixed costs cannot survive three quiet months, do not start yet, build the buffer first. If you start because you hate your job, you are building against something, and that lasts about six months. If you cannot bear the answer that your idea was bad, you will not hear it and you will keep building anyway, and that is the most expensive version of all.
And the strongest objection to my own text: fun is not a business model. The fun comes from the loop, not from being self-employed. You can build that loop inside a job too, if you find a place where ideas get tested on Tuesdays instead of in the budget cycle. Those places exist, they are rare, and if you find one it is probably the better deal.
Why now
The three drawings are the same sentence at three resolutions. The loop says how many answers fit into a year. The chain says who used to vote on them first. The fan says what is left of a pile of attention when nobody passes it on.
Building something has not become safer. It has only become cheaper to find out.
Being wrong costs almost nothing now. The only expensive thing left is not asking.
Related
- The Disappointment Was Priced In: why cheap building breaks a market in which only the demo gets judged
- Fear Is Not a Probability: why the risk you feel is rarely the risk you carry
- A Tool, Not a Miracle: what the boom actually costs and when less is smarter
- Over Budget, Behind Schedule, Off Target: the same chain, seen from inside the project
