The unit economics were broken from the start, I just did not look at them properly. Or I looked and then looked away. Each new customer I was acquiring cost more to serve than they paid, and I had done the math in a way that assumed volume would fix that. Volume does not fix a per unit problem, it makes it bigger. I knew this in theory. I had read the blog posts. It is different when it is your own numbers on your own spreadsheet at eleven pm in a city where you moved to prove something to yourself. I think I was also scared that if I slowed down to actually fix the model I would have to admit it was not working yet, and I had told too many people it was working. Friends from back home, my parents, a few investors I had coffees with. Small lies that compound.