Continuing the discussion from The Comparative Advantage of Knowledge Workers:
I suspect this is a troll, but I can’t help myself.
While statistically this might be true at some level, I feel like it’s trivially easy to weed out software businesses that are far more likely to succeed. There’s a big difference between “Facebook for cats” and “boring, relatively simple SaaS that replaces spreadsheets at 3 companies I’ve worked at, and another 10 I’ve spoken to.” A stat like 99% of all small software companies fail surely is heavily weighted with stupid ideas. (Necessarily so: you can fail at a business in a weekend, but succeeding will take years.)