If you're going down the "Services: The New Software" rabbit hole, I hope you like systems. And I hope you like writing things down. My last business was a consultancy. It was not fun. I envied SaaS businesses for their ability to scale their revenue. I was constrained by headcount. I could pull some levers to improve my margin, but if I wanted to grow revenue I needed more people. Bootstrapping didn't help either. I had to carefully balance sales with hiring. A bad quarter and you lose your profits for the year. Two? You're in the red. And yet, our profits were 2 to 3 times that of most of our competitors. Three reasons I think we were able to pull that off: 1. We simply had much higher standards. I know... But you'd be surprised how complacent our competition was. 2. We managed our clients very carefully. We didn't simply do what they asked. We figured out what they were trying to achieve, and delivered against that. That often meant saying no, pushing back, or refusing to take money from them. 3. We were autistic about our playbook. We spent a lot of time writing things down. SOPs, templates, handbooks... It is very hard to create, maintain, and curate knowledge. But without it, you're building on quicksand [1]. I'm sure we did other things right, including creating our own accelerator/product, but these three things were the most important ones. And while I had sworn I would never go back to services, a lot of what I hated about it looks different with AI: - You can create repeatable processes even when the input is very fuzzy. - You're not as constrained by your headcount. - Creating and maintaining internal products is way easier. The speed makes it more viable too. Looking at 8090 or Distyl, I'm thinking this might just work... Maybe I'll try again [2]. [1]: There's a lot more to it, enough that I'm writing a book about it: [2]: I kind of already am: