A problem you could have
Paul Graham says that good startups solve a problem, ideally a problem the founders themselves had. That framework resonates with me, but I think there’s a notable middle ground here — “a problem that you could have”.
In particular, startups that sell to other startups are solving problems that the founders could have the next time they start a startup. Not only that — the problems they solve are problems that the engineers and product managers they hire could have on their next project or company. This makes it faster and easier to build the right product.
The same is often true for consumer products. The teams working on Amazon e-commerce, Facebook, Twitter, Gmail, etc are either solving problems they have or could have.
The situation is different at an enterprise company like Samsara or Flexport, where the customers fit a completely different user profile than the founder, engineers, or product managers. At these companies, a more rigorous research and domain modeling effort is needed to understand customer needs and build the right thing. This makes product development slower and harder. On the other hand, getting this right means it is more difficult to replicate.