The business equivalent of premature optimization is premature hiring or premature process.
Early-stage operators often build elaborate workflows, hire specialists, and create approval chains for problems that don't exist at their current scale. You're adding complexity to handle load you're not carrying yet.
Same principle applies: measure first, then optimize. Don't build a 10-person ops process when you're a team of 2.
The corollary to Knuth's rule I try to remember: "make it work, make it right, make it fast — in that order." Most teams skip straight to the third step.