I tend to approach engineering management decisions as I do software development: small adjustments most of the time, big refactors only when the potential long term rewards are worth the risks and there’s no other viable solution.
Like in software engineering, the context in which we make decisions matters. The strategies that make sense in a small, 3 dev team don’t really scale for a larger group. The decisions we make to create an initial prototype to prove an idea are different from what we’d do for production applications, but it’s important to remember that doesn’t make them invalid, it makes them good decisions for the context in which they were made. The same goes for teams. The way you manage a 3-person team is not the same as with a 20-person team, and that’s ok.