I think that’s an interesting question. When I try to allocate my own time on a project it’s generally 80% / 20% / 0% — but that’s assuming there’s work that immediately needs to get done. I think that unless you have an almost impossibly good pre-existing architecture, investing time in refactoring almost always pays dividends in the future (which is why the 20% on “tomorrow”). I guess if I had literally no work to do it would probably look something like 0% / 80% / 20% ?