[d0mine]

The general thesis: "to optimize, do less" is correct (even tautological in the time performance area) but specific examples without additional context are troubling: I see an enthusiastic manager reading your post as an endorsement of dropping tests, code reviews, git, any kind of automation, etc (as it doesn't bring direct business value). Obviously, it is shortsighted. You'll pay dearly for any time you initially save by dropping these practices in most contexts.

[Breza]

I run the data science department in a corporation. The best thing I've done is banishing Excel from our reporting workflow and replacing it with various types of automated reporting. Stuff like that is what you should cut, not something that'll end up adding technical debt.