It really depends. You really need to know what you're doing through and through. What to cloak, what to buffer, what doesn't matter, what does, and so on. All though theres definitely a little luck, in my years I've seen sites that focused on all "white hat" techniques get dropped from the SEs and replaced by sites with spam. And vice versa. And the trends change. This year spam might be working for a particular niche and then in 1-4 years that changes and you're being out ranked by sites with topical siloing, YMYL content, established authorship, use of entities and blah blah blah. So when you're talking about the "long run" with Google, who the hell knows?
So it's really looking at the way Google is reacting to your niche by analyzing your competition atm.
AND
Having testing sites where you test links, on-site optimizations, technical optimizations etc etc where you can see how Google reacts to any changes is 100% necessary. I NEVER TEST on sites I care about. Never. You do one change at a time to multiple test sites and see if theres positive movement.
If the optimization moved all sites in a positive direction, well it's conclusive. I honestly only do one test site nowadays. But for those that are really into experimentation and spending a ton of time you don't need, have at it![]()
The same effort can be better spent in building a quality site which stays there for a long time.