I've found that organizational structure and politics become a limiting factor past a certain level of skill and experience.
That is, for one engineer to be obviously worth the price of three or four others, they'll need more responsibility and resources somehow.
In finance, the responsibility is easy to delegate: it's the size (actual money) of decisions they'll be making. In generic software, how do you do the same thing? All the code on X servers? Primary responsibility and merge privileges on Y repos? There are interesting experiments and answers in this space, but most shops aren't oriented to think that way.
That is actually an excellent point. As a senior developer you need to be able to leverage your experience. That is, being able to make decisions, being able to teach and mentor, etc.
Answering from my position. I didn't get responsibility for more systems, or bigger projects, or things like that. They made me responsible for the technical side of production. All of it, zero wiggle room.
Overall, I would not consider myself a manager. I don't have disciplinary rights, even in my team. I guess if I want someone gone, they will leave the company, but I'm no disciplinary superior.
And in my space, technical considerations still dominate. My decisions just cost more or less, affect availability, maintainability, support of our software, alow teams to go faster.
I code less chef and terraform, yeah. I'd like to do so more. But fuck me, it's better for my team to tell people to do as my team wants. It's better for me to keep my head above single resources and servers, so we keep a tactical view.
But lots and lots of senior engineers aren’t high performing. They’re still moderately more valuable than junior engineers, but not immensely so.
People calling themselves “senior engineer” are usually referring to years of experience, not to whether their management chain consisently sees them as high performing.
What I mean is that it is hard to give a number to a person. Sometimes they shine, sometimes they don't.
I, personally, rarely can do really-work more than 4 hours a day -- only if task is very interesting and challenging. Then I can do it for 12 hours a day straight, but I stop myself if I work more than 6 hours, because I know I would burn out, happened a few times.
4 hours a day of really-work sounds not a lot, but my performance is pretty good comparing to my teammates. Never heard managers complain on my performance. The other 4 hours I spend on meetings/lunch/hacker news/socializing etc. So I would look for a candidate who can do it for at least 4 hours a day.
I heard rumors of people who can really-work for 8 hours a day (like John Carmack). Those people are called "100x engineers", I really want to work with some of them one day.
But I can produce just as much as some other developers, so I really see no incentive to do more work unless it's clear I will get more recognition for it, rather than just more work to do.
I have 15 years of experience (both huge companies and own startups), but I keep wondering what if it's really the case: that there's simply no difference for company success? When I hire people myself, I only see value up until mid-seniority. From there what matters is just discipline, readiness to help peers, responsibility and hours candidate can really-work.