Employees Who Stay in Companies Longer Than Two Years Get Paid 50% Less

As I went from "no" experience to my current position my job switches (every ~2 years) always were for 10%-25% (in one case, 100%, but that was Midwest to Bay Area so other factors are at play). I never got more than 5% merit increases otherwise.

Most of my friends who have been at the same place for >10 years in engineering (not software) are just now reaching parity with my base salary (CoL-adjusted) and they didn't spend 6 years in academia before transitioning to industry.

Two year tenures isn't job hopping. It's a reflection of how this industry works. Very few companies offer sufficient breadth and depth of product complexity, career advancement, or other similar things to make it worthwhile. I'd say the sweet spot is 2-4 years, except at very large companies (e.g. Google) (EDIT:) or companies which are developing complicated products with physical engineering or regulatory factors complicating development. Anything longer, especially if there is a lot of long-term stasis on a resume (e.g. "tech lead on product X" for more than a year or two), is an indication to me of someone who either isn't capable of stretching himself or doesn't want to. Anything less, especially if more than one project per job exists, indicates an inability to see a project through to maintenance or someone who is easily bored.

I work at a wholesale insurance company (middleman); we have over a hundred individual systems, some simple, some quite complex. Some of those systems can be maintained by any programmer that comes in. Others really shouldn't be maintained by someone who doesn't have good experience in writing maintainable code.

But many projects require domain-specific knowledge---and not just to specific parts of and companies within the insurance industry, but _our specific company_.

For a few years, we hired less experienced candidates with the hope of training them, because the talent in Buffalo NY is hard to come by (which could also be an outreach problem, but our last three hires were from out of state). Our team's consensus is that it takes approximately two years of experience with our systems to have a good, balanced grasp on them. Because of the small team size (five at our highest, two right now), we place great value on team members who have been around long enough to know the systems, the rationale and design decisions, history, etc.

This relates directly to Peter Naur's concept of Theory Building.[0]

We've had the past four candidates leave in ~2, Because we have the tendency to lose employees at around two years, I'm no longer in favor of hiring more "junior" candidates---having to train in both domain knowledge _and_ general programming is too risky and hasn't panned out well. So hopefully we can hire programmers that can immediately get into the code from a general programming perspective, but it's going to take them quite some time to really understand what the systems are all about.

I've been there for over eight years, and the last programmer that remains with me has been there for six months longer than me. One of our programmers transitioned to technical director, and has been there for over ten years. My supervisor dropped programming for project management responsibilities, and he has been at the company for nearly ten. Our sysadmin has been there for over ten. We just lost an employee that was there over ten years. We're a close group. Would I be making more money job hopping? Probably, yes. My point here is that the company I work for---the team I work with---values experience with the company, and so long-term employment history is attractive.

[0]: http://pages.cs.wisc.edu/~remzi/Naur.pdf

That's insane. Someone who is the technical leader of a project for more than two years doesn't "stretch himself"? I see no reason to believe that's true. If anything, it takes years for you to realize design flaws in components of the earliest-developed systems. The extent to which you can learn about how to design well-designed systems is going to be quite limited if just peace out after a year or two, as you suggest.

But the author has conflated this basic principle with changing companies, which I believe is a false statement in general. The effectiveness of changing company as a strategy to increase pay will vary wildly across industries and professions.