[eldavido]

You're correct that the marginal value of great code vs. good code is...marginal, MOST of the time. But I think the 1% cases are where the big (200k+) comp packages are for ICs. Stuff like perf optimizations at big scale, 1% adwords CTR improvements, optimizing JS runtimes, etc.

I agree that for most organizations management is the highest-leverage (ergo, highest-paid) path.

A better way to frame the problem is to think about the organization's economics and how your contribution fits into it. Figure out the 2-3 key metrics for success at the org and figure out whether you can move those, and if so, how. And realize that in a lot of job roles, you won't move the key metrics, and won't be paid as such.