I'd look at it from a different perspective. A low-level SWE salary might be $120k, and the all-in cost of employment might be $160k (benefits, taxes, office space, lunches, etc.). For a senior person, the base salary might be $200k, and the all-in cost might be $260k. So how does the senior person justify that extra $100k?
Well, they just have to output about 2x more impact. In my experience, it is not very hard to be 2x more impactful than a junior engineer. This is not meant to be arrogant, just that it's very easy for a junior engineer to waste time on useless stuff. In some cases junior (or bad) engineers will be 10x or 50x slower than they should be. I've seen some really bad examples in my 10 year career.
For example at my last job, a team of 2 engineers was using the wrong tool for a task, and this resulted in them taking 5x longer to deliver a worse product than alternates. I was able to do the same thing as them, but deliver much faster and at lower cost.
In other cases, junior engineers will make poor decisions, or write bad code that needs to be refactored later. This can be simple things like thinking on the wrong layer of abstraction. Refactoring the bad code is effectively 10x slower than doing it right the first time.
In all, I think paying an extra $100k for a good engineer is a real bargain for companies, because the impact can be so huge. The real challenge is figuring out who is a legit senior engineer.
Well, they just have to output about 2x more impact. In my experience, it is not very hard to be 2x more impactful than a junior engineer. This is not meant to be arrogant, just that it's very easy for a junior engineer to waste time on useless stuff. In some cases junior (or bad) engineers will be 10x or 50x slower than they should be. I've seen some really bad examples in my 10 year career.
For example at my last job, a team of 2 engineers was using the wrong tool for a task, and this resulted in them taking 5x longer to deliver a worse product than alternates. I was able to do the same thing as them, but deliver much faster and at lower cost.
In other cases, junior engineers will make poor decisions, or write bad code that needs to be refactored later. This can be simple things like thinking on the wrong layer of abstraction. Refactoring the bad code is effectively 10x slower than doing it right the first time.
In all, I think paying an extra $100k for a good engineer is a real bargain for companies, because the impact can be so huge. The real challenge is figuring out who is a legit senior engineer.