[001sky]

"The top developers are 25 times more effective than the average developer, but their salaries don’t reflect that"

For consideration. Two counter examples--(1) A great engineer working on a shitty project is not worth much. (2) And a great project can be executed with competent engineers. People are only as "effective" as the problems they solve are potenially valuable.

Now imagine another variation:

"The top [founders] are 25 times more effective than the average developer, but their salaries don’t reflect that"

Is this more or less likely to be true? And does it have anything to do with engineering and or Job titles? or is this phenomena something else altogether?

[dllthomas]

I'm not quite catching your point. If a coder is able to do 25x as much in the same time, I am willing to pay him substantially more - I get things done sooner and/or need fewer other coders. Most times, I won't be willing to pay him all of the 25x more (even with perfect information), because there is more risk putting all my eggs in the one basket. In some circumstances, where time is especially critical, I may be willing to pay more than the 25x - Mythical Man-Month and all that.

That they might be more valuable on other projects may be relevant to what I need to pay to retain them, but isn't relevant to what I am willing to pay.

[001sky]

Just trying to think through the issues at different levels of abstraction. Not sure i have the answers. But just some thoughts.

Consider a situtuation where you don't know the productivity of X and its subject to 25x variance, a project manager is unlikely to structure a project to allow for 25x variance as acceptable. He's eaither at risk of sandbagging or falling massively short. And so upsetting other parts of the business. So, he needs to have more like a 50% [edit: I like chuck's 3x also] plus or minus understanding of what he can deliver. So, its really hard to pay 25x when you are going to constrain his performance to 1.5x. Likewise, its hard to restructure your projects for 25x as the expected case, unless you have really good data. But, most of the data will be 1.5x ceiling (if he is maxing performance under the previous example). So, its still a leap of faith to restructure projects around 25x expected performance from 1.5x data (order of magnitude jump).

But, for a founder this is not so constrained. There it is a lot easlier to structure your work at your peak capacity. With equity, you will benefit from the performance, etc. But this is true wether or not you are an engineer or whatever else expertise. So, one theory might be that the top employees are the ones that allow the founders to do precisely that -- franchise players that allow strategy to be developed around a 10x or 25x performance envelope--but for the market to be efficient, two things need to happen. There needs to be reliable data of 25x performance, and there need to be jobs that can be structured around this expectation. Without upsetting the apple cart of exectution with the rest of the biz guys (marketing/bd/sales/ etc).

And bridging those gaps in the data is the hard part about talent search. And why people will pay a bit for it.

edits.

[dllthomas]

Yeah, asymmetric and unreliable information plays a huge roll, for sure. That it plays out differently for founders vs. employees is an interesting elaboration.