[xenihn]

Sure, but is it cheaper in the long run when you account for attrition, hiring costs (recruiters, engineer hours spent interviewing), and ramp-up? Let's go ahead and assume that things like code quality and technical debt aren't factors here, and that the replacements for regrettable attrition would receive (and accept) offers with comparable TC.

[timwaagh]

That depends on project duration. Most projects did not go on more than a year or two. If that's the case your ramp up would happen once every two years even if everybody stayed forever. If you're reasonably efficient about it, you can hire someone with maybe a few hours of engineer time. If the average engineer stays even a year it's not that much.

[ntsplnkv2]

You don't bother thinking about that. Over time you have a large enough sample size for attrition rate and you hire based off that. it doesn't matter if a few leave, a few will always leave.

[xenihn]

Why not think about it?

[ntsplnkv2]

It costs time and value when you really can't control it.

No matter what there will be attrition for reasons completely outside of a company's control. You find a comfortable attrition rate for your business. if it's too high you deal with it.