3 ms·
> Well, as a thought experiment; should the CEO of the company know how to do every function he or she hires for within their business? No, but they should at
by nine_zeros 3y ago
> Well, as a thought experiment; should the CEO of the company know how to do every function he or she hires for within their business?
No, but they should at least know how to do the job of their SVPs or VPs in their absence. In fact, many times when the SVP departs, their reports are expected to report to the CEO directly.
- bee_rider 3y agoBut if the job of the SVP and VP includes being able to do the job of the people that report to them then doesn’t this require the CEO to ultimately know how to do everything?
- nine_zeros 3y agoThere is a depth first and a breadth first traversal aspect to it. Which is why a skilled CEO, presents themselves as a backstop, but simultaneously starts the hunt for a new SVP or an interim SVP. But even appointing someone new for that role will require the CEO to understand that role very well. They truly can't appoint someone competent in that role without being competent themselves.
- bee_rider 3y agoA DFS and a BFS both visit every (connected) node, haha. I don’t think the original comment was saying that it is crucial for somebody to fill in for their reports (that might be true, but it is a different aspect). Rather, they were saying that the manager’s skill set must include the skill-set of the person being managed, in order to evaluate the latter. I think there’s some truth to this, but there’s also obviously got to be some attenuation factor going on, otherwise we get this “visit every node” problem. The suggestion of firing every manager that can’t run the test suite seems like it ignores this attenuation factor. It might be a set of command line scripts with funky in-house flags. But they should be able to draw some some of useful conclusions from the output IMO.
- svnt 3y agoRole competency and the ability to assess role competency are two different skills.