It really depends on how intelligently companies run their retention metrics. Rather than measuring if a customer was retained for that phone call, it makes much more sense to measure which customers are retained after a month or so. Of course the first metric is easier to run; also large corporations "competing" in highly uncompetitive markets (internet, health insurance, etc.) have little incentive to provide any sort of reasonable, adequate customer service to begin with.

As a customer I don't care. I just want what I've called for. Comcast is the single worst customer service I've ever dealt with.

Even if that's true, Comcast's bad metrics aren't our responsibility.

The point is that this tactic would be selecting against the nice reps in the long run. Of course, I can't think of any other tactic that wouldn't do so.

Understood, but my point is that it is their responsibility to select for nice reps, not ours. If their metrics combined with the natural customer reaction selects for bad reps, that's their fault and their problem.