If you have a candidate with a good CV and he's given you his stackoverflow account name, then it should take you about 5 minutes to get a good idea of how much time he spends on the site, what kind of questions he asks/answers and how well those questions are asked and answered. What you want to do with information is course totally up to you, but I fail to see why having that information is in any way a bad thing.
I agree that a raw number means very little, but the info gleaned from looking up their stackoverflow account can tell you quite a bit.
Stackoverflow Careers is based on the premise that it's worth $5,000/yr to me, plus $99/yr/candidate for the candidate, to share that information.
Do you really see that much value, versus simply asking all candidates 'Are there any publicly accessible technical forums or mailing lists that you participate on a regular basis? If so, are you willing to share your username?'
I prefer the latter, because it's a question I can ask all candidates, and it doesn't cost me $5,000 to ask. Nor does it cost you $99 to answer.
I guess it's a perspective thing. To me it seems more like you're paying $5000 to get access to CVs by programmers who are dedicated enough to the whole job hunting process and secure enough in their ability that they're willing to pay $100 just to list their CVs. The stack overflow score seems fairly incidental to the whole process.
The reality is that I'd never just use StackOverflow to find candidates, so these are people who have other ways to find out about the positions, and apply.
I'd rather see the candidate spend the money on an ACM membership (or similar professional development) than spend it simply trying to reach me.
That said, I think I now have a pretty good idea of the value proposition as it might exist for other employers, but it seems clear that the equation doesn't balance for me. Thanks for clarifying your argument, I appreciate it.
I bought ACM membership. They never sent me the monthly CACM on time (8 months before receiving 6 issues in a box(!!!) in the mail, then 1 more 3 months later, still missing 5), and they sold my address and now I receive junk mail advertising academic books.
Joining ACM: just a plain bad idea. Only value is older papers not available ungated on citeseer or elsewhere.
I agree that a raw number means very little, but the info gleaned from looking up their stackoverflow account can tell you quite a bit.