| Yes, certainly - I think the thing I mean doesn't conflict with the thing you mean. Referral programs will pretty obviously bring your hires similar to the hires you already have, with all the attendant problems. College insularity, but also demographics and even specializations. What I was thinking about was more specifically finding strong candidates from populations that don't look great. Maybe State University has a weak CS program, but a handful of really good candidates. (This is probably the case for any weak program of sufficient size.) Those candidates probably won't be obvious on paper unless they're hyper-motivated, because weak programs lead to limited opportunities and uninformative grades. But if Jess from State University does Google Summer of Code and makes something awesome, maybe she gets a job. At that point, Google hands her the referral form and she can pick out her strongest classmates better than any recruiter could, even if they haven't done anything flashy. (My experience, at least, was that the top students are at a minimum aware of one another.) So it's not going to solve a demography problem; if your direct hires are 50% Stanford, your referrals will be too. But a lot of those Stanford referrals might be people you'd reach anyway, so referrals are valuable in inverse proportion to the ease of other recruiting. The harder it is to spot which candidates are good, the more benefit you gain from asking their peers. |