Real-time hiring with prediction markets

by ryan_b 1 min read9th Nov 20189 comments


Epistemic status: speculation

I have a couple of assumptions about hiring:

  • It is difficult and often unpleasant.
  • No one is very good at it.
  • It has a lot of hidden costs.

I have a further assumption about the need for labor:

  • The people doing the work now have the best information about the amount and type of work needed.

Most of the time I have seen a new position open up, it worked approximately like this: management notices a shortfall -> they make a case for a new position -> there will be negotiation with senior management/HR/finance about the needs and available resources -> greenlight to post a job opening and collect applications. This seems like costs begin to pile up until management takes notice, and then continue during the hiring process. The goal is usually to get a pool of qualified applicants, and then make the lowest offer that will get accepted. Controlling explicit costs like compensation seems to be the dominant concern; there is little to none for lost productivity.

I think if a prediction market were implemented where current employees bet on the value of applicants (their future colleagues), we could get better hiring decisions. If we combine this with continuous hiring, which is to say always accepting applications, we can cut out the entire above process.

As new applications come in, the market will price them. A high price is worth a look, because it means the team wants to add them. Price inflation overall implies the team is in need - though I am not clear on where additional funds would come from.

It seems like this leads to a condition where a company could hire productive people as soon as it needs them, without even having to recognize that need. This sounds ideal to me.