I was just thinking the other day about the value of unconventional interview questions. You know, questions like "Why are manhole covers round?" or "How many golf balls can fit in a school bus?" (I was asked the latter question in an interview a few years ago). Google's been applying the concept of "Big Data" to their hiring practices, and have recently denounced the value of those kinds of questions. It's hard to argue with data.
But I wonder… while the questions may not be good at identifying the right candidate, is it possible that they bring some secondary value? Maybe they say something to candidates about the culture and values of the company: "We aren't afraid of tackling difficult problems, and we value creative solutions." If you, the interviewer, know that the question is a throwaway question, then is it really that bad giving it anyway?
I suppose the risk is that you inadvertently let the response to questions like these influence your decision. You could end up hiring somebody who's good an answering questions but poor at doing the work you want them to do. It's unavoidable... you're bound to let all sorts of presentational aspects (like how they dress, or how confidently they answer questions) influence your opinion. These questions probably just add to the noise of an already noisy process. In fact, I'm sure there are better ways of signaling to candidates that you value creativity and innovation. Like, actually being innovative is a good place to start.
That's it... I've convinced myself. If you use clever interview questions, it's probably time to retire them. There are better ways to do what you're trying to do.