Hire Lucky¶
Jeff Atwood wrote about FizzBuzz in 2007. Google used to use weird brain teasers, but no longer does. That hasn’t stopped a plethora of companies who think they’re copying Google from asking tricky puzzle questions. No matter what the approach, though, interview questions measure how well someone answers interview questions.
We can try to make interview questions feel more like the kinds of problems that come up in our day-to-day work, but there’s really not much ground you can cover in a 1-hour interview. You might stumble upon the candidate’s very narrow area of competence and completely miss the giant space of incompetence, or vice versa.
And even if you do manage to find someone whose skills are a perfect match for the job they’re going to do, bad things happen. Sometimes you’ll hire the perfect person for a position, and 2 months in they’ll have something happen in their personal life that either makes them ineffective or causes them to leave entirely.
My proposal is to throw all of that away. Don’t waste all that time reading CVs, interviewing, and talking to each other about all the people you’ve interviewed. It’s super expensive and a crap shoot at best. Instead focus on the one thing that’s simultaneously hardest to select for and best to have: luck. Just roll some dice. Hire the luckiest people.
“But Nathan,” I hear you saying, “what if their luck is transient?”
This sounds like a reasonable concern, and honestly it’s probably the only one worth addressing. The solution is simple though. Don’t just roll a single 20-sided die for each candidate. Anyone can hit a 5% chance once. No, roll 2d20 for each candidate. 2 20s in a row, and you have a hire. Anything less, and you keep looking.
Alternatively, if you’re not some sort of weird table-top gaming nerd with lots of icosahedra sitting around, you can roll regular dice. To get the same level of luck as the double-20s, you need between three and four 6s. If they get three 6s in a row, and the fourth die comes up 4 or better, they’re lucky. Hire them. An advantage of the 3\(\frac{1}{2}\)d6 approach is that you can roll the dice on consecutive days. That reduces the risk that you’ll catch someone having a lucky day.
Note
It has been suggested that the truly lucky people will not be hired by an organization that implements this hiring scheme, because then they would be working for a company that selects their employees randomly. Don’t listen, though. That’s just the unlucky whinging about their lot in life.
There’s still the problem that some people’s luck runs out though. So I suggest an additional field in the annual performance review: luck, on a scale from 1 to 6. It’s the easiest one to fill out. Just roll a 6-sided die and record the result. If it’s below 4, time to think about a PIP.