Consulting tasks are rapid fire engagements where the client is charged high hourly rates for a concentrated load of supposed consulting firm wisdom. Oftentimes that wisdom is delivered using paint-by-numbers frameworks that people have prebuilt the powerpoint slides for, but well, companies keep buying it anyway for some reason.
Because that's the case, consultants need to appear to know what the hell they're talking about most of the time -- they need to be quick on their feet. You don't hire a bunch of consultants who fumble for words when presented with a problem only to say, "We'll get back with you."
So, if that's the kind of person you want -- quick witted, good using the back of the napkin, a winsome bullshitter -- how do you interview them?
You ask them bullshit questions! How many gas stations are there in New England? How many marbles fit in Obama's footy pajamas?
If your goal is to hire bullshitters, make them bullshit. And the funny thing is, an interview is an ideal place to observe bullshit. So management consulting firms are set up well to find ideal candidates in interviews.
The frightening thing is that this approach has trickled into other businesses and professions where it doesn't belong. Probably because these firms are viewed as hiring the cream of the crop. Shouldn't we all do what they do?
The famous case is of course Google. These Car Talk style puzzler questions are a natural goto, because they seem to tease out some kind of genius within the short timespan of an interview, right?
If you're not McKinsey though, this person might be the worst person for the job. I see hiring as being not all that dissimilar from attacking a data science problem/product development project. You have to know your goal. I talk about this at length in the conclusion of my earth-shattering data science book.
Recently I learned of a chain of mall-based cookie stores that decided they'd hire MBAs to run each shop. They thought having type-A entrepreneurs in charge of each store was a great idea.
But then the MBAs started doing things like adding extra chocolate chips to the standard recipes. And that didn't end well. What did they expect? You hire an MBA, they're gonna make it rain chocolate Cookie Monster-style.
Only when you know what you want, can you interview effectively. Otherwise, you're going to concentrate on figuring out if someone is excellent in all the wrong ways -- one way this manifests itself is by hiring people that look and sound just like you. You throw them a little brain teaser from your own work. Sure you've got more domain knowledge and it took you two days to figure it out, let's see what this candidate can do in 5 minutes.
But does the position require your clone?
What will this new hire do? Will they need to communicate well with other teams, or are they only tweaking models and code? What's more important in this role -- speed or thoroughness? Are you attacking a solved or unsolved problem? Do you give a rat's ass about publishing and presenting research? Do you want someone who's extremely autonomous or someone who will depend on their teammates? Are there opportunities to advance or will this person need to be content staying in this role for a couple of years?
Contrast the cookie example with a company like Enterprise Rent-a-Car. Each year they hire thousands of recent college graduates to entry level positions. And they don't look for the McKinsey types, "blue chip" excellent students. They look for the students with OK grades and a hunger to prove themselves (go-getters like Bob Maplethorpe below). And this strategy has worked well for them.
But figure out exactly what you want first, otherwise you'll be disappointed in each and every hire.