White board interviews in tech are a scourge. Companies shouldn’t use them, and developers shouldn’t work for companies that do. I’ll talk about why in a moment, but first, let’s discuss what they are and what they are not.
White Board Monolith
White board interviews usually consist of the interviewee in a conference room with a white board and marker, while a parade of employees (either singly or in small groups) come in and pose programming questions for the interviewee to answer on said white board.
The idea is that by having the interviewee write on the board, interviewers will gain insight into the interviewee’s thought process as it occurs and test their problem solving skills in an isolated, neutral environment. For this reason, the question posed are usually challenging problems that can be solved with relatively foundational concepts in computer science, rather than involving platform/tool specific topics.
After the interview, the interviewers usually gather together and discuss the candidate’s performance. Sometimes they take a vote, sometimes they just inform the hiring manager qualitatively, sometimes one senior dev makes the yea or nay decision. Regardless though, this element of the hiring process is usually a large deciding factor into whether the candidate eventually is offered a job.
(Many) Fatal Flaws
Unfortunately, the white boarding process achieves none of its goals. Interviewers don’t get insight into a candidate’s thought process; they don’t end up knowing anything about the candidate’s problem solving skills; and the environment is neither isolated nor neutral.
The problem here primarily stems from the phrase above: “foundational concepts in computer science.” In an attempt to allow candidates to be compared, interviewers try to choose questions that rely on topics that underpin most of computer science today – keyword being “underpin.” They’re not topics you actually have to use day to day as a developer. As a result, they are, paradoxically, a niche subject area, since no one is working with them daily.
The issue here is the same as any standardized test has: how do you separate someone who’s just having a bad day from someone who doesn’t know how to do it at all? Perhaps worse, how can you tell some one who really knows their stuff from someone who got lucky and knew only the questions you asked? You cannot be confident you know one from the other.
Second, for the same reason as above, you’re not actually testing problem solving, you’re testing their memory for academic CS topics. There are books written on how to solve these sorts of questions in interview settings. You’re testing whether the person studied, not whether they can figure things out.
The environment isn’t isolated if there’s someone in there with you (the interviewer). Simple as that.
And in fact, you’re actually isolating the candidate from things they should have access to if you want to know how they will actually do the job: a computer, Google, the internet, an IDE, and, probably most importantly, time. Your employees don’t program on a board of any color; they program on their computer. They Google things. They discuss design patterns and architectures with their colleagues. They sleep on problems and wake up in the middle of the night with the answer.
You’re not testing any of these things at a white board.
This, to me, may be the worst sin. White board proponents insist that it gives them an objective idea of how smart the person is. This couldn’t be further from the truth.
Interviewers are humans too, and they are affected by projected confidence, implicit bias, manipulation, and expectation. As a result, white, cis presenting males do better in these types of interviews than women, people of color, or non-cis folk, simply because they LOOK like the type of person who’d do well. I routinely hear horror stories from women who solve a problem correctly only to have the interviewer tell them they’re wrong and then describe the woman’s solution back to her as the correct answer.
White boarding is well named: it perpetuates white male dominance in tech, while providing companies with a cover of objectivity.
What we can do
I, now, refuse to work for or with a company that uses white boarding interviews to evaluate candidates. I will urge them to change their process, and if they don’t, I move on.
I believe it is incumbent on people like me (white, cis presenting males) to do this. We benefit from this, so it’s on us to dismantle it. My resume is the kind you show to your investors to say “see who we got to work on our product” – so it’s on me to show them that they’re missing out by using a racist, sexist hiring process.