Hi! Can here. Let’s talk about job interviews. This is a longish ramble, but it’s something I’ve been thinking about for a while.

A common refrain you hear in tech is how bad the job interviews are. I don't know if it's specifically a tech industry problem, but I can confirm the premise. They suck. It doesn't even matter what side of the table you are sitting on. Preparing for interviews is not fun, and being in them as a candidate is physically and emotionally draining. Being on the hiring side does make it a tiny bit easier, but not by a lot. Some nice people get a jolt of energy from meeting new people, and others, who are less nice, get their high off of asserting their dominance over the poor candidates. I am more…ambivalent?

I mean, I don’t hate them. More often than not, I look forward to meeting new potential co-workers, getting to know them as the people they are, am happy to answer their questions about the company, and sure, and do my best being objective while I judge them on their merits. I’m reasonably good at it too. I’ve probably given more than a few hundred interviews [1] so far, and no one ever left any of them crying or sent me angry emails afterward. That might sound like a low bar to clear, but you'd be surprised. I have done both. [2]

My main beef with interviews is that people read too much into them. The interview itself is rarely the meat of the hiring decision even though everyone pretends they are. It is a factor, and it's one of the more important ones, but that's about it. There's an entire cosmos worth of chaos that decides whether you get the job or not, no matter how your interviews go.

Quadrants to the Rescue

Let's simplify things a bit. Consider the two possible outcomes of an interview; a candidate does well or not. And then, two binary outcomes for the hiring decision: hire, or pass. There are a lot of gray areas between each of those. Hiring panels are rarely unanimous and there are many cases where a single hire vs. pass decision drags on for weeks and weeks.

But, you know, let's ignore all that for a second and look at a model:

The above is sort of what you want. Good talent gets an offer, and bad talent is shown the door. 

But, there's a more fundamental problem with this chart. It’s all wrong.

Hiring decisions are not rational, and they just cannot be. They are, if anything, like returns on the stock market. Any individual hiring decision is a random toss-up, and at scale they are mostly determined by both skill and luck. This doesn't mean there's no objectivity in the system. There is. But whatever the tiny bit of objectivity, you can only observe it in aggregate.

Again, there are patterns. A good candidate does end up getting more offers than a bad one, and good companies that approach the process with more rigor do end up making more optimal decisions. However, even those are tempered by chaos.

A more accurate version of the chart, for a single interview, above would probably look like this:

Bad News for Good People

Is this really a bad thing? I am not actuallly sure.

One way to read this is that if you ever got dinged, that is rejected, after an interview that you think went really well, it probably wasn't your fault. It could be, and it probably was to some degree. But it also could be that you just got lost in the frat. That should be somewhat consoling.

If you think this is a stylized exercise, let me list you a couple of cases where I think we should have hired a candidate yet we did not. I am going to blur and fuzz some of the details but these are all real cases where I have first-hand knowledge.

• The recent graduate who excelled at every interview, blew everyone's mind away with his answers to even the hardest problems, was exceptionally kind, really interested in what we were working on. The candidate had the poor luck to have scheduled his interviews before the team decided to only hire senior talent for a year. We tried to get our director to make an exception, but alas. We passed.

• The candidate who managed to pass a phone screen, get called in for an interview and then passed them with flying colors, only to be passed later. It turned that he was on a visa at his current place, and we didn’t sponsor visas back then. We never explicitly mentioned that in our job descriptions, but the recruiter was supposed to check for it. We considered making an exception, but the organizational embarrassment coupled with the complexity was too high a bar.

• The designer who had an incredible portfolio knew how to code (we did have this as a requirement), and excelled at the interview process too. This time, the candidate had the bad fortune of not having a degree from a school in the US. It didn’t matter the candidate literally spent all their life in the US except for an undergrad degree abroad. Somehow, our hiring manager was convinced that the lack of a US-degree would mean the candidate would not be able to capture the essence of American culture. I mean, are you fucking kidding me? We passed.

• The senior candidate who excelled at all the interviews with flying colors, showed quite a bit of interest but then later dropped off the radar and whose position was filled quite abruptly before we could reconnect. We had the extremely poor judgment to read into his radar silence as unprofessionalism while it turns out they had lost someone in their family and were out of the country. You could argue this was sort of a force majeur case, but I still think we could have waited more than a few days before pulling the trigger on a much weaker candidate who ended up leaving in less than a few months.

This is all horrible.

Good News for Bad People

But, there is also another way to read the chart above. You could also use this chaos to your advantage. If you are, for example, a candidate who often gets called in for interviews but never receives an offer, this ambiguity should be good news to you. Especially if you aren't really looking for The Job, but a job, the strategy is to simply get more interviews, and at some point, you should get an offer.

Some people make the argument that the default case for an interview is a pass (a no-hire), and it's up to the candidates to prove their worth. This is largely true, and false positives do happen less than false negatives (I think?) but they still happen. In fact, again, let me tell you about a few slightly altered examples:

• The candidate who we hired because they had gone to a school that was like...the toughest school in the world, and honestly you hire anyone who shows up with that diploma,

• The engineering manager, who had never written line a single of code in their life, never managed a team that worked on production systems, but was hired because every other engineering manager was on the cusp of quitting if we didn't hire someone new,

• The candidate whose recruiter had accidentally said yes to, and then we decided later that, they aren't bad enough to embarrass ourselves over our organizational blunders (they turned out to be bad),

• The candidate whose hiring manager had accidentally said yes to, and then we decided later that we would rather take the bullet instead of risking yet another organizational blunder (this one turned out to be really good!),

• The tens of hundreds of people who had been hired because there were empires to be built, headcounts wide open with money aplenty coming from investors, and it's better to hire ahead of the curve even though that means we are adding engineers faster than we are increasing revenue by 2X

Again, this is both depressing, but really, maybe it’s all fine?

Mismatched Expectations

It's unreasonable to expect that much rigor when you have tens of steps in any process and multiple humans with feelings and emotions at every step. You can try, but it just won’t happen. You can try to remove as much of the human as possible from the decision making, but even in that meta-process are embedded human decisions. What are the things that matter in a job interview? What do you ignore, and what do you measure? Do formal qualifications matter, or not? How do you measure someone’s potential contribution, for example, to a team’s psychological safety? You can make things data-driven as you want, but pure objectivity in hiring is not a reasonable goal. Extremely qualified people fail to contribute at bad workplaces, and candidates who appear as weak initially can excel at companies that help them grow and succeed.

Why you are hiring people also matters. There are a couple of models here too. The capitalist way to look at it is to look at workers as tiny productivity producers, where if you sprinkle over enough capital (or capital over them), that can build widgets for you to sell. If this is how you approach hiring, for example, you would want your interviews to be as dehumanizing as possible. I am not using the word in a derogatory sense, but in that, the goal is to remove the human from the equation. The logical conclusion of this model is that hiring people is a sign of failure.

I have had interviews like this, and I have to say, I enjoyed them. I felt as human as an aluminum slab that makes it way through a metal press to become a piece of car would feel throughout the entire process, but the feigned objectivity was amusing to watch, if only to how truly bought everyone was into the game. I am not sure if my performance was genuinely relevant to the work I'd be doing (it wasn't). I even managed to score a few offers.

The other end of the spectrum, of course, is that you hire people because you want to work with people. The productivity is sort of a nice added bonus, but it's not really the end goal.

This actually happens, especially when there’s too much money sloshing around at a company. For example, you might be making monopoly margins for years, and then one day you wonder what happens if you bring in a bunch of smart people and give them a vague direction like "why don’t you invent immortality?" or "should cars have drivers?." You might even get away with it for a long time until you bring in a CFO from Goldman Sachs, and she tells you to stop paying people hundreds of millions of dollars in bonuses.

Or, you might just want to hire people to increase your headcount while you can, because that's how you get promoted and by the time people figure out you have way too many people on the payroll, it is someone else’s problem to lay those people off a couple of years later.

The majority of hiring decisions, of course, fall somewhere in between. Many big companies have departments that operate on one end of the spectrum. Smaller firms swing from one side to the other as they grow, raise money, or market and macro economic conditions change.

And don’t get me wrong. The interviews are important, and every company should strive to make them as good, and as objective as possible. But the decisions are too random on an individual level to either celebrate or lose sleep over.

[1]: At first this number seemed too high. But then, a quick Gmail search showed I’ve interviewed at more than 30+ places over the years. And given how many years I’ve spent at fast-growing companies, where I interviewed several people a week often, it seems low now. The big red flag, I guess, is of those a few hundred, I can only distinctly remember at most 4-5 people.

[2]: Not that I should have.