· all posts

Hiring is an intractable problem

One thing that I’ve consistently seen in hiring is that people can hack the interview process, often quite successfully. This shouldn’t be a surprise since most companies have procedures in place for firing people. If people didn’t hack the interview process, there would be almost no need to have procedures on letting people go, except for misconduct or a layoff.

Because companies have such procedures in place, hiring good talent seems to be an intractable problem. Often among technology companies there is something called a “performance improvement plan”, and why would that even exist if the interview process could select for people who would never need a performance improvement plan?

It’s almost as if the industry has acknowledged that people who may not be a good fit for the job will slip through the most rigorous interview process and is compensating for it on the other end.

So what can we do?

Although it may not be possible to have a perfect interview process, we can get reasonably close, at least in the engineering field. Essentially, the idea is to create work-sample tests which generate objective, unbiased information about the candidate’s abilities, are resistant to hacking, and which provide enough information to almost disregard what’s written on the resume.

For the last several years, work-sample were our most important hiring decision factor. We relied on them almost completely, and in doing so, we multiplied the size of our team and retained every single person we hired.

Resumes are good, but people are not and sometimes lie. And regardless of what people say on their resume, it may just be that they won’t like the work which they are applying for, so it’s best to give them a taste of it as early as possible.

Resumes are a great example of being able to optimize for the hiring process. If two people have equivalent experiences, one person may be able to sell themselves and their experiences better. The same information presented in a different way can have a huge difference in public perception. It’s why you hear people say they have to “work on their resume” when preparing to look for a job. If people couldn’t hack the resume, they’d be practicing their actual skills while looking for a job, not practicing the art of resume writing.

There’s the argument that a person who can’t sell themselves wouldn’t be a good hire, and therefore the process does accurately select for the best people. This is certainly true for sales positions, where the skill of selling yourself can be a great proxy of your ability to sell other things. But for other positions, it simply isn’t the case. If a company is hiring a woodworker to do nothing but make exquisite pieces of art out of lumber, it doesn’t really matter if the person can sell themselves well. The company who hired the best woodworker will have wonderful pieces of wood to sell, with the help of the sales team. The company who hired the “woodworker” who could sell themselves now has two salespeople, zero woodworkers, and will soon be filing for bankruptcy.

Is it really an intractable problem?

As I go about hiring, I want to keep in mind the possibility that hiring may not be an intractable problem after all. If it isn’t, the techniques for such a hiring process would result in an objectively happier place to work, with everyone enjoying their role, growing their skill set, and staying around long enough to learn from their own mistakes.

An aside: lots of people in the software field tend to leave a place too quickly, which prevents their own work from coming back to bite them. But I haven’t met a better teacher than my former self. After a while I can look back and understand why I made a decision and how well it scaled (or didn’t). Staying around makes it more likely that you will slam your fingers in the car door of life. And I bet you won’t make that mistake twice.

Looking forward

I’ve seen that places which hire slowly and intentionally have this process somewhat figured out, and I’m excited to figure it out for myself as well. Hopefully in the future I can share what works and what doesn’t.