Programming Puzzles Are Not the Answer – How Spreedly Does Work Samples

I applied a few months ago to a position at Spreedly and welcomed the code sample process. It is only one part of their hiring process, like the article mentioned. I loved getting feedback on the results too, it turned out someone had a better score and that's ok, I knew exactly why and actually learned something about that problem. Most jobs you apply for and don't get you just hear deafening silence back.

It also gives you an idea of what the company focuses on - delivering working software. I'm not saying companies that don't do work samples don't focus on that but as a candidate it makes you wonder if playing games is part of the job. You gotta talk about real code, look at real code and discuss it.

I have also been a manager and sat on the other side of the hiring table and it is hard to judge the technical chops of people. We used quizzes of deep technical items that should have been known by someone if they had done the work before (e.g. MTU settings for a db cluster) and that helped weed out people but it really wouldn't give a sense of if they were a 100% good fit, you need other ways to check that too.

I like the analogy of developers to chefs and cooking. Programming puzzles is like judging a chef on cute tricks to slice an onion. Would you hire a chef based on their onion slicing tricks or their ability to cook an amazing meal?

I recently applied for a software engineer position at spreedly. I looked into what spreedly did and thought I was a perfect fit, and I still think that. I thought I did good on my work sample all things considered (see below). Spreedly’s metrics did not agree. One thing that I really liked about spreedly was the fact they told me why I failed and what I failed on. Other companies are treating their hiring process like a trade a secret; if you’re trying that hard, it’s probably not working IMHO. Most of the spreedly work sample was great. You had to dig into source code, reason about the code base, etc. What I did not like was the gotchas. There were not a lot, just a few, but it’s part of the reason I wasn’t considered. I don’t have time for games, pure and simple. I’m never going to put the time or effort into a code sample that I would as if I was getting paid. Not when I have real work to do, not when I’m playing the job application lottery and I might go bust vesting too much time with one company’s code sample. It seems almost narcissistic of a company to ask for that much. It’s kind of analogous to the fruitless question: “Tell us why you want to work for Hooli?” The other feeling I consistently walk away with is that I’m a guiena pig; I’m a beta tester for your hiring process and that does not feel good one bit.

All of that said, I have to say spreedly’s work sample was the most realistic. During the process I did feel like there was a purpose to what I was doing, which is great for someone like me who has a hard time doing “puzzles” and the like because I see no utility in it.

Do I think blind hiring works? It could, but not the way it's being utilized. When I first heard of blind hiring, I thought this is great. No bias, no bs, just code. I've been on the hiring side of the table before and 90% of the applicants were full of it in regards to what they put on their resume vs what they actually knew. The worst part is after you spend the whole day interviewing worthless candidates you have all your regular work that has piled up during the process that must be done. A lot of time and energy gets expended in that process and wasted. I truly understand both sides. As a tool to filter out "those people", it's great. However, blind hiring seems like it being used to try and find the perfect candidates. Perfect candidates do not exist; just as perfect humans do not exist. Great test takers do exist though. Personally, I have never correlated good test takers with good employees or good programmers. In fact, it’s always seemed to be the opposite from my experience.

  So, what’s a better way to skin this cat? I honestly don’t know, but if I think of it I’ll be sure to make a proprietary app and sell it to all the companies I’ve done blind samples for : )