I do almost the same but a bit differently. Like a lot of people have suggested here that they can't share their company's proprietary code (neither can I). So I have cooked up some sample questions asking people to code for a app involving REST and CRUD (because that's resturant what we do at office). It's not much work and can be done in 2-3 hours. Then I get down to discuss their answers and the 'why' questions around their approach. Always gives me a pretty good insight on their work style and doesn't give the candidates any opportunity to cheat (in these remote times) because eventually they would get caught whole explaining their code.
This approach has given me excellent candidates every single time and also led to a lot of time saved otherwise wasted.
> It's not much work and can be done in 2-3 hours.
I bet it’s much more work than that. It’s maybe 2–3h for you, who has reviewed dozens of submissions, have designed the problem, and know what the actual solution is.
For the rest of us, it takes trial and error, implementing, polishing (because of course you want to show your absolute best code when being judged solely on your code) and it’s probably actually taking 2–4x as much time as you think it does. No applicant will ever tell you that because they don’t want to look bad.
Try and have people on your team do the challenge, you’ll see how much time it actually takes.
I personally pass on these. If it’s more complex than “implement array.flatten”, it’s going to take way too long and I decline to go further. If it starts with “implement a web app that…” or “an api that…” I don’t read any further and bail.
It's 2-3 hours for a guy having a 2-3 years of experience (which is what I'm looking for). And yes, I get those attempted by the guys in my team before sending it out. And yes, most of the time, I end up flushing out a few bits of the question before actually sending it out based on their suggestions.
In reality though, the time spent by the guys varies. I always ask them how much time did they spend on the question after selection. The mileage varies from 30 mins to 4 days (because they had office work/weekend trips and attempted the question only when they were absolutely free). No, the hiring HR makes it specifically clear to not try to give a polished code during the initial call and only work 2-3 hours on it.
I see where you're coming from, lengthy problem solving questions are definitely not worth anybody's free time. I've seen problems that can take 1-2 weeks to solve. But for me 2-3 (or even going by your maths of 4x3... 12) hours... is definitely worth your free time... Because
1. that's the amount time you spent while giving your 4 other interview rounds in 4 other companies paying ¼th the salary.
2. Or for a similar company paying an equivalent salary that takes 4 rounds of interviews spread across 4 different days (counting your commute time as well).
Implement array.flatten can be Googled in under a minute. That beats the purpose of a test. I'd rather take a telephonic round and be over with it. The core idea behind my shift was to ensure candidates can't cheat. (And I hate video calls because a lot of them end up having connectivity issues during the interview, especially during the difficult questions).
> I don't read any further and bail.
Yes. This. You're the kind of guy I don't want my time wasted with. There are a lot of intelligent candidates who'd have definitely aced a telephonic round (or even this written test for that matter) but bail out at the sight of "implement a...". In my experience, you're the kind of a guy suffering from Dunning-Kruger syndrome. The good work is there, but the headaches that such candidates cause on the other 90% of the times makes it not worth my time.
We can argue this to the world's end but in the end what matters is the demand and supply market forces. You can call my process a waste of your time. I'm sure you'll end up getting a better job the next day. And on the other side of this coin, even I'll end up getting a better candidate than you. In my experience, all the candidates that I've hired in this format have been marvelous, so I'm gonna stick to this.... At least until I keep getting great candidates.
Sure, I'll make sure you get your salary on time in case you get selected. You're worth getting selected, yes? Why don't you sit for the interview and we'll find out.
Or may be since your time is so precious, we can just skip this written-test-hogwash altogether. Let's just do the traditional telephonic round, shall we? Ohh and by the bye, so long as we are on this topic, you'd pay me for the time I had to waste on you in case you turn out to be a garbage of a candidate, fair?
Well, if they made me stand in line for 2-3 hours, then decided not to give me the food that I paid for because I'm not a culture fit for the grocery store, then yeah probably.
And do you ask payment for the comp-off against the time-off that you had to take to visit for the interview.
And do you ask payment for the gas spent while driving to come attend the interview.
And do you ask compensation for your intelligent views presented during the interview when they asked you a technical problem related to their real life scenario and your answer ended up solving that saving them millions.
Oh, you went way overboard just for the sake of your argument. Interviews at such positions have a different format. And different expectations. Heck, even the job contracts are customised at those levels. Nobody's gonna give those candidates a written technical round.
Not necessarily. If the HR is short-sighted, and that happens a lot for a lot of even famous companies, they'll say you have already saved the million and that your salary expectation is not worth it. Trust me, shit like this happens.
Theoretically yes (and they'll sure as hell contest that claim). But I've never seen anybody EVER do that except _very rarely_ for positions in Legal.
This approach has given me excellent candidates every single time and also led to a lot of time saved otherwise wasted.