Ask HN: Is it reasonable to have job applicants complete a test before applying?
The test is something like this: I've described a fairly simple json specification in words and requested that applicants submit a cover letter that conforms to that spec. The spec has an optional value called challenge_checkvalue which can be derived by completing a challenge - querying a url with an id between 00 and 99 for json payloads that may or not conform to the same spec. The checkvalue is derived by combining the ids of all the ones that are not compliant.
So it's a decent amount of work. I've deliberately made the number of payloads to check high enough that no one is tempted to check them manually. Validating JSON, even making http requests like that is not something that every valid candidate will have ever done, but I think that these concepts should accessible with some research.
My question is; is this too much to expect people to do before I even look at their application?
19 comments
[ 3.3 ms ] story [ 45.4 ms ] threadIf someone is applying for a lot of positions at once, an up front homework problem can be seen as unreasonable before even a basic semblance of role compatibility is established.
Why not offer it as a HW problem after a resume and phone screen?
If you haven't completed it or had someone complete it then you shouldn't be sending it out at all.
FWIW, speaking as an SDET lead, I'll do something like that voluntarily, time allowing.
But I A) don't submit applications, I send resumes, so let's start there; and B) wouldn't do something like you describe as a formal prescreen. Before I've talked to you, you have maybe half an hour of my time if I've heard of you, maybe five minutes of my time if I haven't.
After you've convinced me I'd like to work for you, then I'd give you an hour or two on a project, interest depending. Any more than that, and you need to pay me. You'd also probably have to be on my short list of final employer candidates. I'm interviewing y'all too, you know. I'm more inclined to do homework after a successful round of interviews than beforehand, mostly because there'll probably be a job out there I'm equally interested in that doesn't require this. I do pay attention to how much of a pain in the ass a company is to apply to, and it does weigh into my decisions to accept (or even continue).
I realize this comes across maybe a touch arrogant, but it's honestly accurate. Ask yourself this: Do you want the type of developer desperate enough for a job to jump through hoops?
SDET is an extremely competitive niche for good developers, since the QA path tends to filter people out into other disciplines before they hit a good amount of SDET skill. A good SDET can afford to be choosy. This is more friction than I'd put on a process to hire for my team.
However, the universally negative response to having the challenge this early has made me rethink a bit. I think I may make it much more optional and/or apply it later. Maybe look for places to hang it out on the web...
I wouldn't use this as your only feeder--as someone else mentioned, puzzle affinity isn't going to be a 100% crossover with the people you want to track. I'd see a puzzle as one of a few tactics to all try simultaneously.
That's probably another argument against making something like this mandatory--it essentially becomes your only feeder/filter if everyone goes through it.
I can understand your desire to pre-filter somehow. But: I would skip this job ad as it is.
You need to have a real good selling point for your company/product/project that makes the applicant really crave the job. (You left that one out to anonymize your post, right?)
You might deny yourself a pool of smart people who either dislike jumping through hoops or have easier access to jobs they deem equally (or more) attractive.
You might attract folks who are more desperate than skilled but nonetheless able to solve your task.
On the other hand: I am not a lawyer and i don't know under which jurisdiction you are operating. Some countries have laws requiring you to consider all incoming applications. So please consult a lawyer, too.
1. sell the opportunity 2. have candidates self-select out
Every job pipeline is going to follow traditional funnel math and you are starving your funnel.
The test you've proposed is more appropriate for later in the funnel when they are more willing to trade their time for the potential opportunity.
In terms of economic efficiency, it's much better overall to do resume-screening before the coding challenge. Both you and the candidate spend something less than 5 minutes each checking if it might be a good fit before committing to spending much more time on the process. With the challenge that early, you either make people waste time when you could have found out more cheaply, or you convince people that it's too expensive to find out more about your compatibility.
If you advertise a salary 50% over market I'll do it.