The Technical Debt of Hiring(skillsapp.com)
skillsapp.com
The Technical Debt of Hiring
http://skillsapp.com/blog/2010/12/2/technical-debt-hiring/
7 comments
We decided to focus on the employer side of things at launch, for several reasons:
* We have to start somewhere
* Easier to ship a tightly-focused product, easier to build a message around it.
* Not sure where the revenue is on the candidate side.
* Honestly, it's just scratching our itch :)
We've given some thought to the candidate side, and hope to do something there in the future, though. You're spot on about the recruiter firms.
* We have to start somewhere
* Easier to ship a tightly-focused product, easier to build a message around it.
* Not sure where the revenue is on the candidate side.
* Honestly, it's just scratching our itch :)
We've given some thought to the candidate side, and hope to do something there in the future, though. You're spot on about the recruiter firms.
As someone who is looking for a job, I'd be more than willing to shell out a couple of bucks if I knew it would put me in contact with people I'd want to work with.
Right now on the candidate side of things it's pretty much the same situation. Go through job listings, send resumes to any that sound interesting, do about 4 phone interviews a week, pass on roughly 3 of them because they either aren't working on something I find interesting, the position itself isn't something I'd enjoy, or my gut tells me to walk away due to the impression I got from the person I talked to. I will pay you to not have to deal with this.
Right now on the candidate side of things it's pretty much the same situation. Go through job listings, send resumes to any that sound interesting, do about 4 phone interviews a week, pass on roughly 3 of them because they either aren't working on something I find interesting, the position itself isn't something I'd enjoy, or my gut tells me to walk away due to the impression I got from the person I talked to. I will pay you to not have to deal with this.
This may be totally stupid, but couldn't you have a database of companies looking for a job ("need 8+ PHP, 6+ MySQL") and developers ("have 7 PHP, 7 MySQL, 9 Ruby on Rails")? Of course, actually getting that going would be difficult, but Monsterboard et al. seem to be making a lot of money...
If it is your startup, and if you're in the bay area, and even moderately social, I don't understand why for your first ~50 employees you'd be posting jobs to the general public at all (or at least, treating it as a very low priority thing, like how VCs react to random bplans mailed to them without an introduction).
If you as a founder don't actively go out and meet and recruit your core team (maybe the first 20 people?), or meet them through referrals from other team members, trusted friends, etc., you are doing it seriously wrong.
The only way I'd hire a random person is if it was for a specific role where I didn't know anyone (say, CFO for a complex ad network, although I know plenty of non-CFO accounting and finance people). Even then, I'd want references from people I already trusted. Mining piles of resumes is not the efficient way to find good people, and really, for a lot of roles, I want to know someone is at least a B player, and would love an A; I'd rather a B than someone who may be an A or may be a C. This is especially true for support, operations, etc., and anything where I don't have domain expertise to evaluate competence directly.
If it's a specific domain, and you've exhausted all your 1-2 degree connections, maybe find the internet nexus of discussion for that topic and participate for a while until you can see who is great, then try to figure out who is available. Pull, vs. push.
Maybe this is joshu's new secret startup :) (as speculated on Quora)
If you as a founder don't actively go out and meet and recruit your core team (maybe the first 20 people?), or meet them through referrals from other team members, trusted friends, etc., you are doing it seriously wrong.
The only way I'd hire a random person is if it was for a specific role where I didn't know anyone (say, CFO for a complex ad network, although I know plenty of non-CFO accounting and finance people). Even then, I'd want references from people I already trusted. Mining piles of resumes is not the efficient way to find good people, and really, for a lot of roles, I want to know someone is at least a B player, and would love an A; I'd rather a B than someone who may be an A or may be a C. This is especially true for support, operations, etc., and anything where I don't have domain expertise to evaluate competence directly.
If it's a specific domain, and you've exhausted all your 1-2 degree connections, maybe find the internet nexus of discussion for that topic and participate for a while until you can see who is great, then try to figure out who is available. Pull, vs. push.
Maybe this is joshu's new secret startup :) (as speculated on Quora)
As one of the founders, I readily admit that this product is not for every hiring situation. That's fine! We stil think that there's a large use-case/market here.
There are many startups, and many more that aren't "social". There are lots of small technical shops that aren't startups at all, and don't have a deep network of people from which to fish out hires.
Eventually, everybody's well runs dry. How to do hiring effectively when you no longer have a trusted reference is a thorny problem. Even if we don't build a perfect solution, there's lots of room to make things better.
And no, we're not Tasty Labs. :) Check our about page.
There are many startups, and many more that aren't "social". There are lots of small technical shops that aren't startups at all, and don't have a deep network of people from which to fish out hires.
Eventually, everybody's well runs dry. How to do hiring effectively when you no longer have a trusted reference is a thorny problem. Even if we don't build a perfect solution, there's lots of room to make things better.
And no, we're not Tasty Labs. :) Check our about page.
Ah! I didn't see a link to the main site at the top of the page so I didn't check out the app description itself (I was not really that focused on it either).
The actual solution you have looks like it should be interesting, and even if you have a pretty broad and deep personal network, could still be useful.
I'd love it if you took outside assistance or some kind of social mechanic to create skillsapps for domains outside your team's expertise. I'd be really interested in crypto, security, protocol design, and satellite/rf network engineering. Maybe partner with existing domain specific certification entities?
The actual solution you have looks like it should be interesting, and even if you have a pretty broad and deep personal network, could still be useful.
I'd love it if you took outside assistance or some kind of social mechanic to create skillsapps for domains outside your team's expertise. I'd be really interested in crypto, security, protocol design, and satellite/rf network engineering. Maybe partner with existing domain specific certification entities?
We are taking outside assistance; for any given topic we engage somebody prominent in the field to write the exam.
"Somebody prominent" is somebody whose name is common knowledge for those familiar with the langualge/framework/etc in question. Could be the inventor (if we're lucky) or a popular developer who blogs on the subject.
If we limited Skills to testing subjects we know, it would be a very limited app indeed. :)
"Somebody prominent" is somebody whose name is common knowledge for those familiar with the langualge/framework/etc in question. Could be the inventor (if we're lucky) or a popular developer who blogs on the subject.
If we limited Skills to testing subjects we know, it would be a very limited app indeed. :)
How will Skills be different from http://www.brainbench.com/?
Regardless of how social you are, your network is always going to be relatively small, by limiting your hiring to that small pool of candidates you'll likely be missing out on many more talented candidates.
There's also conflict of interest issues that are raised by hiring through your network, you need to be very careful you don't let the fact they were introduced through a friend (or a friend themselves) bias your judgement of their abilities. It's very easy to find yourself brushing away the flaws in a candidate because "you know they're good anyway".
And of course there is the issue that if your in the startup community, most of your contacts are probably quite happy working at a startup of their own or one where they have options (hence don't want to leave).
There's also conflict of interest issues that are raised by hiring through your network, you need to be very careful you don't let the fact they were introduced through a friend (or a friend themselves) bias your judgement of their abilities. It's very easy to find yourself brushing away the flaws in a candidate because "you know they're good anyway".
And of course there is the issue that if your in the startup community, most of your contacts are probably quite happy working at a startup of their own or one where they have options (hence don't want to leave).
I can agree with the definition of the problem, but I am not convinced that you can find a technical solution to it.
Currently the only information on the site is : "Skills is a tool for identifying expertise and passion in developers."
I honestly doubt you can do that in an automated way, but hey, we'll see.
Currently the only information on the site is : "Skills is a tool for identifying expertise and passion in developers."
I honestly doubt you can do that in an automated way, but hey, we'll see.
I honestly doubt you can do that in an automated way
I have no idea whether they'll do it, but I am confident we can use software to improve the process by an order of magnitude,and given the amount of money sloshing around the process, I have great confidence someone is going to get rich working on this problem.
I have no idea whether they'll do it, but I am confident we can use software to improve the process by an order of magnitude,and given the amount of money sloshing around the process, I have great confidence someone is going to get rich working on this problem.
Bingo. Hopefully we're also the people you describe in the latter part. :)
When I try to subscribe with my email, I get to this page here:
http://skillsapp.com/subscribe/
which is blank, you might want to look into that. :)
Happening with FF 3.6.12 and Ubuntu.
Happening with FF 3.6.12 and Ubuntu.
We'll be posting previews and screenshots as we continue to develop; there are a lot of great minds and we're happy to get feedback as we go along.
In the end, we're doing testing, but we've given the actual testing methodology a lot of thought. It is obviously impossible to just put a score on "passion", but we've come up with some ways to give employers a way to getsome idea of the candidate's knowledge of/engagement in a field beyond just their coding skills.
We look forward to earning your respect for our product :)
In the end, we're doing testing, but we've given the actual testing methodology a lot of thought. It is obviously impossible to just put a score on "passion", but we've come up with some ways to give employers a way to getsome idea of the candidate's knowledge of/engagement in a field beyond just their coding skills.
We look forward to earning your respect for our product :)
I can think of some approaches:
- Gattaca-style DNA checks for the "coding passion" gene
- The fMRI scans that the Daemon uses to evaluate candidates for faction membership
:-)
- Gattaca-style DNA checks for the "coding passion" gene
- The fMRI scans that the Daemon uses to evaluate candidates for faction membership
:-)
True that interviewing is a resource hog and the post nicely articulates the problem. But I don't think it really counts as a technical debt, if it were you would gain something from it in the short run. For example, by not testing your code you have a short term gain (faster deployment), but accrue long term debt (slower development time as the % of new code vs existing functionality becomes smaller).
I'm also going to argue against the word "technical" here; taking time out to interview isn't affecting your code, it's affecting all the uses of your time and just happening to catch the code as a particular subset. It's not a technical debt or anything fancy, it's just a plain old-fashioned unproductive use of time, one of those things startups can't afford because their biggest disadvantage against larger entities almost by definition is a surfeit of (wo)manhours.
I think the point was that you accrue "hiring debt" as you stop reading details in the resume so you can get through them faster (the benefit), then stop being as detailed in the screens, again so you can get through them faster. Each step, you are increasing the debt, because it moves on to the development team to pay that debt by taking more time out for interviews.
Of course, in this case, all the benefit goes to the manager and all the pain goes to the developers.
Of course, in this case, all the benefit goes to the manager and all the pain goes to the developers.
I can see that... but kicking the can down the road is would be an obviously bad form of debt, like a payday loan or something... Technical debt is a term that to me is more interesting when you might have a payoff as described in Eric Ries article:
http://www.startuplessonslearned.com/2009/07/embrace-technic...
Exactly
I'm really wondering how this is such a problem for them. I've seen my share of tech resumes at this point, and I can't remember a single one focusing on being a "motivated self-starter with a passion for solving problems" or any such similar gobbledygook. One of the big benefits of a career in the tech field is that you don't have to resort to that sort of nebulous MBA-speak to get across what you know and what you can do - and most developers out there seem to be smart enough to take full advantage of this.
I'm wondering why they get so many bad resumes?
I'm wondering why they get so many bad resumes?
I've been using codility.com to make the initial screening less cumbersome and more scientific. It has worked aweseomly well so far! (I'm not affiliated with them)
Resumés say very little about the applicant but seeing them in action through actual programming tests says a whole lot!
Resumés say very little about the applicant but seeing them in action through actual programming tests says a whole lot!
I tried the demo problem and solved it in six minutes (missed only the integer overflow issue). It seemed from the screenshots that time to solve wasn't very prominent when comparing applicants. Is this an issue?
I just tried the demo problem, site's great, hopefully I can use it next time we're hiring.
You know what? - We need something that's an amalgam of a dating site and a resume site; take a look at okcupid sometime - they have a mass of questions that users can add and answer that then provide a percentage match. Something like this between employers/recruiters and prospective employees could trim the dross, and help tailor the results for the specific job and individual. Sounds like a startup idea to me.
I agree this may be a good way of making money, but I'm not convinced that Q&A is a good way of matching applicants with employers. This is as much a Game Theory problem as it is a Classification problem. Candidates will write whatever they think employers want to read.
http://weblog.raganwald.com/2008/03/games-people-play.html
http://weblog.raganwald.com/2008/03/games-people-play.html
Then to top it off, when a hiring manager realizes that he doesn't have time to deal with hiring, he passes it off to a recruiting agency -- something that makes it even harder for me as a job seeker to help sell myself to the company. Am I the only one that hates the secrecy that recruiters use in describing their clients positions? They make it sound like I might possibly be working for the CIA or something with their lack of details on what matters (e.g. Company name so I can research their history and what they need) and they give me an exhaustive list that nobody could ever possibly have had experience with unless they had already worked at the company! No I don't have 10 years experience with $_OBSCURE_TESTING_PLATFORM, but give me a few hours and Google and I can figure it out. But how can I relay that via a resume? Especially one that will be filtered by either a computer or an HR drone that knows nothing other than buzzword bingo.
This is an area that is ripe for someone to come in and revolutionize the arena.