You could say:<p>"It is a funnel, you look at the stages, and at the end there is a great engineer that is usually working for a long period for the company.<p>Each level has expenses and conversions. Go an optimize keeping in mind the constraints."<p>But what I noticed is that is all about your company specifics and how you overcome the trade off...<p>For example the average new startup in unsexy industry has problems to get any body in the funnel. Getting somebody that know how to write some code can be a challenge... interview is more about persuading them to accept the job.<p>Great "sexy" companies have ton of people applying and would do everything they ask them to do. The problem for the company is to efficiently select those people.<p>When you have enough candidates, somebody may think: "We can save time and make accurate evaluations by giving a screening test offline to evaluate most of the relevant skills of the person!". Unfortunately the test takes on average 10 hours to finish.<p>Result: mediocre candidates desperate for a job try so hard to get everything done, but the great engineer has plenty of options and does not want to spend 10h on a test. In the end all great people that you would like to hire quit in the process.
I've suggested to recruiting firms that they should hire (or retain) a couple enterprise architects and have them meet with the technical managers looking for developers and let these highly experienced engineers pre-interview candidates based on the hiring company's needs and only refer people who can pass the enterprise arch's review. The time saved by the hiring company to not have their in-house archs, team leads, and engineers tied up interviewing candidates should MORE than make up for the premium such pre-interviewing would demand.<p>(Something else that would help is paying employees a lot more than $500 for making an in-house referral. I know of cases where locally where in-house referrals could have happened but by referring through a chosen recruiter got the referred a $1000 referral fee instead of his company's $500 referral fee... but that's a separate rant.)
Why use the cost of an engineer hour?<p>The correct metric is the opportunity cost of that hour - the lost productivity for that hour.<p>As an approximation take total revenue and divide by total engineering hours.<p>If you are hiring engineering because engineering is a bottleneck (not just hiring due to churn), then the metric is realistic. It would still be an underestimation in his calculation because he isn't including the hours taken to get new hire up to speed (oppositely assuming hire is due to churn).
Don’t forget the candidate’s time as well. You should only bring them onsite if you think (from phone’s screen etc) they’re probably worth hiring so if they don’t work out you should be improving the upstream screening for everyone’s benefit.
I don't think that time is very hidden. You have to sift through a ton of resumes and then spend hours and hours interviewing and phone screening. If you have actual work to do it's pure misery. I also don't have time to really think about my interviewing technique.<p>I would much prefer if we had a "hiring department" who are technically knowledgeable, talk to applicants, maybe give them tests and build up a pool of candidates who then can be interviewed by the actual engineering teams.
A bigger question about recruiting and engineer time, is whether they are effective at weeding out unproductive candidates. I'm not a recruiter, but I have interviewed a number of folks, and I still find it very difficult to determine if someone is good or bad. Folks that sound good during the interview, may just be good at communicating ideas, which is an important trait, but is still just one trait.<p>Even after you hire them, and you see them under-perform, it can be difficult to know if it's because of some intrinsic problem (e.g., just not so smart, or laziness), or if the problem is with the employer (e.g., not managing work, lack of communication, poor team integration). Yes, there are obvious stars and duds, but the middle is a vast gray area.<p>As engineers, we naturally try to turn hiring and performance assessment into an analytical questions that have clear answers, or that at least can follow a clear process. However, it still seems like voodoo.
Great article! Add to it the moral cost of interviewing multiple subpar candidates and improving the fit of candidates at the top of the interview process has a huge impact.<p>This is one of the reasons that many companies work with us at Insight (insightdatascience.com), as the interviews to join the program are done by Data Scientists and Engineers.
Sometimes interviewing a candidate as a team can have small benefits such as strengthening / making explicit what values are important to the team among the participants. This assumes that the team is operating in a full transparency, which is often seen in a small startup. I had the experience that after being eng. interviewer a few times I felt more glued and aligned with the team.
To be fair... $150 an eng hour? That's a 312k salary. I don't think that is accurate. And a recruiter makes 200k salary on average?<p>These numbers are a bit nuts.<p>Agree with a lot of what is written tho, but not this crazy number.