I have been involved in recruiting at trivago for two years now, and, during that time, I have read a lot of articles in order to improve and refine my skills in that area. Each article added a little bit of knowledge, some new thoughts, or some new set of questions to use in interviews. However, recently, I came across an article that made me think more and deeper than others, because it touched one of the core principles I thought a recruiting process should follow - and the author definitely has a point.
The principle I am talking about is the “One no means no” way of making a final hiring decision: If, say, four people interview a candidate, she will only get an offer if all four interviewers are in favour of this. If one of the four has doubts and says “no”, then no offer is made. Everybody has a veto right. In other words, the decision is made by committee.
Joel Spolsky has a variation to this rule in his Guerilla Guide to Interviewing, where he recommends that any two negative votes (or one negative vote by a senior developer) out of six interviewers in total cause the candidate to be turned down. In both cases, however, the decision is made by voting, and a positive decision needs a far greater share of the votes than a negative decision.
False positives are bad
The reasoning behind this method is that false positives - hires who turn out not to be a good addition to the team - are worse than false negatives - candidates you turn down, who would have been great. Joel puts it like this:
“… it is much, much better to reject a good candidate than to accept a bad candidate. A bad candidate will cost a lot of money and effort and waste other people’s time fixing all their bugs. Firing someone you hired by mistake can take months and be nightmarishly difficult, especially if they decide to be litigious about it. In some situations it may be completely impossible to fire anyone. Bad employees demoralize the good employees.”
Thankfully, I have never faced the legal issues related to hiring or firing that he mentions, but I might add two more points:
- It costs time and energy to train a new colleague and bring her up to speed, so you naturally want to be sure that she is worth the investment.
- If a new hire does not fit the team personality-wise, he can cause the team atmosphere to change for the worse. Even if he is good technically, this can ruin the team’s focus and productivity.
False negatives are worse?
The principle of voting on whether to make an offer or not has been entirely natural to me for some time now, but has recently been questioned by a brilliant article by Henry Ward. While some of his thoughts seem targeted specifically at hiring in startups, most of it is relevant to non-startup companies, as well.
First of all, Henry has a completely opposite opinion when it comes to false positives vs. false negatives:
“We should not be afraid of False Positives. We can quickly fix a False Positive hiring decision. However, we should be afraid of False Negatives. We can never fix a False Negative mistake. And the cost is unknown and uncapped. Facebook passed on Brian Acton (WhatsApp cofounder) and it cost $8B and a board seat.”
While the example that he picks is quite dramatic, he correctly points out the nasty things about false negatives:
- In most cases of false negatives, you will never know that the candidate would actually have been great. He will move on, possibly to a different city or a different country, and you will never hear from him again.
- Because you do not know about the false negative, you cannot fix it.
- Maybe worse: Because you do not know about the false negative, you cannot even learn from your mistake.
The last point is what bothers me most, personally. You make the same mistake of sending great developers away over and over again, and you cannot help it. And because people are all unique, you cannot even hire a “similar” person five weeks later, and draw conclusions about the guy you passed on.
Henry must have had a similar feeling, and, because he wanted to avoid false negatives more than false positives, he designed the recruiting process at eShares in a different way. Instead of voting in a committee, one single person - the hiring manager - makes the decision on his own. This sounds like dictatorship, but of course that’s not the whole story yet. Before the decision is made, every interviewer gives her opinion and final verdict. The difference to “decision by committee” is that a weak “no” might easily be overruled by a passionate “yes”.
For example, if one interviewer is a bit disappointed with the candidate’s ok-but-not-great SQL knowledge, but another one is absolutely thrilled with her passion for quality and thoroughness, then the hiring manager will probably still say “Yes” in the end (see here for an explanation on why you should hire for talent, not for knowledge or skills). In fact, Henry claims that some of their most important people, who made all the difference in the world to the company, had at least one person saying “no” to them after the interview. Some even had the majority of interviewers against them.
Things can go wrong
Having been in doubt multiple times myself after rejecting a candidate, I have no trouble believing this. People are complex, and there are millions of things that can affect an interview, or an interviewer’s impression. Maybe the candidate was nervous and it showed during some problem solving exercises. Maybe one of the interviewers is extremely fond of Haskell, and the candidate unknowingly made a bad Haskell joke. Maybe the candidate, while great at software architecture, just cannot do live coding in an interview, because the situation freaks him out. Maybe the candidate just sucks at highlighting her strengths, is very modest, or shy, and some interviewers misinterpreted that. Even if you are a great interviewer, you will not achieve 100% certainty about your decision. I have witnessed several people develop in a stunningly great way after I was very close to rejecting them after the personal interview. I am sure glad I didn’t.
Ultimately, the two approaches - hire by committee vs., let’s call it “benevolent dictatorship” - are driven by different mindsets. Hire by committee is rather defensive, while benevolent dictatorship is, colloquially speaking, the YOLO way of hiring. Hire by committee is geared towards risk minimization, while benevolent dictatorship aims at maximizing impact. Henry claims that you will lose many 20x employees by consensus hiring, because they tend to be controversial.
Personally, I do not know if that’s true, but I have been in situations where I secretly wished we would not vote, but go by the strongest conviction. I stood back, because I believed that every vote should count equally. Maybe I should talk this over with my fellow interviewers, because I feel that breaking the consensus rule, at least once in a while, is okay.