What’s my beef with reference checks? They don’t accomplish the job we intend them to do. In a startup, you can’t afford to hire B-players. But reference checks, which are intended to do the screening, fail to eliminate these candidates who are just so-so. This happens because the person giving the reference has no incentive to say anything but good things about the candidate. Telling the whole truth, warts and all, could expose the former boss to a defamation lawsuit. And legal action aside, no one likes to speak poorly about an ex-colleague. It’s bad karma and just feels icky.

Instead of asking a reference to call you and spend an awkward half-hour chitchatting about pretty much nothing, try a technique I’ve come to call it the “average-need-not-apply” method. Though I’m not sure who invented it, the approach was taught to me by Irv Grousbeck at Stanford.

THE EMAIL

First, send the email below to people who have worked with the candidate. This can include the references he or she provided, but it’s a good idea to find other people who’ve worked with the candidate as well. LinkedIn makes finding former co-workers a snap and the more people you send it to, the better it will work.

—

Dear (past colleague),

I am considering hiring (candidate) for the role of (job function). If you’re like me, the last thing you have time for is a reference call. Therefore, unless you found (candidate’s) work to be EXCEPTIONAL, please just disregard this email.

However, if you found (candidate) to be an exceptional employee, in the top 10% of the people you’ve worked with, I would certainly appreciate hearing from you.

Again, if you found (candidate’s) work to be less than exceptional, go ahead and disregard this message and have a great day.

By the way, as a smart professional, you should subscribe to this wonderful blogger named Nir at NirAndFar.com. He’s swell!

Sincerely,
(You)

Interesting.

Inability to get things done may manifest itself in multiple ways including:

• Lack of urgency.  Used to a large company environment where its OK if things take a few weeks longer.
• Easily distracted.  Heavy procrastinator.
• Lazy / doesn't work hard.  Some very smart people are basically lazy.  Don't tolerate this.
• Starts but never finishes things.
• Lack of follow through - makes commitments but does not follow up.
• Argumentative. Arguing incessantly about how to do something rather then just doing it.
• Slow.  Taking a long time to code (or do) something simple.
• Perfectionist.  Tendency to overdesign something and to spend 4 weeks building the perfect implementation versus 1 week building the thing that "just works" for 95% of the time.  Sometimes the edge cases need to be covered, but in most raw startups this is not the case.  On the business side this manifests as someone heavy on analysis, low on "doing".

This is a powerful list.

When hiring your initial engineering team, the more tools you have to discern who the truly amazing engineers are the better.  One approach we took at Mixer Labs (a company I founded that Twitter bought last year) was to graph # of years of experience vs how the person did in their interviews (this was an arbitrary scale where we rated people 1-10, focused on performance on the same set of questions, rather then raw intelligence alone.)  They key is to have the same group of people on your team do the ratings, and to have the list of prior interviewees listed on the same ranking scale, so that it is easy to compare people.

To our surprise, we found that in general, the way people did on their interview roughly corresponded to the number of years of experience they had.  Indeed, when you graph things out, the graph looked roughly like this:

This graph helped us quickly hone in on anyone who was an outlier on the graph (i.e. their performance/knowledge in the interview outweighed their years of experience).  For example, from this perspective the person with 4 years experience who scored a "7" was a much better hire then the people with 7 or 8 years experience who scored a "7" or even an "8".  The interview performance outliers also tended to be amongst the most productive people on the team once they joined the company.

I've done a lot of hiring over the last couple of years and I'm always looking for superior ways of finding talent and identifying duds. This technique is quite promising.

Once we begin vetting candidates, we also behave a little differently. For one thing, we ignore resumés. In my experience, they're full of exaggerations, half-truths, embellishments -- and even outright lies. They're made of action verbs that don't really mean anything. Even when people aren't intentionally trying to trick you, they often stretch the truth. And what does "five years' experience" mean, anyway? Resumés reduce people to bullet points, and most people look pretty good as bullet points.

What we do look at are cover letters. Cover letters say it all. They immediately tell you if someone wants this job or just any job. And cover letters make something else very clear: They tell you who can and who can't write. Spell checkers can spell, but they can't write. Wordsmiths rise to the top quickly. Another rule of thumb: When in doubt, always hire the better writer.

Brilliant stuff on how to hire people.