Why I can’t work for Google

Why I can’t work for Google by David Hardtke, Chief Scientist at Bright.com

Update 2/19/2013: This article was picked up by VentureBeat

The Times recently published an article describing how employers increasingly rely on internal referrals in hiring. In the age of “spray and pray” job seekers, where the cost of submitting an application is a few recycled electrons, this may seem like a rational solution in the war against resume spam. Trusted current employees can utilize their networks to bring in commensurate talent. Many companies are building technology to exploit this trend. Last week, for instance, Google integrated their career site with Google+.

When the herd starts doing one thing, the smart sheep sees an arbitrage opportunity. Hiring should follow rational economic principles; the goal is to hire the best workers as cheaply as possible. Finding candidates using unusual means is the way to score big in the labor arbitrage market.

Which brings me back to my story. I am now the Chief Scientist at Bright, leading the group that is refining the Bright Score technology (the Bright Score is a scientifically driven way to quantify the match between a job candidate and an open position). In 2007, however, I was a research scientist at UC Berkeley doing experimental nuclear/particle physics. I decided to make a career change and targeted two areas where my expertise in big data would seem a natural fit: hedge funds and web search. In retrospect, I am lucky I didn’t get a job at a hedge fund.

Since I decided to go into web search (information retrieval would be the technical term), I immediately applied to Google. To their credit, they dug my application out of the pile and arranged for phone, and then in-person, interviews. Along the way, however, the same question came up over and over: “Who do you know at Google?”

I didn’t get that job, but did get a job at a search startup (Surf Canyon) that builds enhanced search functionality on top of Google. We even published research showing a technique that could dramatically improve search results within Google search (why they don’t implement this idea is a great mystery, but that is another post).

Every year or so, a clever Google recruiter contacts me. I’d really like to work at Google at some point. As an IR (Information Retrieval) practitioner, working at Google is like a Muslim making the Hajj.

Every time they contact me, however, it is always the same question: “Who do you know at Google?” They beg me for names.

Here is the problem: I don’t know anyone at Google. They are remarkably thin in terms of former high energy nuclear physicists on their staff. At IR industry events, they are usually poorly represented—when I attended one of the premier research conferences in the field (CIKM), there were about 30 people there from Microsoft, but exactly one researcher there from Google. I tried to befriend some Google IR researchers I got paired with on the golf course, but they refused to even have a conversation about their work.

From my interaction with Google recruiters, I can tell that by not knowing anyone at Google, I am unfit to work at Google. This reminds me of the classic Economist article on teacher credentialing that showed a picture of Einstein with the caption, “Unfit to teach.” A hiring heuristic that gives massive preference to internal referrals can cause recruiters to throw away some of the best candidates.

In Google’s case, reading all 3 million unsolicited resumes is certainly not the solution either. Using technology like the Bright Score, (or TalentBin, Entelo, or LinkedIn) smart companies can mine the pool between “Monster.ugly” and internal referrals. It is surprising, however, that the company that revolutionized the way we find information is still basing their hiring decisions largely on internal word of mouth.

Good people that don’t have internal connections do apply for jobs. But finding those great candidates among the crowd is not easy. Reviewing every resume is difficult, so a common trend is to increase the importance of internal referrals. This is dangerous.

Ideally, a company finds a healthy balance between searching for the “needle in a haystack” candidate and relying on internal referrals. Companies that want to be the most creative and profitable need to recognize this and adopt practices that keep a healthy mixture in their workforce.

, , , ,

3 Comments on “Why I can’t work for Google”

  1. Allison Wilhelm (@AWilhelmSourcer)
    February 7, 2013 at 6:03 am #

    It’s a little odd that you’re being sourced, only to be turned down for not knowing anyone. Why use recruiters at all if they only want people in the Google network?

    But I do see the merit of referrals. If a hiring manager is trying to decide between two candidates, a referral can be a good tie breaker, since it means an already trusted employee can vouch for them, whereas no one really knows the non-referral.

    I’m curious as to what it’s like to work in their hiring department. But alas, I don’t know anyone at Google either.

  2. Veronica
    February 7, 2013 at 7:50 am #

    Awesome post, David. And I went to Maryland, so I rarely offer compliments to Duke grads Just kidding.
    Employee referrals are great, they are an awesome resource for recruiters, but they should in no way be the “end all, be all” of new additions! I think it’s awesome that you’ve joined Bright and I will follow your posts from now on to learn more about technology focused on finding the right “fit” for an open position.

  3. Frank Pacheco
    February 7, 2013 at 12:33 pm #

    You should come and apply at Microsoft, don’t forget that we have Bing!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: