The day Google stopped hiring non-engineers...

... or how mid 2000s era hiring at Google was pretty wild.

Pretty much my second week at Google, my manager, Bill Coughran, asked me to start attending the hiring committee that hired me. It was *quite* the introduction to the hiring process at the then 2000 person Google¹. I would go on to attend 2 or 3 a week for the next many years. I estimate those committees I was in reviewed some 10,000 engineering hires, and extended offers to maybe a quarter of them.

The process back then was extremely, some might say exhaustingly², detailed. The committees would sometimes consist of up to 13 senior engineers. We would go through 6 to 10 candidate ‘packets’ in an hour. You were expected to have done the reading ahead of the meeting, too.

The packet, which was a 20 or 30 page collection of the candidate’s CV/Resume, their phone screens³, and the onsite interviews they would have completed. In 2004, these numbered from 8 to 13 interviews. (We would pull back to a maximum of 8 in later years, when we found that the additional interviews didn’t change the hiring decision significantly enough to justify the extra time taken by the candidate, the interviewers or the committees.) 

As a result the packets would sometimes balloon to 30 or more pages. So for a given meeting it wasn’t unusual to pre-read 250 pages+ of candidate information. And if you were in three committees as I was, you would often be reading over 1000 pages a week. And remember we wanted to do it well! To do otherwise was , well, pretty disrespectful of the candidate, the committee and your colleagues who had done all those interviews. 

Keep in mind that you wouldn’t want to hire badly because, even back then, Google was pretty bad at firing people. We had what I’m convinced was a lot of bad no-hire/false negative decisions back then, simply because Google was so bad at performance management⁴. 

Back then, I only worked on engineering hires. I reviewed software engineers (SWEs), platforms/hardware engineering, network, and the rest. I didn’t work on the site reliability (SRE) hiring committee; I’d sometimes see SREs as overflow or for other reasons. I would sometimes see technical program managers, program managers and other oddball roles that fit in engineering as well. 

I would not work on Sales or G&A hiring at all. I’d sometimes refer people into those roles, but I was never in their committees. I’d sometimes do candidate interviews for our legal and policy teams, but that was about it. 

This is all to set the stage for a very unusual committee in 2005⁵ or so. Our HR/recruiters we worked with in these committees were unusually ..unusual⁶ that day. We were being presented with more candidates than normal and were talking about extending our meeting to 1.5 or even 2 hours. Imagine your brain is already tapioca from 1000 pages of candidate reading a week and then looking down the barrel of a doubling. 

We asked what changed, and maybe we can reduce the number of interviews, etc. And one of the HR folks told us “Well, Larry put the kibosh on any further hires until Eng has the majority again.” It seems that G&A and Sales had outstripped engineering hires and Larry refused to sign off on any new hires until Eng/SRE was in the majority again. 

Back then , every engineer we approved would go to the Bill Coughran level, then to LSE⁷ for approval. I seem to remember that back then it was usually Larry Page doing the final review, rather than Eric or Sergey. 

Anyhow, for our part, we reduced the interview load⁸ and pushed back on re-interviewing⁹. We started a number of new committees by cleaving off committees from the ones we were in, etc. I can’t imagine any of this was fun for the recruiting staff. I will say that it was refreshing to see what was important to the engineering focused Google that I really liked working for. 

I was reminded of this when watching the always entertaining Rory Sutherland speak about “how pen-pushers destroyed the office”. I recommend watching that if you find yourself wondering if you actually manage employees, or if HR/Legal does, and if you actually report to your manager or if you really report to a line in a spreadsheet held in HR, at Bain, or at McKinsey. It’s pretty interesting. 

1. I would later do the same to Fitz, who has maybe forgiven me for it at this point?

2. And they’d be right.

3. And for fresh-outs, their grades. It was nuts. Fresh-out PhDs we would even have their dissertation for us to ‘casually peruse’. I assure you we were not sassy mean people about your thesis though. In some ways I think you were better off being ABD than having your doctorate.

4. Google was vulnerable to charming scoundrels, too, but that’s a whole different essay

5. Honestly, it could have been before or later… it was before we spun up the Android committee, I think?

6. Angsty? Anxious? Aggressive? I really don’t know how to express the emotion on display that day so long ago….

7. Larry, Sergey and Eric, obvs.

8. This was ahead of the Laszlo Bock’s inspired analysis that proved what we felt about interview quantity and diminishing returns after 8 interviews. He was a great guy to work with.

9. Yes, sometimes the interviews weren’t…confidence inspiring and we’d ask for a re-do with a more experienced interviewer with some guidance on what we wanted to know about.