Bill’s Blog

This past week I attended the Society of Information Management’s monthly get together at the invitation of my friend Aaron.  The featured speakers were a couple of guys from Google, Brian Kemler and Steve Benson.  Their topic was innovation.  I found the discussion interesting and took a few notes.  Read on if you’re interested.

- True innovation means both technical innovaton and business model innovation.  They said Page Rank was an example of technology innovation while Text Ads on search pages was business model innovation.

- 50% of the employees at Google are engineers.  From what they said I could also add “and doing engineering work”.  I don’t know what the percentage is elsewhere, but I thought this was a great formula for success.

- When they talked about resources, they talked about people and machines.  I found it interesting that they didn’t talk about money.

- Google uses a 70-20-10 model for where they focus.  70% of their resources focus on core businesses like search, ads, and apps.  20% focus on adjencies or related businesses, and 10% of their resources are focused on exploratory type work.

- Flat management structure.  8 layers from bottom to top. (this is flat?)

- Hire for culture fit (one person I spoke with later in the week said he had 24 interviews when he was hired.  I also learned that Larry Page still reviews every single hire before an offer is made.)

- Dedicate time for inovation.  All engineers can spend 20% of their time on anything they want.  Though this guy did say that that 20% time usually means 120% time meaning that a lot of the innovation work is done after hours.  I asked him how they manage the 20% time.  I was met with a blank stare.  He eventually said that they hire smart, passionate people who want to work on interesting things, so the 20% wasn’t a problem.

- Experiment and realize that some things won’t work.  He gave an example of the electrical substation in The Dalles data center.  The power company was going to take a couple years to build it so Google decided to do it themselves.

- Encourage people to act on new ideas

- Big Award culture.  Their founders awards are in the millions (wow!).  They have Executive awards in the several hundred thousand dollar range.  Smaller awards are ~$1500.  They also have spot awards that anyone can give.  These are $175 or so.  One thing he said about the big awards was that they provided good incentive for people to not leave and work for Facebook or Twitter.

- locate engineering centers near the International Colleagic Award Contest Winners (or something like that – think very smart people)

- managers have up to 50 people reporting to them.  They don’t have staff meetings, but instead do weekly snippets.  They are stored in an internal, public db (as is everything at Google).

- TGIF meetings Friday at 4.  The execs answer employee questions.  When the company got too big for this to be done in person, they started a system where employee questions are submitted electronically and voted on.  They most popular wind up at the top of the list.  He said that isn’t always comfortable, as some of the questions can be difficult to answer.

- iterative design, constant improvement.  They release early and release often.  They must have metioned cloud computing and its advantages 100 times during the presentation.  This was one of them.  Cloud computing allows them to add or remove features anytime they want.

- crazy “org” chart.  not top down, expect people to do the right thing.  don’t dictate technology.  can use macs or pcs, ppt or google docs.

- Internally they use search for everything.  They don’t bother to organize information.

- External sites all run gubuntu, GFS, and Big Table (their DB).  This horizonal platform, purpose built hw, and their massive scale is used by all google apps.  Also, there are UI guidelines but no hard and fast rules.

- They talked about customers not being locked in.  They felt they had to earn people’s trust and their business.  They view competition as a good thing.

- Internal tool called Ideas.  New ideas are entered into the system.  Peopl vote on their favorites.  Engineers work on the top ideas.  Many thigns at google seem to work this way.

- budget time not money

- “cool” projects get the engineers

- when they kill projects they kep them alive for the existing users

- they turn features off if users don’t use them

Since you’ve read this far, I’ll give you a bonus.  Later in the week I had lunch at the Google campus in Mountain View.  I was in a small group meeting with Ivan Ernest, the Head of Global HR, Engineering and Operations. There were a few additional things he shared.

- you can put yourself in for a promotion anytime you want, even if your boss doesn’t agree with it.  They build a “package” of peer reviews of your work and have peers who don’t know you make the decision.

- new hires are often brought in as “MTS” – Members of Technical Staff.  After a year or so their peers provide reviews about them.  Then peers who don’t know them decide what level they should be.  It’s a strange culture for sure, though it seems to be working for them.

- Seems like many many things are voted on, or managed by peer reviews.  An example are their quarterly objectives and key results.  They are done bottoms up with very little tops down input.  Ivan said that Google’s approach was to hire the smartest people they could and then ask them what they should be doing.

A final note.  I had lunch in the Google cafeteria.  It is free for everyone, including me.  The food was great, there was plenty of variety, but there was a certain sense of disorganization to it all.  It was a wonderful trip.