Some thoughts on innovation from Google’s Astro Teller

I have been in Silicon Valley this week for our annual DFJ partners get together.  There are eighteen funds in the DFJ Network and once a year we all congregate in Half Moon Bay for a conference, to which our CEOs and LPs are also invited.  As always there were good speakers and chances to meet with folk from companies like Facebook, Google, Amazon and Twitter – you might have seen some of my Tweets earlier in the week with some data that Amazon shared.

Perhaps my favourite speaker was Astro Teller, now at Google carrying the title of Director of New Projects where he focuses on keeping the company innovative.  Prior to joining Google a few months back Astro has founded a number of very interesting and successful companies, including AI based hedge fund Cerebellum Capital and Body Media.

Astro gave the dinner speech on Tuesday night and since then I’ve been meaning to blog a couple of his ideas on how to keep a company innovating.

First a definition – innovation is about step changes, not incremental improvements.  The ideas below are good if you want to deliver step changes in a company’s direction/performance.  And they apply as much to startups as big companies.  As Astro was talking I ran through our portfolio in my mind and I came up with a number of companies that are successful with their first product and as they have focused on building sales they have stopped innovating.  That isn’t usually a problem in the short term, and if you are growing revenues you have bought yourself some time to fix this issue, but radical product extensions or new product lines are necessary for all companies at some point.

So, to Astro’s tips – and I particularly like these because they are practical:

  • Focus only on truly great ideas – everyone understands that killing off unpromising projects early is an essential discipline, but to get a company really focused on great innovations it is important to get behind only the very best projects – and that means only keeping the great projects and killing projects that are merely ‘good’ just as you kill the bad ones.
  • How do you know a great project? – well partly you just know one when you see one (like good jazz and pornography in Astro’s words), but a good test is whether you would be really upset if your competitor launched the same innovation
  • When someone comes with an idea for a project, insist that they also show you nine other ideas they had and discarded.  Partly this is because people shouldn’t just be coming with the first idea they think of, but also because forcing people with one good idea to go through a ‘I’d better think up nine other dumb ideas for my boss’ process, will generate a lot of creativity.  One of the ‘dumb’ ideas might turn out to be the best one.
  • Similarly – insist on seeing three ways of doing a prototype (ten would be too many, given the extra work involved)
  • With all these ideas floating around and a need to only focus on the great ones, killing projects early becomes an organisational necessity – so the company should reward ending projects
  • When agreeing targets with project leaders make the objective to hit XYZ milestone by ABC date or kill the project – do it in writing and make it clear that ending the project will be looked on as favourably as hitting the milestone
  • The drive to innovate needs to come from the CEO – if she doesn’t buy into the requirement nothing will happen
  • Make someone responsible for innovation – could be a product person
  • Insist that every idea for an innovation is accompanied by a story which explains how it will help the customer
  • Task someone with deleting code – deleting code without breaking the system requires a full understanding of the system

I’m going to finish this post with Astro’s close, which was to say that some people are natural innovators, and that teaching people to innovate is a bit like trying to teach them to be funny.  You can’t make people funny, but taking the funniest people you’ve got and putting them in funny situations will reap dividends.  So it is with innovation – and the pointers above are some ways to create an environment in which innovation flourishes.