Google sowing seeds of own self-destruction

Posted in Technology trends at 10:34 am by ducky

Robert X. Cringely has an article where he says that Google will ultimately be killed off by employees who have left to start their own company.

While I don’t own any crystal balls, I don’t think Google is as doomed as he does. His reasoning is that the 20% time Googlers have to work on other projects will result in 4,000 new product ideas per year, of which 400 will be good, but only about ten of which can be turned into new products. He says that the people who had the other 390 will be bitter and go start their own companies.

I have to argue with some of the math.

  • Not all 20% projects are for new products. I bet most aren’t, in fact. If I were to start working at Google tomorrow, I would probably try to work on:
    • user-generated thematic maps
    • spam reduction
    • better contact management in Gmail
    • programmer productivity tools
    • a to-do list manager

    80% of the the projects on my list are either internal tools or add-ons to some existing product. It is way, way easier to think of a feature enhancement than a completely new product. I would be really surprised (and disappointed) if they aren’t already working on a to-do list manager, so my list probably has 0% new products.

  • Not everybody will be working alone. If, for example, I started a new to-do list manager, there is no way that I would be able to productize it all myself in 20% time. I would want to recruit others to help me on it. This means that there will be fewer ideas than people.
  • One of Google’s great strengths is its hardware infrastructure. Their 2006 financial statement showed $2.4 billion (yes, Billion with a B) worth of property and equipment assets. That gives the potential defectors a reason to stay: they have a whole lot more toys to play with if they stay (and a real disadvantage if they try to compete directly with Google).
  • I never worked on a 20% project, so I don’t know if they ever get canceled. I suspect that it’s very rare that you’d get told that you had to stop working on it. Thus if you really believed in something, you’d keep working on it as a 20% project because you were sure that if you just added a frobnitz, then the powers that be would see how incredibly cool it was, and would push it. Eventually, something else that was shiny would come along and you’d put aside your wonderful thing just for a bit… and your project would just wither away.
  • Working at Google is awfully pleasant. In addition to the food and stuff, you get to hang out with really nice, really smart people, and other people take care of nuisances like facilities, payroll, tech support, etc. You get to work on fun stuff that you want to do. Why would you ever leave?

While Cringely figures there will be 390 worthwhile projects per year that will get killed, I figure that the number of worthwhile new-product ideas will be less than 20: (3700 coders in FY 2006/ 3 people per team) * (1 new product / 10 projects) * (1 product that is worthwhile / 10 proposed products) = 12 worthwhile products.

In 2006, as near as I can tell, they launched nine new products: GCalendar, GDocuments, GSpreadsheets, GCheckout, SketchUp, GCo-op, and about three versions of existing products for mobile phones.

Only four of the things I mentioned were really new (vs. a port to phones) and came from in-house (vs. acquisitions). GCo-op doesn’t seem all that major to me, so really there were three major new in-house products: GCalendar, GSpreadsheets, and GCheckout. If my estimations are right, then that means that there were nine new products that got orphaned. Probably less than 10% of the people who have orphaned products will leave, so that means less than one project would leave. If that product required Google’s infrastructure, then chances would be even lower that it would escape.

The fact that two of the products that were released in 2006 (SketchUp and GDocuments) came from acquisitions says to me that Google doesn’t have enough new product ideas internally to keep up with the number they can release and support. I don’t know, but I suspect that I was being optimistic in my estimate of new products per 20% project. It’s probably much lower than 10%. This would mean that Google actually has quite a ways to go before they start losing people who are frustrated that their pet project got cancelled.

I expect that there are a non-zero number of people who will quit and start their own companies, but I think that will be because they see an opportunity in an area outside of Google’s business. They will decide to open a restaurant, or consult, or design video games, or set up a charter bus tour company. Some people will step off of the treadmill and raise kids, go into the ministry, or become forest rangers or professional snowboarders. While Google might miss those people, I don’t think that the professional snowboarders will be a threat to Google’s continued existence.

Comments are closed.