A few weeks ago I reviewed Peter Sims new book on experimental innovation, Little Bets. Since then I’ve received emails from friends asking me for concrete examples of businesses doing experimental innovation as well as why this is a better approach to innovation.
For those of you who haven’t read Little Bets or want to familiarize yourself with the concept, Tim Kastelle wrote a great write up about how little bets work.
With that said, internet companies are exemplary of taking a little bets approach. Because of the nature of the net, internet companies can do trial and error in real-time. And because of the size of their user base, they can get results very quickly.
Two of these companies are Google and Mozilla.
Google is probably the best known for their evolutionary approach to innovation. Their management philosophy is to drive evolutionary innovation through a combination of large mix of little bets and a small mix of big bets.
For example, Google Labs is a repository of ‘experimental ideas’ for new products and services which users can use before they are officially launched.
One of their most recent experiments is the +1 button you see on their search results. +1 is Google’s version of Facebook’s Like button, which will help Google further personalize your search results with the intent of using a users social network to present more relevant search results.
Google deliberately puts their experiments online and invites users to participate in these experiments because they understand that having such a huge user base enables them to see results and iterate very fast.
They know that most of these experiments will not work, but understand that most of them will end up ‘informing’ them about ideas that can then be used for other experiments. Tom Davenport did a thorough write up about how Google innovates for HBR a few years ago, you can listen to his interview below for more insights:
Mozilla takes the same approach. With such a huge user base using their Firefox browser, Mozilla releases experimental features in the form of extension which users can install on Firefox. The capability this extension brings combined with user feedback, it may end up becoming a permanent feature in the next Firefox release.
For example, experiments that ended up becoming permanent features on Firefox are Personas (be able to change Firefox ui) and Sync (be able to sync browser across multiple computers and devices). Both of which started as extensions for Firefox 3.5 to 3.6 and became a couple of the main features in Firefox 4.
An experimental mindset is crucial
As I said above, the open nature of the web makes evolutionary experimentation possible. Most, if not all, organizations are becoming knowledge intensive because the web is becoming a lot more pervasive in how they operate. Social networks are the most obvious drivers of this, and if looked through the lens of the internet, organizations are becoming huge intranets that don’t just include employees but also customers. That user base gives organizations the ability to conduct tests with users very fast and efficiently.
They key, is to bring these principles of testing and failing fast into your own environment. Having a high tolerance for failure, treating failure as a path to learning and using this knowledge to inform your thinking is crucial.
We could make the argument that most internet startups take a similar approach to developing their product or service. This approach is better known as agile, from software development, and embodied by the mantra: Fail often, fail fast, fail cheap.
Is this a recipe for better innovation?
I think so. It’s certainly a lot more intuitive and a lot more easier to understand and apply than the more analytical approaches. There are people who disagree with the Fail Fast mantra, arguing that it’s an intellectually lazy way to build businesses (again not very analytical). I think it holds true, but most big companies are driven by analysis. They don’t test their assumptions. If anything, they keep on confirming their assumptions.
Young startups are not driven by assumptions. They don’t have the time and resources to validate just about everything about an idea. Fail Fast is a way to test assumptions. To get out there quickly, test it with customers and see if whatever you have works and keep refining. Sure you want to have a financial model from the beginning or figure one out early but as we have seen, even that takes time.
The real challenge is in figuring out the business model.
What do you think about the Little Bets approach? What other examples would you add?