Build big data adaptability through rapid experimentation
If you’re trying to waltz when your customers want to quickstep, they’ll choose your competition. This anxiety-inducing situation often points leaders in the direction of big data, which is a smart move. However, most organizations approach big data with too much trepidation — they’re dipping their toes in the water when they should be plunging into the deep end.
Adaptable organizations experiment rapidly with their offerings (products, services, and relationships) and build strong discovery capabilities. There’s no better way to accelerate discovery than to embrace big data in your strategy — the key is to attack this with purpose and vigor. To dramatically increase adaptability, you must build an organization that experiments in a big data fashion: with high volume, velocity, and variety.
Big data is classically hallmarked by high volume (exabytes), high velocity (real-time data streams), and high variety (video, audio, and unstructured content), and your approach to experimentation should be no different. To start, you should plan for a high volume of experiments running at any particular point in time. You should focus your experiments on finding the right product (or service)/market combinations; advanced companies are also experimenting with processes, value chains, and strategies (i.e., the strategy of no strategy). You need a high volume of experiments to hedge your bets; there’s no way to know what will click, so you always need experiments in your innovation funnel.
The best way to do this is to have multiple teams work on different ideas and assign everyone the responsibility for coming up with new ideas. Google is the paragon for innovation; employees are free to spend 20% of their time working on whatever they feel is a good idea. A product manager I know at Apple was assigned hundreds of resources to build whatever he thought would work. The combination of culture (i.e., everyone is generating ideas) and structure (i.e., multiple, parallel teams are working on experimentation) should be enough to generate a wealth of experiments. Now you need to move these experiments through the innovation funnel.
The traditional sales funnel inspired me to come up with the idea of an innovation funnel. You should have a slew of ideas that move through a pipeline to eventually become offerings, similar to the way prospects work through a sales funnel to eventually become customers. The difference between a sales funnel and an innovation funnel is that, with an innovation funnel, you have much more control over how fast ideas turn into products or services. In the spirit of big data, your experimentation process must be extremely fast. If it takes you 18 months to build a product that serves a fickle market, you have a huge problem.
