"Plan B"

We met with Randy Komissar again from Kleiner Perkins. He talked to us about a framework that startups should use to make a lot of progress quickly. His working title for the framework (and book that he is writing) is called "Plan B."

The process works like this: the founding team should identify all relevant analogs and "anti-logs" for the new opportunity that is being explored. Analogs are examples of successful companies--not necessarily from the exact same space, but relevant enough that we can glean lessons applicable to our opportunity. The "anti-logs" are companies that were not successful. We want to draw upon the experience of others before us to identify the "knowns" regarding the given opportunity.

Then we identify the unknowns, upon which we wish to take a "leap of faith." These leaps of faith are pivotal for the start-up, and they are where the company should be focusing all of their efforts.

To address the leaps of faith, the startup should follow a 5-step iterative process.
1. Identify what the key questions are that need to be answered regarding the leap of faith.
2. Develop hypotheses regarding each of the key questions.
3. The company should go out and do testing to validate or invalidate the hypotheses.
4. Interpret the data from the testing to generate insights.
5. Refine the original hypotheses based on the insights from the testing.

Iterate steps 1-5 until you have resolved most of the key questions/leaps of faith. Once you have done that, you will have backed into a business plan which you can then execute.

Seems like a pretty simple, straightforward process? The challenge in executing this process well lies in the judgment that needs to be applied at each step.

• What analogs/anti-logs do you select for comparison?
• What are the most important leaps of faith? What are the key questions that need to be answered?
• What are the hypotheses for each key question?
• How do you create and execute the tests to validate the hypotheses? How do you run the tests as quickly and inexpensively as possible?
• How do you interpret the results, and what insights do you draw?
• When do you refine your hypotheses, vs. when do you throw out your test results and try again?

The most successful companies, according to Randy, can get through the process quickly because (1) their hypotheses are usually correct, so they don't have to spend a lot of time and energy iterating their hypotheses; (2) the hypotheses that are wrong fail early, fail often, and fail quickly. So you have to have the experience and judgment to develop hypotheses that are mostly right, and/or you have to be lightning fast with your iterative testing and have your hypotheses fail quickly up front.

Randy then took us through the example of Steve Jobs and the iPod. Jobs had a few different analogs: (1) the Walkman (people were willing to listen to music on headphones in public places), (2) Napster (people were willing to download and share digital music), (3) VCRs (media industry had to settle for "fair use"). He also had a couple of anti-logs: (1) The Rio (a poorly designed MP3 player), (2) Napster (got sued by RIAA because the record labels felt they encouraged pirating). He knew that people would listen to music on-the-go, and that they craved digital music. He also knew that he could design a much better user experience than the Rio, and he could create an ecosystem that would be friendly to the record labels. His biggest leap of faith--would people be willing to pay for digital music? Jobs didn't believe that they would. So what did he do? He hedged his bet. He decided that he wouldn't make money off of music, but off of the hardware. He created a "fair use" case by charging money for the legitimate music, but he also enabled users to download pirated music onto the device as well. The result? Only 3% of music on iPods was actually purchased from iTunes. But the record companies weren't able to sue Apple, and in fact, they cooperated with them.