Monday, August 1, 2011

Business Model Analysis, Part 10: Getting Started

This post is part of a series on business model analysis for entrepreneurs. The first post in the series presents a comprehensive list (available as a downloadable PDF) of issues entrepreneurs should consider when designing a business model. Others posts in the series delve into specific issues; this one aims to provide some practical advice for entrepreneurs as they get started with business model analysis.
Why Bother?

Since the list in Part 1 of this series includes four dozen questions, entrepreneurs may see business model analysis as a daunting task and an unwelcome distraction at their venture’s outset. They may prefer to conduct a few interviews with prospective customers, then start building, in order to learn by doing in an improvised manner.

Eric Ries explains the dangers of this “Just Do It!” approach in his book, The Lean Startup, and I won’t repeat his arguments here. However, if the sheer length of the list of questions above is a deterrent, entrepreneurs should keep this in mind: due to serial interdependence between the questions—some cannot be considered until others are addressed first—it is not possible, necessary, or even desirable to answer all of the questions immediately and simultaneously. In the spirit of the lean startup, business model analysis is an iterative and ongoing process.

Most entrepreneurs begin the process with an insight about an unmet need and a potential solution for that problem. They explore the opportunity through customer discovery interviews, and use early feedback to refine their concept. Until their idea settles down, it makes no sense to push for deep understanding of pricing options, customer acquisition costs, working capital requirements, etc.

While entrepreneurs should avoid over-investing in detailed analysis of downstream topics, at some point early in the process of evaluating an opportunity, they should make a quick pass through all of the questions. Their goal should be to articulate hypotheses for as many questions as possible, and to gauge their team’s confidence that these hypotheses are on target. Back-of-the-envelope economics are adequate at this stage.

This quick but comprehensive scan is intended to ensure that the entrepreneur has not ignored any important business model elements. Likewise, the scan should surface potential “deal-breaker” issues early—in particular, any lack of internal consistency between business model elements—and to stimulate a search for ways to address them. For example, based on the initial scan a team might conclude: “We envision a complex new product that would probably best be sold through a face-to-face demonstration, but our ballpark unit economics suggest we cannot afford a direct sales force.”

Write It Down!

When conducting this first pass through the business model questions, entrepreneurs should write down their hypotheses, along with the key assumptions behind them. As their model evolves, entrepreneurs should adhere to this “write it down!” discipline. They should consider displaying a summary of their model as an information radiator—a prominent visual display that can be readily referenced and amended by team members—like the wall displays that many teams use to track the status of programming projects or progress with their product roadmap. For example, a wall-mounted white board could depict a business model “canvas,” following the approach suggested in Alex Osterwalder’s book, Business Model Generation. For some terrific examples of Osterwalder’s canvas in action, see the presentations done by Steve Blank’s students in his Lean LaunchPad course at Stanford.

Using the information radiator approach, members can affix different-colored Post-It notes in the cells of the canvas, showing the status of hypotheses for each business model question, for example: 1) green = the hypothesis has been validated through a decisive test; 2) yellow = a test that could validate the hypothesis has been identified, but not yet executed; 3) blue = a hypothesis has been advanced, but not a way to test it; and 4) pink = the question is important, it is too early to offer a hypothesis. As the team makes progress, pink notes will be replaced with blue ones, blue will be replaced with yellow, clusters of notes representing competing hypotheses about a single question will be pruned, etc. And, when the team makes a pivot—that is, when it changes its business model in significant ways due to a failure to validate a key hypothesis—large swaths of the board may revert from green notes back to yellow, blue, or pink ones.