Write Down Your Best Guesses: The Lean Change Canvas

The first step in discovery is to get all the hypotheses you have about your innovation written down in one place. The Lean Change Canvas in Figure 4.1, adapted from Alex Osterwalder’s book Business Model Generation, is a good tool for putting all your best guesses on one page. The nine boxes of the canvas represent the key elements of almost any kind of business or operating model. The outcome, a one-page summary of all the hypotheses to be tested, becomes a scorecard you’ll use until your innovation is proven and scaled. It is also an explicit, nine-part model of how your innovation will run.

Figure 4.1 Lean Change Canvas

I’ll cover each box in the canvas in some detail, but first here are some practical guidelines about how to use it:

1. Do it fast. Don’t give the canvas too much thought initially. Simply sit down and write out your vision, filling out as many boxes as you can in a fifteen- to twenty-minute session. You can do this with a group or by yourself, but in the spirit of experimentation it’s always more interesting to start with individual versions. Then, if you have partners or colleagues who generated their own versions, you can compare notes and learn up front how many of the newly recorded hypotheses are already shared by the team and what to do about those that aren’t. (Hint: Test them!)

2. Different boxes are more important at different times. You don’t need to have a great hypothesis for every box. You must start with a compelling value proposition and a problem or problems that need to be solved. From there you can fill in the boxes as appropriate given the stage of your effort and your data-gathering. A good guideline is that the hypotheses in the canvas will be tested from right to left, starting almost immediately with the hypotheses about the targets and gradually moving across the grid to things like which partners are essential and what the expense structure is likely to be.

3. Put all your ideas on one canvas to start with. You may have several ways to solve your chosen problem and/or several different types of targets requiring completely different channels. Start by putting them all in one place and then divide them up logically into separate canvases later. You can also use color coding to do this, coding each set of approaches with the same color.

4. Keep it short. The discipline of brevity forces your hypotheses to be clear and testable. If they are long and elaborate, use this as an opportunity to break them down into simpler sequences and start testing at the beginning. Watsi didn’t start by testing whether people would give to help strangers with medical expenses. They started first with whether they could deliver any money at all to strangers, using their own money and waiting to test whether others would contribute until later.

5. Work incrementally from the present. Another way to keep your hypotheses direct and to the point is to start with the here and the now. Ask yourself, “What is the immediate next thing that must be true for this model to work?” For example, rather than hypothesizing “Foundation X will fund our initiative,” try “Program officer Y will take a call to discuss our idea.” Remember, the canvas is a living document. As each hypothesis is confirmed or rejected, note that learning on the canvas and add the next set of critical hypotheses.

Finally, if you aren’t coming into this exercise with a very clear idea of how to solve the problem you’re trying to address, there are a number of approaches for generating operating models online at our website www.leanchange.net.