Innovation: Established Organizations & Startups - Learning Process

In my earlier post I mentioned that Govindarajan & Trimble describe a rigorous learning process.

The effect of a rigorous learning process is that estimates become progressively more accurate over time. They illustrate it with a diagram. The diagram is precisely what softwarewallahs call "The Estimation Funnel"! (See my post on Estimation).

Estimates are based on hypotheses about cause and effect; they are of the form if < action.> then &lt;result>. Hypotheses, as in the scientific method, must be tested by experimentation. Evaluation of results must be free of biases (the authors describe these biases). Adhering to the sceintific method is what makes the learning process rigorous.

The genesis of all startups too are a set of hypotheses. These concern pain points, market size, value provided, revenue streams, costs, ... All of that goes into the business plan. The business plan is a set of hypotheses about business reality. It is not business reality. The rigorous learning method is required to evolve the business plan to conform more closely to reality.

The authors provide a simple graphical tool for recording the hypotheses. They call it a "hypothesis of record". They illustrate it with an example of how Analog Devices actually put it to use.

Further, the acts of making the business/innovation plan and rigorous learning cover, in my opinion, the first three steps of the Shewhart PDCA cycle. Not surprising, as continuous quality improvement too is based on hypotheses.