Intro: The Lean Experiment Canvas

Lean Experiment Canvas

*http://bit.ly/LeanExpCanvas

The Lean Experiment Canvas (LEC) will help you run better business experiments with better results. The cornerstone of the lean startup movement and Toyota’s lean transformation is the ability to create a learning organization and cycle through PDSA cycles quickly.

Your goal is to uncover truths about your business and customers and it all starts with great questions and a falsifiable hypothesis. Don’t leave your learning up to chance and hindsight bias. Solidify your assumptions, risks, and metrics before your experiment. Optimize collaboration and share what you have learned.

The LEC is made up of 12 individual blocks which collectively take you systematically through the scientific method. For the astute reader, you will realize the LEC is a manifestation of the OODA loop developed by John Boyd (more to come later on another post).

It is hard to say which block is the most important since they are mutually reinforcing but as you proceed to fill in the information there is a tendency for a natural pathway to develop.

LEC Pathway

The 12 Blocks

1) Background

2) The Question

3) Falsifiable Hypothesis

4) Prediction

5) Assumptions

6) Risks

7) Key Metrics

8) Key Stakeholders

9) Experimental Steps

10) Timeframe

11) Results of Experiments

12) Next Obstacles

1) Background:

The background information is essentially the diagnosis and reason for action of the experiment. This information contains the “as-is” status of the current operating environment. The background information is used to understand the greater business context and need for an experiment. After reading the background one should be able ascertain why an experiment would be necessary.

2) The Question:

The question is the spring board which will feed your hypothesis and is the broad question of interest you want to study. In science this question can reach finer levels of granularity. In business, the granularity reached in science is difficult to achieve. The more specific your question the less variables that are in play which will get you closer to cause and effect.

3) Falsifiable Hypothesis:

This is the cornerstone of a scientific experiment. A good hypothesis is falsifiable, meaning it can be proven false. An example of a subject that isn’t falsifiable is astrology. Can any experiment prove astrology to be false? A good hypothesis follows an “if….. then….” format with explicit cause and effect.

4) Prediction:

This is a more explicit prediction of what you think will happen. In your hypothesis you will create a prediction in the “then….” portion of the hypothesis. In this block you want more detail. Predict how you think your measures will be influenced. Predict how your colleagues and co-workers will react. The more you predict the more you will learn, especially when you results are contrary to your prediction. Never say, “let’s just see what happens.”

5) Assumptions:

It is hard to say if there is a block more important than the assumptions block. This is the bedrock of failure for most improvement efforts and team dynamics. The assumptions are shaped by the mental models we have created. Among team members mental models can be vastly different. Filling out the assumptions block attempts to solidify those models which can be shared, understood, and worked through. You want to expose your assumptions to the real world. The worst scenario is to hold assumptions that are false and you not to know.

6) Risks:

The risk block allows you to think through worst case scenarios. Where as assumptions are internal view of how things will occur, risks are the things that could happen which can derail your experiment. Being aware of the risks will allow you to deploy stop-gap measures to prevent the probability of their occurrence.

7) Key Metrics:

These are the most important metrics or outcomes you are trying to measure. It is very important to have a system of measurement set-up before running the experiment. Step through and understand the different types of metrics (process, balancing, and outcome) and what you need to identify to accelerate learning.

8) Key Stakeholders:

This block is straightforward. Who will be affected by your experiment? Think about those who will be directly affected as well as those who will feel the effects downstream.

9) Experimental Steps:

These are the explicit steps you need to take to prepare for and run the experiment.

10) Timeframe:

How long do you expect the preparation to take before running the experiment as well as how long will you run the experiment?

11) Results of Experiment:

When capturing the results of the experiment make sure to compare them to your assumptions, risks, and predictions. What you are most interested in is any assumptions that have be invalidated. These anomalies are massive opportunities for learning. Cherish them.

12) Next Obstacles:

As you are running the experiment you will inevitably find new constraints or problems you weren’t planning on initially. You will want to capture and feed these issues forward into the next experiment. The next obstacles section feeds into the next experiment which begins the LEC stacking and the compounding of information.

Below is an example of a completed LEC. Note most LECs should be completed by hand and in pencil.

Completed LEC

See SlideShare for more details: http://bit.ly/LeanExpCanvas