Direct observation remains one of the most effective lean skills for leaders to master. I have written about its proper use in "Going to the Gemba." Like problem solving, there is a difference between the skill and a management system for using it.
There are four distinct levels of observation, each with a degree of abstraction from the truth. A management system of observation determines the pattern with which these different levels are utilized, and for what purpose. Determining the right level of observation should not be done on autopilot; it requires purposeful decisions.
The less abstract our level of observation, the more we have to invest our time and energy. As a result, getting to the ground truth for every decision and insight would be all-consuming.
1. Stories and anecdotes. This is the most abstract level of observation, yet we rely on it heavily for understanding the current state and making decisions. Stories and anecdotes come in meetings, emails, and hallways. But they are heavily filtered by the person, or persons, telling the story. They actually tell us more about the person and their view than about the truth.
While this makes storytelling a weak tool for understanding the current state, it does have a purpose. First, it is a powerful tool to get our view across, when the point being made is more important than the facts behind it. Second, it is a great insight into the storyteller's viewpoint.
We must be careful about making a critical decision when all we have is storytelling and we can't be sure of the personal filters through which the story passed. As an example, a university criminology professor conducted an experiment. During the class, a fake mugging was staged. The students were asked to immediately write down a witness statement. They were generally less than 50% correct. This exercise serves as a warning about the accuracy and effectiveness of storytelling.
2. Data and graphs. Data and graphs are treated in some circles as the top of the heap. But they greatly depend on the assumptions made when building data gathering and classification. Data is fundamentally about results. It's very useful to help us understand if we have a problem, and if used well, where we have a problem. But it's weaker when seeking the why. It is most useful in a stable environment, where using data for comparison from one period or experience to the next reveals useful insights.
3. Pictures and diagrams. Pictures and diagrams should be used purposefully. The inherent goodness of these techniques is that we are working in the realm of process. Once we start drawing a picture, or building a map, we are more likely to be engaging the process.
When we draw activities, connections and flows, we are already closer to the root cause of why we are getting the current results. Process generates results, so if we want to understand the why, we have to work in the domain of process. Pictures are meant to be a simplification of what is really going on, so they still have a degree of abstraction.
4. Direct observation. This is the least abstract form of understanding the current state. It is going to the direct source, with as little filter as possible, and no assumptions at all. This is how you understand how things are really working. We don't care how things are supposed to work; we only care how they actually work.
Understanding the pros and cons of these different levels helps us make good decisions about which level to utilize. For example, if you have a highly aligned lens for observation with another person as a result of many hours of practice, then storytelling may be very effective because their eyes are your eyes. Charts are great when conditions are stable, but when the context changes, the charts cease to have as much value. Most importantly, when charts and data do not give us the right story, we have to go and see for ourselves.
Jamie Flinchbaugh is co-founder and partner of the Lean Learning Center in Novi, Mich., and the co-author of "The Hitchhiker's Guide to Lean: Lessons from the Road."
