When discussing and coaching on analysis with teams (which of course means the general subject of analysis in the end) there seem to be two directions individuals take. First is the “analysis is hard” and the desire to run away from it, and the “single data point” and leaping to conclusions.
Two the first I answer “yes, it’s hard, but with practice it becomes easier.” The second, which is arguably a version of ‘running away’ from analysis, is much more difficult to work with and correct. Single Data Point analysis is different from “aha” analysis (see prior article) – the ‘aha’ moment is usually when people have, for a long time, concluded that something exists a certain way and suddenly they find it’s not true. Technically, this is, well, called “being false.” Science is littered with “aha” moments. Imagine with a supply chain process, a team gets together and begins a capture session. One team assumes that they are the only ones performing a vital process, say entering an order into an order management system, and there’s no way to streamline or improve the process. But lo and behold (Aha!) a second team is found to re-enters the order (or re-validates the order), very much a duplicate process, and very much an excellent target for combining and streamlining. Or capturing supply chain processes and finding warehouse A which feeds warehouse B which feeds warehouse C (Aha!) nobody knows why they have three warehouses! Everyone grasps that there is no reason for maintaining so much inventory, and we can streamline and consolidate. The Aha moment is when a hypothesis (either assumed – the warehouses were important, or explicit – there is no opportunity to streamline order management) is proven false. Proving the idea false unleashes teams to look at alternatives to a supply chain process which suddenly were previously forbidden.
For students of science, and logic, I refer them to the writings of Karl Popper of course, and the concept of ‘falsifiability’ as critical to types of reasoning process (science) and to proving universal ideas. Essentially, it’s very hard (perhaps impossible) to prove something is universally true, but you can unambiguously prove certain things are false. So with analysis, you like to create hypotheses or explanations of what you observe with a supply chain process, and then attempt to validate (i.e. falsify) it. I may say “80% of the cost is associated with only a few supply chain processes”. By creating a pareto of cost process-by-process in a system, if we’re lucky you’ll find that indeed, you can’t find the exception to the statement, and therefore you accept it as being a reasonable explanation. Until, of course, you look at how cost varies by time, and find that the statement is indeed false, but you create a ‘bigger’ pareto-type explanation which includes time, and now it generally explains the situation again.
The “single data point” teams don’t attempt to disprove statements, they attempt to prove statements, which creates the problem.