I suggest reading this excellent paper on Improving Problem Solving by Ian Bradbury and Gipsie Ranney. As they note, problem solving is not a substitute for innovation and improvement as solving a problem solving only returns you to the status quo. The paper also considers the relationship between problem solving and system improvement.
Their paper provides an excellent overview of a very common problem: only looking at defects. The paper discusses the vivid example of the NASA O-ring failure where focusing on all the data may well have avoided the tragedy that befell the Challenger shuttle.
If one only considers the “problem” category of results from a system, one may either miss an important causal relationship (as in the Challenger case) or erroneously infer presence of a causal relationship. To avoid these pitfalls, it is advisable to consider all events or outcomes… The same process that delivered the defects delivered outcome or results that were not defective. Study of the defects alone presumes that the defects are necessarily produced by a special cause. Such a presumption may frequently be invalid.
Fully appreciating this quote and incorporating it into your practice is extremely important. In my experience however, many people don’t do this and certainly very few people do until they have many many years of experience applying Deming’s ideas. Reviewing this excerpt, and their full paper, every 3 months, until you have it ingrained in your problem solving thinking, will help improve the results you get.
Another important point made later in the paper:
Classify results in such a manner that theories about cause and effect can be tested.
This is a simple point but one that is still often not done. If the organization using the PDSA cycle well, this will be done. But often people decide the cause is obvious, the solution is obvious and our fix has solved the problem so we don’t need to waste time confirming it works. Only the “solution” often didn’t solve the problem and failing to effectively test that the problem has been solved means bad results will continue.
Related: The Art of Discovery – Distorting the System, Distorting the Data or Improving the System – Experience Teaches Nothing Without Theory – Inspection is too late. The quality, good or bad, is already in the product. – The Improvement Guide