During my career I have found it beneficial to compare extremes whenever possible to help me solve problems. The first thing you want to do is plot your data. You can use a run chart or histogram to help you do this. Lets take a look at an example using a run chart.
We have a process where we want to reduce the number of defects. Maybe the only data you have is the number of defects produced daily. You put it into a run chart and it looks like the following:
There are many days when a low number of defects are produced. The key is to identify the conditions that exist when a low number of defects are produced versus the conditions that exist on the days when a high number are produced. Take a look at day 18 when the highest number of defects were produced and day 19 which is one of the lowest. The question you have to ask is what changed in the process from one day to the next? Looking at the data this way can help you identify potential clues.