* Multiple solutions naturally result. At least three different solutions will naturally result, and probably many more! When solutions are being devised, I focus on each subsystem and ask myself how I can alter the situation to improve response time. For example, when focusing on the Oracle circle, I ask myself, "How can I alter the Oracle instance to reduce response time?" The answer may be as simple as increasing the buffer cache, but the focus is clearly on changing Oracle. It is common to develop multiple solutions for each subsystem. The realities of availability and change control mean that some changes are not immediately possible, need to be scheduled, or are simply not an option. Multiple solutions will be attacking the same problem, which obviously increases the chances of a significant performance improvement.

* A story develops organically. Your analysis will naturally develop a story. Yes, I said a story. If you can turn a raging technical catastrophe into a plausible story that both technical and nontechnical people can understand, you have a very good chance of people believing you and therefore embracing your recommendations. Why? Because it makes sense to them, and people are much more willing to believe in someone they understand. Don't underestimate the power of a simple story. Throughout history, stories have repeatedly demonstrated the ability to change entire cultures and countries. Your technical analysis may be staggeringly accurate, but if you can't convince others, they will not accept your recommendations. A simply story can transform a very complex problem into a tangible and understandable situation. The beauty of the 3-circle analysis method is that the technical analysis naturally creates the story for you. I'll show you how this works after we look at some case studies of 3-circle analysis.

* Avoid finger-pointing; breed cooperation. Unless there is a gross misconfiguration, everyone involved can contribute to the solution. Since everyone has something to contribute, finger-pointing stops. It also helps put pressure on vendors or groups who state there is nothing for them to do, because essentially they are saying that there is nothing they can do to help. No one wants to say they don't want to help. For one group to place the blame entirely on another subsystem is usually not appropriate, demonstrates ill will, and will result in a suboptimal solution. When discussing solutions, look for every circle's group to contribute significantly. There must be a very good reason why a circle's group does not wish to participate in solving the problem.