If data are available from both before and after a planned change, a plot of data over time can be used to see if the change resulted in improvement. If the data depicts a random pattern within a predictable range, we should not infer that a change in performance has occurred.
Shewhart's concept of variation can also be used to help develop a change-that is, help to answer the fundamental question of the Model for Improvement, "What changes can we make that will result in improvement?" Plotting data over time can reveal when the variation in the data no longer follows a predictable pattern. The chart may show an isolated observation or two that are outside the predictable range, or show a new trend.
If the random variation in the data is disturbed by some specific circumstance, improvements can be developed by understanding what these special causes are. People can make changes to remove or overcome these causes if performance is worse, or continue them if performance is better.
If the pattern of variation seen in the data is random within a predictable range, more fundamental changes are usually needed to bring about an improvement. If you don't have baseline data, don't wait to begin testing a change. It often takes time for a change to affect performance. Start collecting data when you start testing, and use the beginning data to understand the current level of performance.
Langley, Gerald J.; Moen, Ronald D.; Nolan, Kevin M.; Nolan, Thomas W.; Norman, Clifford L.; Provost, Lloyd P. (2009-06-03). The Improvement Guide: A Practical Approach to Enhancing Organizational Performance (JOSSEY-BASS BUSINESS & MANAGEMENT SERIES) (Kindle Locations 904-909). Wiley Publishing. Kindle Edition.
|