Tips for Effective Data Measurement in Six Sigma Projects
Data measurement is a major component of the Six Sigma methodology, since decisions are made based on data and not on anecdotal evidence or assumptions. Data plays a prominent role in the Measure phase, when baseline metrics are established, and in the Improve phase, when the result of implemented process changes is assessed. It may seem that collecting data is a simple task, but doing it correctly requires understanding and following certain principles. Two basic principles can help you ensure that your measurement process is effective.
Know What to Measure
One of the key aspects of the Six Sigma philosophy is that customers define quality. How does that relate to data measurement? Quite simply, businesses and other organizations too often think they know what is important to customers, without really confirming it. So the first principle of data measurement is to measure the right things. In other words, if you want to improve your customer experience, you need to know what it is that customers are dissatisfied with. Keep in mind that customers can be internal customers as well as external customers and the same principle applies to measuring employee satisfaction as well. Even with process measures, it is important to focus on the right metrics to get an understanding of process performance.
For instance, let’s say customers are unhappy with the length of a customer service call. It would not be wise to focus on the amount of time an agent spends actually talking with a customer, if in fact the real problem is that customers are repeatedly transferred from agent to agent, have a long wait time before speaking to an agent or are on hold for a long time while the agent tries to get assistance.
Know How to Measure
Start with a good operational definition, which specifies the method of gathering data and performing calculations in a way that is unambiguous. Your definitions should be precise enough that if different people are given the same instructions to follow and the same thing to measure (for instance the same set of customers, the same set of product samples or the same data set from a database) they will get the same results. This is known is inter-rater reliability or repeatability. You can also confirm intra-rater reliability or reproducibility, which means that the same person will get the same results each time if he or she conducts the same measurement multiple times. For certain types of data, the Gage Repeatability and Reproducibility test is helpful for verifying that your definitions pass muster.
Having good operational definitions is particularly important when data collection and analysis will be performed by different individuals and/or at different times. For Six Sigma, measures are always needed at different times, because trends over time are important, as is the ability to compare metrics before and after a process is changed. Be sure to document the start and stop points for measuring cycle time, so as to remove the possibility of variation in how cycle time is measured. When appropriate, create visual examples to illustrate how things should be categorized or rated.
The more specific your operational definition is, the better the information you can glean from your data and the more success you will have improving your process performance.
