Be Careful What You Measure For. 3 Keys to Data Collection Success.

Many project leaders become overzealous when it comes to collecting statistics in hopes of characterizing a process to the fullest. Instead of a strategic and focused approach, some choose the buckshot method by sampling everything under the sun. While the idea of being as comprehensive as possible is good, sometimes it is just not practical. Therefore, it is important to develop a Data Collection Plan that clearly outlines goals and objectives along with operational definitions and methodology for the data collection.
When developing a Data Collection Plan here are some typical pitfalls to be weary of:

Distinguish Between Need to Have From Nice to Have
This is where you refer back to your goals and remind yourself to stay in scope. Sometimes data that may seem useful turns out to be irrelevant. This is non-value added information and adds to the noise when trying to interpret data. An easy way to determine if the information is relevant is to think about how you are going to qualify good and bad results and what potential actions you could take. If you have no influence over the data yielded and it does not provide further insight into the process then it may just be a Nice to Have.

Consider How You Will Collect The Data
This one is related to the above note but is focused on what resources are required to collect the data. Is this a manual or automated process? Managers sometimes ask for data without fully understanding how the data is collected and unintentionally add needless manual labour into the process. Often the added work can be avoided by simply talking to the Subject Matter Experts and making modifications to the data format requested. In contrast, it is also a good opportunity to implement automated systems for critical data points if required.  

Time Period of Data
There needs to be an established time period or sample size for the data in question to avoid collecting superfluous information. How many production cycles do you need for comparison? Are you comparing like to like? Are the number of samples representative of the normal process?

Having a well defined data collection plan eliminates confusion and more effectively utilizes the resources required resulting in a streamlined process.

Discussion
How has establishing a Data Collection Plan helped you in your projects? Do you have any other tips on Data Collection to share?