BACK TO BASICS
Eight Simple Steps
Improve your data collection process with a concise plan
by Mohit Sharma
Collecting data is a critical and important step in putting a project together. In experience reviewing projects, I have seen Black Belts often make mistakes in collecting data. This affects the overall result.
A Six Sigma practitioner must first create a comprehensive plan to facilitate data collection. Simply stated, the beads must be with you before you sit down to string a necklace. Here is my eight-step process for data collection. We will call our example project "Project Y." (Steps one through five are shown in Table 1).
Step one: Identify whether Project Y is continuous or discrete.
Step two: Write down the operational definition of Project Y. This is important because everyone involved on the team should understand Project Y in the same way. For example, a project on reducing the cycle time of a transaction can have different interpretations by different users if the operational definition of "cycle time" is not clearly articulated. Importantly, if required, the operational definition must have a mathematical formula to ensure uniformity.
Step three: Identify unit, defect, opportunities and specification limits for Project Y before you proceed on the data collection plan.
Step four: Identify potential X’s set for Project Y. Again, it is important to collect data on X’s set along with Y in one attempt.
Step five: Write down the operational definitions for X’s set as well.
Step six: Identify a statistically validated sample through Minitab or any other similar tool. Remember, a selected sample must represent the characteristics of the population. Using different sampling techniques to identify a sample is acceptable. (See Table 2 for steps six and seven).
Step seven: Create the data collection plan: 4W1H is a simple lean Six Sigma (LSS) tool that can be used to create the data collection plan. The 4W’s are: What?, When?, Where? and Who? The H is How? or, the method used.
Step eight: After the data collection plan is ready, create a data collection template in which the data will be captured (Online Table 1).
To ensure uniformity in data capture:
- Capture the unique wire transfer number in column one (Online Table 1).
- Capture the account number and relevant details of the same wire adjacent to "Wire No." in the same column (Online Table 2).
- Repeat the same process for the rest of the wires picked in random sample (Online Table 3).
This is a simple eight-step process, but following it provides remarkable results in the analyze phase of any define, measure, analyze, improve and control project; it’s a tried and tested, practical solution that avoids rework in data collection. Aside from preventing added work, it also provides authentic results. It is a solution that is all about putting the horse before the cart.
Waddick, Patrick, “Building a Sound Data Collection Plan,” iSixSigma, http://tinyurl.com/data-collect-pla.
Mohit Sharma is assistant vice president for Genpact in Gurgaon, India. Mohit earned his master’s degree in marketing from Symbosis University in Pune, India. He is an ASQ member and a certified Six Sigma Black Belt.