Let’s be honest about it. In many cases, collecting data (especially when done manually) can be tedious and viewed by some as a ‘pain in the backside’. This is understandable to a degree but imagine a situation where after spending 6 weeks collecting data we find out that it is inaccurate, it can’t be used and is in effect a waste of time. This issue can be due to the fact that we put no thought or effort into how we defined the metric in question.
E.g., a Food Processing Company was trying to baseline the Cleaning in Place (CIP) Process. In order to understand if here a difference in the CIP time by shift, product type, CIP types, etc. they set about collecting data over a 6 week timeline to answer some of these questions.
When the Project Team examined the data after the 6 weeks, they found there were some major differences by shift and the other aforementioned factors. Importantly though, this was not due to a difference in performance but by how the Metric was being measured.
- Shift A was interpreting the CIP time as ‘from the time the equipment was stopped until it was started again with the CIP complete’
- Shift B was interpreting the CIP time as ‘from the time the equipment was stopped until an acceptable micro test result for the CIP was back from the Lab allowing the equipment to be restarted’.
- Shift C had another interpretation altogether
Unfortunately, it was then back to the proverbial drawing board!
The morale of the story is to agree on a very specific Operational Definition for a metric, include it on the Data Collection Sheet and even go as far as to give the Data Collectors a fictional pre-completed data collection form to use as a guideline.
Submitted by Éamon Ó Béarra, SQT Lean Six Sigma tutor