We live in an era in which we’re inundated with data. Is it helping us as leaders to make better decisions?
Ironically, when people don’t like the data, they can fudge the numbers, sweep them under the carpet or look for different numbers. All too often the conclusions presented are agenda driven rather than data driven.
The quantity of metrics generated by even a single visit to a website can make your head spin. From budget decisions in Washington to the recommendations of your local ad agency, reams of data can be presented to support a proposal you’re asked to endorse.
Yes, there is. According to Dr. Ronald D. Snee and Dr. Roger W. Hoerl we should be examining what they call the pedigree of the data.
Because data quality is so important, it’s my aim here to share several insights from an article penned by Snee and Hoerl titled Show Me the Pedigree. What they propose is that when analyzing data we not only look at results, but also the origin of the data, how it was collected and how it has been handled. Data quality must be paramount if it is to have any value at all.
The word pedigree is a familiar word they’ve chosen to borrow from animal husbandry. “The value of a yearling racehorse depends significantly on the quality of the pedigree.” The concept is applicable to a wide range of disciplines. In the legal system the integrity and reliability of the evidence is of vital importance. Similarly the Food and Drug Administration (FDA) requires complete, consistent and accurate data.
The Snee and Hoerl article centers around one of the passions I myself have associated with continuous improvement. That is to say, in order to have sustainable continuous improvement, the reliability of program data automatically generated from your business systems is critical. First and foremost, you must understand the type of questions you want answered. Next, you must have the correct content of that data in order to answer those questions. That is, the words and numbers must have the values you need in order to answer the questions.
Their article deals with the crux of that last sentence. Without the proper pedigree of the data, the answers you generate may be of no value. Drawing inaccurate conclusions from data you wrongly believe is accurate can lead to wasted efforts and a whole bunch of other bad things. Some might even be critical to the survival of your business.
Way too often I see charts and data presented in a very lackadaisical manner without any explanation of how that data was collected or analyzed. Often these things are presented as facts and my skepticism often peaks. This is why Snee and Hoerl refer to the importance of maintaining a “chain of custody” regarding the data.
Another critical aspect for data pedigree is how the data was collected and how it was measured. In my experience, people all too often draw conclusions based on assumptions that may or may not be correct. Proper sampling, operational definitions on how to collect that information, and understanding how the measurement is to be “measured” must be addressed and shared for effective conclusions.
The authors note the importance of including a description of the processes used for taking measurements. This is another of my core convictions. A well-defined process will have three core elements–documentation, repeatability and predictability. We’ll explain a well-defined process in another article.
Bottom line: we’re buried in numbers. The correct analysis of this data is situation critical. What kinds of checks and balances does your organization have to confirm the pedigree of the data?
-Read the article “Show Me the Pedigree” here: http://asq.org/quality-progress/2019/01/data-quality/show-me-the-pedigree.html Note: Registration is required.
-Request the article without registering by emailing firstname.lastname@example.org.
# # # #
“Show Me the Pedigree”—Ronald S. Snee and Roger W. Hoerl; QualityProgress.com, January 2019