Data Management Is a Science
I’ve often theorized that the same concepts used in pathology, which is the study of the causes and effects of disease in medicine, can be applied to the practice of data management in our businesses.
As data stewards, we should think like the pathologists who work tirelessly in labs to understand and fight disease. Studying the organization, processes, and usage of our data is integral to ensuring that the data is working favorably to help meet its associated business goals across teams, departments, and systems. Performing an analysis affords a better appreciation of what is working and why, as well as identifies areas where we face struggles. It can also help uncover the underlying reasons for these struggles.
Our data has many functions. For example, it can be used to prioritize accounts, identify opportunities, and build customer engagement. What’s more, if we’re not meeting our objectives and/or our data isn’t living up to our expectations, we can often identify the underlying reasons by seeing what is “hiding”—sometimes in plain sight—in the data.
Matching Is the First Step
Well-executed data management is an ongoing process, it’s never finished.
At its core, data management is really about managing processes. The data is merely a manifestation of the processes and rules that exist to generate and maintain it.
In most cases, optimizing our data begins with matching—a process that includes working with a trusted outside dataset and then aligning that external data with our internal datasets. The goal is to match our internal records with the same or a similar entity in the more robust, external dataset, thereby acquiring pertinent attributes for data enrichment and augmentation. We also find opportunities to refine our internal data (be gone, unwanted duplicates!).
But matching can be a complex process. Not every record is going to match, and for many reasons. Don’t be discouraged! These no-match records hold opportunities to understand our data and to revisit how we use data throughout our internal processes. Edison was onto something when he said, “I have not failed, I’ve just found 10,000 ways that won’t work.” Believe it or not, “no matches” can reveal the path forward for your data management success.
The Pathology of Matching proposes a framework for studying the challenges and potential to be discovered in our data. Performing a “no-match forensic analysis” can help our ongoing efforts to understand our data and discover opportunities to turn no-match results (what the uninformed might call “failures”) into actionable solutions. It’s no longer enough to say “why” something isn’t working. Our businesses stand to gain so much more when we leverage “what” our analysis informs us that we can do better.