Do You Have Bad Strategies or Bad Data?
Q: Why do most supply chain strategies fail?
A: It comes down to data. Most companies have data quality issues. Companies make decisions with no data, partial data or, worse, misleading data. The misleading data is the worst because it looks right, but is inaccurate. Decisions made with incorrect data lead to no cost savings and compounding incorrect data throughout the supply chain.
Q: What is the definition of data quality?
A: A quick research project will reveal there are two definitions of data quality. Data is of high quality if:
1. It represents the real world that it describes;
2. It is fit for its intended use.
Really, you have two criteria, not two definitions, because they are not mutually exclusive.
For example, a freight invoice record may have the right data to calculate a rate. If you are talking with the accounts payable team, they would say the data collected is of high quality. The freight invoice has the origin, destination, ship date, weight, and miles. From a financial perspective, this data accurately represents the real world that it describes, and is fit for its intended use.
When the supply chain department is tasked with finding a more cost-effective way to package their products, they ask the accounts payable team if they have high-quality data they can use to perform the transportation piece of the analysis. Based on our above review, the answer would be a resounding, "Yes."
However, the freight invoice is missing data about the shipped product, so it can’t be used to do a financial analysis based on changing the packaging of a given product. In this case, the data is unfit for its intended use.
Q: What is the challenge of creating data quality?
A: The challenge is the many different uses a company needs for what appears to be the same record. Back to our example, we have two departments with very different needs.
Most companies ultimately create data silos. The accounts payable team will copy the invoice data and match it up to another shipment dataset with product information. The team sorts the data and "cleanses" it by actually removing assumed anomalies. As a result, the company now has partial data at best and misleading data at worst.
Q: How can a company create data quality? What is the payback to having data quality?
A: Bad data is estimated to cost companies more than $3 trillion annually in the United States. Companies have to create a data strategy that defines the different intended uses of their data along with the data elements required to fulfill these uses. With data-quality tools, companies can achieve 10:1 ROI or greater. The payback is enormous.