The importance of accurate information in supply chain systems.
One of the lessons I learned very early on in consultancy was that however large a data set may be, there is always value in a manual examination to assess its quality. The project that taught me this lesson relied on dimensional SKU data that were apparently well populated – there was nothing obvious to indicate that anything was amiss. Eventually, however, a couple of things didn’t seem to add up, and when we started to examine individual records we realised that, although all the fields were populated with precise looking measurements, they were in fact garbage. The credibility of the outputs took a knock and significant additional work was required to recover the situation.
In very large and complex organisations there are often multiple systems at work, with some overlap in the data fields in each. These companies never really achieve a steady state when it comes to systems, as the pace of change in the business is such that some major systems project is always in progress. It is quite common for parameters such as ‘cases per pallet’, dimensions etc. to be inconsistent across systems in the same organisation.
Even in companies that have implemented ERP systems across all functions, the master data files (SKUs, Customers and Shiptos, Suppliers etc.) often contain many records that have missing or erroneous data points, are partial duplicates, or are obsolete and should have been purged altogether.
Selling a business case for spending significant resources on sorting all this out might prove difficult. But dodgy data sets carry hidden costs. That expensive supply chain optimisation system you invested in needs accurate data to deliver its promised benefits. And when you undertake strategic projects, the supporting analysis may be rendered faulty by our old friend ‘Garbage In Garbage Out’, leading to doubts about what actions to take or even the wrong actions being taken.
So the advice to organisations is work to understand and improve data accuracy in supply chain systems. If you have multiple instances of the same data field across various systems, decide which one is the ‘master’ and ensure there is a process to synchronise the others. Implement proper training so that the people entrusted with entering and maintaining data understand the relevance and importance of what they are doing. And surely we should also be working with our suppliers (and customers in B2B situations) to ensure data is consistent throughout the end to end supply chain?
Finally, if you are undertaking data analysis, remember that however big the data set, it is always worth having a look at it. At the very least sort it in different ways and check the top and bottom few records and sample of those in between. If you find anything amiss, dig deeper!