We recently had a senior data analytics professional from a large financial services organization speak in my MBA Supply Chain relationships class. The scope of the datasets his team was working on was limited to supply management, but the sheer volume of data was staggering in its complexity and fragmentation.
In sourcing, data must be tracked over the entire life of a supplier. In this case, it must be done for all suppliers and for every product and service line purchased over the history of the institution. This amounts to tens of millions of line items, millions of payees, hundreds of thousands of contracts, hundreds of billions of records, millions of invoices, and many, many paper checks issued (still!).
This individual was quick to point out that to “do analytics”, you need to have a solid dataset to work with. And much of the work in analytics relies on what that data looks like, and the quality of the data. Thinking about data in the context of the “life of a supplier”, provides some context for how challenging this really is. First suppliers have to have a profile, which means they have to register. Next there is a request for quote or information, and information on the supplier’s capability has to be identified and stored. Then the evaluation of the supplier must be documented, and linked presumably to a category strategy (although not in all cases…)
Then you have to write a contract – pricing, volume, and then collect that information. And an active supplier will also start sending you things – but are they doing it on time, and what are the metrics with the delivery of the contracted product or service. So you have to measure this and collect it in scorecards you are using. And oh yes, you also have to pay suppliers, assuming you have a correct invoice from them that is validated against the contract and the purchase order. But does the invoice manage the contract – and who is checking that? That is a whole process during the procure to pay process called reconciliation. And when you terminate a vendor, you have to close the contract, and pay them and archive them.
Oh, and don’t forget that you don’t just manage price – but quantity! This involves an activity known as “demand management” – or seeking to throttle consumption. This means forecasting demand for the product or service, and determining if the right level of consumption is being forecasted and controlled in some form.
In every stage of this lifecycle, there is a LOT of information that has to be collected, stored, and then later accessed. Now think about millions of vendors, and different languages, different country codes, different product SKUs, and different currencies….This is making my head swim….
This is going to take a lot more than a big Excel spreadsheet…