The chain of data can be seen to start with the purchase order. That would be a fair assumption if we're including only what has been considered to be 'data'. The traditional definition of data considers only values that exist in some kind of definable table or database. That is, data that has defined meanings. EDI documents are considered data. Databases are considered as data. Spreadsheets cross the line with regard to data because the definitions of their content are only local to that particular spreadsheet. And all the other information stored in computer systems is considered 'unstructured' and has traditionally been seen as only to be interpreted by humans. 

The difference with Big Data is that all the information existing in computers is fair game to be used as data. That includes spreadsheets, text documents, images, voice recordings, Twitter messages, and all other social media, and lots more that is yet to be created. The tools needed to convert these varied files to data do exist - for the most part - and are being leveraged aggressively to deliver the resulting data to analytical engines that provide insights to issues that were never thought answerable before.

So where does the supply chain figure into Big Data? The string of activities includes all the documents used to develop forecasts, create budgets, present plans, and discuss options at the beginning, and continues after the sale including register data, warranty information, customer service requests, repair services, Facebook comments, and every identifiable activity that relates to the product. When all these items are available as data and incorporated into the pot of Big Data we start to see the overall impact of the supply chain. And as the processing of the data becomes more available and the tools more accessible we will begin to see an expansion of supply chain visibility that goes beyond locating a delivery.