I interviewed Tina Groves who discussed Big Data Innovations for Supply Chains.
Please provide a brief background of yourself
Certainly, and thank you, Dustin, for this opportunity to share with your listeners a bit more about Big Data. I’m Tina Groves. I work with IBM as a product manager. I’ve been with the company now for about 15 years. About a year and a half ago, I got involved with Big Data, and, man, has it been a great ride. Just to give a bit more background, I’ve been a consultant, I have a degree in computer science, I hold a couple patents, then started to getting in to some more authoring, an avid Twitterer, so I hope some of your readers will follow me shortly after this recording.
Can you talk about Big Data and some of the technologies that are enabling fraud detection?
Certainly. Fraud has been huge in the Big Data space, particularly for supply chain, for both the vendors, as well as government. One area that has been an interest is looking at advanced analytics on these vast stores of information to look for new patterns. Even for the data that is already being collected, looking at patterns and, for example, how prices are set. One of our customers at IBM is a medical supplier, and they detected price collusion amongst its vendors by looking at the timing between when contracts were sent out and the types of bids that were received. Another area is entity analytics; particularly, government’s looking at this for looking at relationships between members, whether these are gang members, family members, organizational entities; looking at these relationships to understand how companies are moving goods around, particularly across borders and between countries, looking at exports.
One government in Asia, for example, was looking at how groups of family members were traveling to particular countries and how they were declaring goods on their way back, those different customs duties levied, depending if those goods are being used for commercial use or personal use, and they were able to detect, for example, that these family members were bringing goods back for resale and not for personal use, as originally declared.
A third area is just to dupe itself. A dupe is providing a way of storing data in a very cheap way. When you think about how a dupe allows them for data to be dumped in a single spot and a mind for using these advanced analytics and predictive analytics allows supply chain vendors to then look at their supply chain horizontally. In particular, and get past just looking at financial metrics. You can learn to look at how, in pharma in particular, anything that’s been tampered with can be very easily detected right down to which retailer that the drug was sold from. Those are the three technologies: advanced and predictive analytics for pattern detection; entity analytics to see relationships; and cheap data storage in the form of a dupe.
3. How can processes be optimized in ways that were previously unattainable?
That’s a great question. Telematics has done a huge amount of advancement for helping organizations optimize their processes, particularly in supply chain, so this is really the culmination of several technologies. Seeing how RFID and sensors are now very cheap, they’re used extensively across many organizations, you add in cloud, the effectiveness of sharing information through the Internet and the security that now allows for that type of data interchange, and then you throw in Big Data, which has that combination of cheap storage, very fast processing with more advanced techniques.
You can get down to patterns, for example, that detect, think about a grocery-store chain and an ice cream supplier. When the freezers are set at a certain temperature for meat, for example, if there are freezer fluctuations, which can happen in their freezer, knowing that premium ice cream in the freezer versus some other type of frozen good can affect the quality of the ice cream. The ice cream supplier can then monitor how its ice cream is being maintained through a data interchange with the grocery supplier and then decide whether or not to, for example, validate a warranty. That’s just one type of process of assuring a quality throughout the supply chain. Another one is around transportation; that’s very common, are becoming more common.
Looking at telematics from the vehicles to optimize fuel usage and shaving off dollars by optimizing roots and orders by coalescing orders earlier so that drivers can pick them up. They sound like they’re minor changes, but these incremental changes all throughout the supply chain add up to hundreds of thousands, if not millions, of dollars for very large, larger logistics supply chain organizations.
4. What are your recommendations?
Well, if you're early in your process, earlier in your Big Data journey, we often—by “we,” my colleagues and I—often recommend organizations look for opportunities where there’s business process that is fragmented by different applications, different software applications, and look for opportunities, then, to combine the information flows so that you can have greater visibility from end to end of that business process.
Another one that I look at is if you're very, very early and analytics, for example, is not yet part of your organization’s DNA, look for opportunities where the people in your organization are picking up the phone to talk to people at other organizations just for simple like a status. “Tell me more about the status of this shipment. Where is it? How is it located? Why is the cost of this bill so high?” If those are the kinds of questions your organization’s asking, particularly of the same vendor or the same types of services, then there’s definitely an opportunity, then, for optimization.
If you're more advanced, you’ve had an opportunity to leverage analytics, you’ve got a very good sense of transparency throughout your supply chain, you’re now looking for something more innovative, then we recommend people look at the advanced analytics particularly around entity analytics. You can look at how various parts are put particular, how a system of parts or services are assembled, and look for optimizations, then, in those areas.
About Tina Groves
BI Product Strategist, Big Data at IBM