I interviewed Richard Sherman who discussed Supply Chain Trends 2017: Analytics.







What about analytics? How are the analytics done in a supply chain ecosystem?


Great question, Dustin.And I think, basically, where the analytics come in is as we become systems, as we become connected via the internet, as we become connected their mobile devices, so we're an increasingly mobile group... So as individuals and markets and companies and businesses all become connected seven days a week, 24 hours a day, and we begin collecting more and more data in real time... So, for example, I may have a wearable health-monitoring device that monitors all of my feelings, all of my hunger, all of my heath activities. It may determine, before I do, that I'm going to crave a meal, or I'm going to crave a beverage, or I'm going to have need for some type of a medication, or I'm going to need to have some type of a treatment. Or maybe I need to step up my pace a little bit because I'm not burning the calories that my goals have been set in my wellness system.


And so as I'm collecting all that data, that data is transmitted to all of the nodes in that ecosystem and then now companies have data about every individual within their market echo system, every intermediary that they're dealing with within their channel. So we have an integrated channel-wide data collection system, which is now feeding that entire ecosystem with real time data about everything and anything in real time, including where it's occurring, because — don't forget — we've got a GPS system within all of our mobiles and all of our wearables. So we can also service by location where that individual is and where that person is. So I know who they are, what they are, and that kind of led to the discussion we were getting into last time we talked, about security, privacy, and data ownership.


But from an analytics perspective, I now have more data than I ever have been able to dream about. So master data management really becomes a business discussion and becomes vital to survival in the market. Because now I'm sharing data. I'm collaborating with data more than ever before, and so I have to be able to harmonize that data not only across all the different functions in my enterprise, but I also have to harmonize it with all of the other participants that I engage with in my market ecosystem.


Regulatory compliance is going to drive more and more digitalization. So automotive parts tracking, food ingredients tracking, pharmaceutical ingredients tracking, counterfeiting, all of those processes and all of that information are going to become more and more governed and regulated than ever before. So again, I have to be thinking digital. I have to be thinking ecosystem. I have to think of all the participants of my ecosystem.


So from an analytics perspective, what I have to first do is I first have to...the first level of maturity, if you will, is having harmonized, descriptive analytics. And the descriptive analytics really define all of the characteristics of product and supply network attributes. So it's very descriptive about all of the data and all of the product and all of the movement, all of the networks and nodes, all of the transportation lanes and characteristics. So I describe all of the attributes of everything within my ecosystem. Then I can begin to apply sensors and monitors and other Internet-of-Thing types of devices that can monitor the health of that ecosystem, the movement of goods within that ecosystem, any changes of control within that ecosystem, are beginning to basically now collect data about the condition or environment or ownership of all the physical materials that flows and transforms through the network.


So this is where blockchain comes into play. And companies are beginning to look at blockchain, which is the underlying technology that supports the capability to use bitcoins. So what the blockchain does is it audits and logs and tracks every transaction that occurs electronically within that market ecosystem and validates it and validates it for its accuracy. So now I'm using blockchain to address some of the security and data ownership of privacy issues.


Now that I'm collecting all this information and all this data about the condition and the environment and the ownership of the physical materials, now I can begin to apply logic to that data to begin to assess performance and compliance or deviation to my plan. So now I'm beginning to move into a more diagnostic use of analytics.


So now I'm beginning to diagnose where am I having a problem? Or what's happening in my network right now that requires my attention? I was planning for a spike in demand to occur associated with a promotion, and that spike is growing faster than I anticipated. I need to be alerted to the fact that the variation from my plan is great enough for me that I have to take some action to that.


As I begin to understand those diagnostics, I can now begin to develop predictive analytics. And what predictive analytics do is they leverage business intelligence, artificial intelligence, cognitive analytic capabilities where now we're starting to create decision-support applications that begin to kind of think like human beings and begin to mimic the human brain. And now we begin to start getting into predictive analytics.


So now I start looking at trends, patterns. I'm bringing in external data like the weather, conditions like maybe there's a huge sporting event that's about to occur. So anything that can cause a dramatic change or deviation from my plan, I can now identify and begin to predict what's going to happen. So now I can not only identify at the time something happens that I have to take action, but now I can predict to say, "You know, New England just went into the Super Bowl against Atlanta. It's going to held in Houston. There's a really good chance that we're going to see a demand for all kinds of entertainment related goods. So we're going to need more Patriots hats and t-shirts down there. We're going to need more Falcons hats and t-shirts down there. We're probably going to need more beer down there. We're going to probably need more adult beverages. We're probably going to have to divert a lot of chicken wings into Houston. Well, the Super Bowl in itself is a global event. We may need to take a look at what the consumption of chicken wings are all over the planet.


So now I can begin to predict where and when I'm going to need things to happen. Or I could, in the automotive industry, I can use it to predict when certain parts are going to fail, and I can route the vehicle to a service center before it actually breaks down on the road.


Then once I begin to collect the information about how we resolve all of those predictions and all of those deviations from norm, now I can begin to create analytics that are prescriptive. And here's where I'd begin to take a look at the historical resolution actions, and I can apply an advanced-process control algorithms that are developed not only to predict when something is going to happen but also prescribe a series of actions or options a person can take to resolve the problem that I've predicted in my analytics.


So the sum it up, analytics maturity goes through basically five different phases, and there's no skipping phases. So you've got to develop first, descriptive analytics based on master data management. You have to then develop monitoring analytics, which takes advantage of all the new data collection devices that are being deployed across the network. I then have to begin to develop diagnostic analytics that will enable me to assess performance and identify potential problem areas. Then I can use predictive analytics to bring in all the external data based on those diagnostics to begin to predict when conditions arise that will cause something to happen different from what I've been planning.


Finally, I get the Holy Grail, where now I've got prescriptive analytics, where now we're doing digital collaboration. I'm no longer... I'm basically now managing my supply networks as a process, and my systems become process control systems, and they're constantly monitoring the different flow paths the goods are taking, and they're constantly monitoring if those flow paths are operating smoothly, and they're digitally alerting either systems or people to take action if there's a deviation from plan. That's all going to occur in real time, and so supply chain becomes a digital, smart supply network, or a smart digital-supply network.


Does that make sense?


Yes. Great. Thank you for sharing today, Rich.


Hey, Dustin, it's always a pleasure to speak with you. I hope I shed some light on some of the trends I talked about last time, and perhaps we can plan a time in the future where maybe we can get in and start talking a little about robotics and some of the issues around segmentation and optimization of new and emerging markets.

I look forward to that.


Sounds great, Dustin. Thank you.


Thank you.


About Richard Sherman




1edda92 (1).jpg


Richard Sherman



Senior Fellow, Supply Chain Centre of Excellence at Tata Consultancy Services


LinkedIn Profile