Interview with Ray Smith, supply chain chief at Microsoft on black swan events and data lakes.
Towards the end of last year, Microsoft launched its first supply chain disruption solution in the form of Microsoft Supply Chain Platform, which bundles together AI, collaboration, low-code, security and SaaS applications in a composable platform. Coming with the service is Microsoft Supply Chain Center, a command center experience carrying data from across existing infrastructure supply chain systems, such as data from Dynamics 365 (naturally), and other ERP providers, including SAP.com/uk/index.html">SAP and Oracle, along with standalone supply chain systems.
With war, pandemic, delivery disruptions and energy crises having defined the decade so far, the Microsoft solution makes easy business sense. So easy, in fact, one may wonder why it’s only coming out now. But as Ray Smith, vice president supply chain at Microsoft, tells Giacomo Lee, Redmond doesn’t see supply chain disruption as a blip nearing its end. Instead, disruption is here to stay into 2023 – and beyond.
Giacomo Lee (GL): What are you trying to achieve with the Microsoft Supply Chain Platform?
Ray Smith (RS): We’re really trying to create an interoperable approach in the market. What is fundamentally challenging are many disparate solutions processes, which result in fragmented data largely being isolated in businesses. The last couple of years have really exposed the frailty and failings of these systems, and customers are simply craving end-to-end visibility.
Our platform (unifies) our assets and some key industries. It has an ecosystem, unified data, and unlocks some AI capabilities, while creating more disruptive collaborative workflows, whether that’s through automation or connecting across users.
GL: Microsoft is touting this as a first, but does it build on anything that’s come before?
RS: We manage two very complex supply chains, one for data centers, another for provisioning of hardware for devices. We’ve had our own journey of a number of key investments in third-party solutions, and frankly we’ve struggled; we lived in spreadsheets probably five years ago, and we’ve struggled to get into visibility and connectivity.
So internally there was a journey to invest in kind of data lake technology, ways to build and develop analytics. And when we looked at the market, we saw customers were trying to do the exact same thing. With all our assets across our investments in Power BI, and in core supply chain execution products, the opportunity we spotted was that if we stitch these elements together for ourselves, and make that much easier, we’d be better able to offer that to customers who have very similar challenges of disparate multi-vendor solutions, lacking that end-to-end visibility.
Frankly, Microsoft has struggled – we lived in spreadsheets five years ago, and struggled to get into visibility and connectivity.
GL: Before Microsoft, you had experience with SAP, Accenture, MIT. How does this past experience equip you with tackling supply chains today for Redmond?
RS: Customers typically talk around unlocking end-to-end visibility, and I often challenge back a bit. I have a background in startups, and if 15 years ago, I went in to pitch a VC and proposed helping to unlock visibility in the CRM market, they would kind of laugh you out of the room because that’s the most lowest level of access to your information.
So it’s really around not just visibility or deriving intelligence, it’s [more] around disrupting workflows. 15 years ago it was all about bringing fast value to customers with novel solutions. Fast forward to now, and I can see supply chain startups find it difficult to break in.
I think my experience on the CRM side and building a startup really highlights the need for an open, platform-based approach that creates an ecosystem and leaves customers to choose the optimal solution.
GL: What kind of supply chain problems are you seeing currently? I’m asking that because supply chain can mean so many different things.
RS: It’s really about disrupting workflows. How do you get those business outcomes rather than having some magical data sitting in a database or a dashboard? For me, it’s really a case of how do we unlock and unify data across all these providers.
Customers really want to have a better understanding of the inventory at hand. A good view on demand planning based on many more signals than they’ve ever had before is a critical problem.
Problem number two is, now I’ve got a good handle on my demand forecast, how do I make an optimized inventory? How do I see across the chain to spot risks, shipping issues, blockages on ports?
Lastly, how do I get a unified field of all my orders across multiple channels, and then orchestrate those orders? It’s optimizing the inventory, but also minimizing the logistics cost. Logistics, particularly for large items, is expensive, and that’s not mentioning the sustainability impact.
Some have invested in kind of dumping all that data into a data lake, but then it stops being operational or actionable. So that’s really where we see our solution stepping in to bridge that gap and interoperate with all these solutions.
GL: You mention port blockages, so how sensitive can something like this be to outside factors? A lot of what’s happening now is due to geopolitical forces, for example.
RS: There are a number of factors obviously, from materials shortages, which, as you said, are mostly stemming from geopolitical tensions, to domestication of supply chains in various initiatives across the world.
Expect these challenges to continue, not to mention “black swans”. A little ripple has a pretty big effect, and then there are more and more ripples, more and more disturbances, which we expect to continue.
But I would say a lot of businesses are realizing that old systems which relied on just historical transactional inputs were really kind of rigid and didn’t offer the opportunity to explore new markets, or optimize margins by shipping in a certain way, or fulfilling products in a certain way.
So people need more signals, more data. Instead of just that transactional data, there’s also shipping lane information, all these external signals from trading partners and suppliers, (signs) if your crop is going to fail.
That info hasn’t been harnessed to that scale. And that’s where AI comes in to really unlock all this volume of data, and get it to the fulfillment manager or supply chain manager.
Gl: How’s AI going to solve this?
RS: AI is really hitting a new age of disruption. But if you don’t start with the data, AI on top of that is almost meaningless. So unification of data is critical.
Now we can do really, really exciting AI algorithms around how we restock, how we do container repositioning. With the help of machine learning and algorithms, we can start getting more predictive on what we should remediate, what is the risk level, what is the weather pattern and how’s that likely to cause disruption in our logistics and so on.
But we know customers have unique supply chains, and we know they have been investing in data science themselves. So, customers can bring their own algorithms to the table and plug that into the platform.
GL: What are your predictions for 2023?
RS: The question that gets asked regularly from people is, do you think we’ve just had a few ripples together and it’s all going to go even keel and flat again? And can we kind of get back to the way we were with existing systems?
The short answer is that the level of disturbances we’ve seen, the level of geopolitical tensions we still see in the world, mean that it’s something supply chain leaders and managers need to be able to handle.
[It’s not] a scare tactic, it’s just the reality of what we’re seeing, and it’s only increasing.