Darren Coil is a Principal PM at Microsoft in manufacturing supply chain business technologies. He describes himself as a “hybrid technologist, business process evangelizer and implementer, and program manager all in one.” Recently he has been a key player in transforming business process technologies in Microsoft’s own supply chain.
Darren Coil, Principal PM, Microsoft
Colin Masson, Microsoft’s Global Industry Director of Manufacturing Solutions, recently sat down with Darren to discuss how the Microsoft supply chain leveraged the Microsoft Azure IoT Connected Factory solution accelerator to drive its digital transformation.
Colin Masson, Global Industry Director of Manufacturing Solutions, Microsoft
Let’s start by reminding everyone that Microsoft is a manufacturer and operates a complex supply chain. Can you share a brief overview of Microsoft’s history with manufacturing?
Microsoft’s first real venture into manufacturing came with Xbox in 2001. We worked with one of the key component manufacturers in the world to build the Xbox, and in the process, we started to understand what it takes to run SMT lines, box build lines, and product tests.
But it wasn’t until the launch of Surface in 2012 that we got heavily involved in manufacturing operations. That was a fast and furious learning experience for Microsoft. We expanded our manufacturing expertise and evolved an entire org dedicated to supply chain operations, manufacturing, product design, and equipment design.
Since launching Surface, what challenges has the Microsoft supply chain faced?
Today we are facing challenges in four key areas.
The first is margin pressure. Because we build premium products, we invest big in the research, development, and design phases, yet we can only charge so much for our products before the consumers refuse to buy them. We build iconic experiences into our products, but we have to temper that with the costs.
Customer expectations are also a challenge. While ten years ago this was less the case, today the customer experience is a key differentiator between competing products. A company must have the data and resources to provide highly personalized customer service to be competitive.
Our third challenge is how to visualize a highly complex, highly distributed supply chain network. Our supply chain is spread across over a hundred countries. At any given time, we can have raw materials moving between tiers of suppliers, builds occurring at the main factory, and finished goods moving from the factory to distribution centers to customers and retailers. Being able to visualize this entire process and deliver the data to the right people in time to make essential, on-the-fly decisions is a key supply chain challenge and necessity.
Finally, product lifecycles are rapidly decreasing. It requires a complete shift in mentality to move from a long-running product like Xbox to the Surface programs, where you only have twelve months to get the product into the hands of consumers before you are launching the next big thing. That puts a lot of pressure on the supply chain.
These four challenges brought to light inefficiencies in our data modeling, handling, and understanding. We had so much data available, but our data feeds existed in disconnected systems, so we were exerting considerable effort to bring together the data necessary to answer even a single question. We realized that we had to find a better way to manage our data to effectively address these challenges.
Can you share an illustration of how the Microsoft supply chain managed production data prior to the digital transformation?
Yes. Every week our corporate vice president in Redmond would hold an operations meeting with our primary Surface build factory in China. To create a PowerPoint deck with the content for that meeting, the team in China pulled data from any number of disconnected systems—from an SAP system, from shop and control systems, from a supplier’s ERP system, from an engineer’s laptop, from a material planner’s spreadsheet, etc. They were stitching together all this data each week, spending twenty to thirty man-hours just to build a PowerPoint deck.
And, by the time the meeting was held on Friday, the data was already outdated. The yields and quarantines that the factory was reporting Thursday night had already changed by Friday morning. The problems they had thought were paramount had already been supplanted by newer, more pressing problems. This system was never fast enough to effectively manage the factory from Redmond, or even on premises for that matter.
And this was no isolated problem—this inefficiency existed throughout the organization.
How did the supply chain get started with its digital transformation and how long did it take?
Our VP sent a core five-person team to the Surface build factory in China and instructed them to automate the way that the factory was preparing this weekly report with the goal of eliminating the need for an operations meeting altogether. That was the first tangible milestone that we aimed for.
Looking back, we see our digital transformation as having a three-wave approach. The first wave was getting connected, the second was moving from using our data reactively to using it predictively, and the third was implementing cognitive tools.
All in all, it took the team under five months at the factory to develop and implement the first two waves of digital transformation driving tangible results, although it was spread out over the course of a year. The third wave has successfully demonstrated a couple of specific use cases and is being expanded to solve more problems for us as we speak.
How did the Microsoft Azure IoT Connected Factory solution accelerator contribute to the overall transformation?
Microsoft Azure IoT Connected Factory addresses a specific set of needs within our digital transformation. It is the avenue by which we move equipment data from the factory into Azure. We needed to connect around fifty data systems and business systems across the company and put them into the cloud to create a single location where the data from across the supply chain is tied together.
We started out using basic PowerShell scripts to move the data, but we needed a way to do it faster. That’s where the Microsoft Azure IoT Connected Factory solution accelerator came in. With a product lifecycle of only twelve months, we need to reconfigure our factory quickly. We wanted to be able to set up new equipment and quickly get data from that equipment into the cloud without spending a lot of IT calories.
Before implementing Microsoft Azure IoT Connected Factory, we had to go through a lengthy process working with the equipment supplier’s IT group to move data to our side, validate the data, adjust their extracts, and settle on data latency and transformation issues.
Since implementing Microsoft Azure IoT Connected Factory, the process is much simpler. When a new machine is installed, we simply enter the IP address into Microsoft Azure IoT Connected Factory and indicate how often to pull data, and that data is immediately flowing into Azure. We no longer worry about how long it will take to get data because now it’s nearly instantaneous.
And with the Power BI backend set up, as soon as data is flowing into Azure, we have processed data flowing back to the factory. This near-real-time data feedback enables the factory to accelerate its qualification cycle with a new machine.
How does the availability of data from Microsoft Azure IoT Connected Factory influence operations?
The data is used by the process engineers and equipment engineers who are responsible for the machines. It enables them to make intelligent decisions in managing the equipment day to day and helps them plan for the future. They can identify, say, which machine builders are producing better machines and use that insight to drive their equipment design strategy.
What substantive results has the factory seen from the digital transformation?
Within hours of implementing the digital transformation, we uncovered inventory that was headed towards obsolescence. That data was always there, but we weren’t seeing it until Microsoft Azure IoT Connected Factory highlighted it for us. The factory has saved countless units from being scrapped by using data to make line improvements and identify quarantines. Ultimately, it saved upwards of $5 million in a twelve-month period just based on what our five-man team accomplished. And finance is reporting that we’ve reduced inventory costs by $200 million.
Beyond the factory, how is this transforming the supply chain?
Good question. Although our digital transformation focused on the factory, we’re collecting data from the entire supply chain as well. Microsoft suppliers are now tied in to our neural network and supply us data directly. We no longer have disputes with our suppliers over data, because when there’s an issue, we can review data that is coming directly from the supplier. This generates great pace in the supply chain.
What lessons did you learn from the process?
They say hindsight is 20/20 and that’s certainly true here. The biggest lesson we learned was that we were over-cautious. When we implemented the digital transformation, we did it in six-week chunks, waited for results, and then started up again once we confirmed that the benefits were real. If we had done the entire implementation in one go, we would be so much further along than we are today.
Another key lesson was that it’s not necessary to upgrade any business systems. Often IT will tell you that you need to upgrade your SAP or your SQL Server before you can tie data from your systems together. With Microsoft Azure IoT Connected Factory, the whole system of intelligence sits on top of your existing business systems, so your systems can go through their normal lifecycles. The lesson is this: don’t let anyone stand in the way of digital transformation by insisting that you need a series of expensive and complicated upgrades. Just start small, break off one chunk at a time, and digital transformation builds itself quickly.
How has implementing Microsoft Azure IoT Connected Factory laid the groundwork for future transformation?
Microsoft Azure IoT Connected Factory has enabled a number of exciting projects that are coming down the pipeline.
One is deploying digital twin to create next-generation automated test equipment. In the future, if we need to push a new test sequence to the factory floor, the engineering team in Redmond can deploy the new test code, observe the results, and have a full understanding of the status on the factory floor. Digital twin in combination with the Microsoft Azure IoT Connected Factory IT gateways makes this possible.
We also plan to implement more machine learning. We already know that with machine learning we can build a better process, but can we use our data and machine learning to build a better product? We will use machine learning to look at things like, is there a better combination of components from different lots that tends to produce the best results?
Finally, we are planning to deploy Cortana to eliminate errors in the factory audit process. Even today, in virtually every factory, there’s still at least one person who walks around with a clipboard, taking down information. We will eliminate those clipboard operations in our factory by supplying the line manager with an Azure Cortana-based interface. The line manager tells Cortana the information and Cortana automatically updates the back-end system. This sounds like just eliminating a small nuisance, but ultimately this will eliminate a lot of faulty data in our information system.
I’m excited because this journey never ends. There will always be new suppliers coming on board, new machines to hook up, and new challenges to face—and today we’re in a comfortable place where we can take it all in stride.
Start your journey today with Microsoft Azure IoT Connected Factory
To learn more about the Microsoft Azure IoT Connected Factory solution accelerator, check out our solution demo, or contact us today to discuss how Microsoft Azure IoT Connected Factory could transform your business.