Digitizing supply chain operations with machine learning, analytics, and the Internet of Things

Using digital technology to improve our response in an ever changing market

Supply chains are rarely in a steady state. That’s what makes them fun and challenging. A recent example is the whipsaw we experienced two years ago. At a time when the company, including Supply Chain, was focused on transitioning to new business models – software subscription and SaaS – we had to return to the basics. The market was hot, components were on allocation with extended lead times, and we were in the middle of “the fastest product ramp in Cisco history” with the Catalyst 9000. We found ourselves in a crisis, scrambling to fulfill product and customer expectations, and to not be the reason momentum stalled.

After what felt like years in this mode – but was really quarters – the market flipped. Industry demand fell, inventory rose, and we shifted to mitigating cost and exposure. It felt like the sharpest market rise and fall in a decade.

We weathered this storm with a few bumps and bruises, but given our ongoing investments in digital tools and automation, we fared better than most. We had better access to real time data, which enabled better decisions and faster execution. With this as a backdrop, I want to give you some examples of how this digital transition in supply chain is fundamentally changing the way we work, improving our performance, and delivering superior results.

Greatest hits: payoffs from digitization

Surface Mount Analytics. In printed circuit board assembly manufacturing, yields and throughput are critical. A minor issue that happens early in the process can create significant problems downstream. Catching potential yield issues early and inline can pay off significantly. What started as an experiment, using Cisco technology, looks to have promising results. By using real-time data streaming and analytics to monitor inline processes and trap defects early in the assembly process, we can improve yields on complex assemblies by as much as 25%.

Adaptive Test. For decades we captured every test record, every serial number, and every component lot code in case there was a manufacturing issue we needed to root cause. Through the use of our big data platform, we have been able to use machine learning to mine this data, adapt and optimize test processes. If a test can be tweaked to make it more valuable, it is done automatically. If a test is shown to not improve quality, it is eliminated. This is done without human intervention, empowering our test engineers to do higher value work. Through this effort, we saved tens of millions of dollars in test capital costs and significantly improved throughput. The Catalyst 9000 would not have been able to ramp as quickly as it did without adaptive test.

Leveraging data to decrease RMAs. Return Material Authorizations (RMAs) due to a field hardware quality issue are costly and frustrating to customers. What’s worse is when customers return a piece of equipment when the issue is really a software problem. Cisco has tons of useful quality data: Technical Assistance (TAC) information, engineering data, manufacturing data, software bug data, field data, etc. The problem was that the information was not organized and orchestrated. Through an effort called Quality Orchestration, all of this data was cleansed and organized in a way to drive action. Now, before issuing an RMA to a customer, we are able to leverage this data and validate whether the problem is truly a hardware problem or not. In addition, we create a Digital Signature that allows our TAC organization to recognize it in the future and correct it in the field based on recommended actions. This work is also being digitized as part of IRE (Intelligent RMA Experience), where AI/ML will help guide faster resolution for customer issues through automation. Customers benefit because the effort is put on the right problem, and at the same time we’re saving millions of dollars on avoidable RMAs.

Integrated Materials. One of Supply Chain’s primary goals is to balance often competing priorities among revenue attainment, supply chain costs, and customer experience through trade-off decisions. Optimizing the materials planning and execution requires a robust and dynamic Integrated Business Planning cycle that is powered by real-time supply chain data visibility, digitized workflows, and analytics. By creating a single source of truth foundational data lake, we have enabled end-to-end material visibility from component suppliers through the manufacturing partners to the logistics network. We have invested in applications and analytics, linking sourcing, planning, and execution to drive an optimized result.

Digital Twin of a Factory. In August 2017, Hurricane Harvey shut down much of Houston, which is the location of a significant manufacturing site for Cisco and our largest distribution hub. In the past, we would assemble our normal risk management response, with war rooms and daily sessions to manage through the crisis on spreadsheets. However, based on an investment in a program called UOV (Universal Order Visibility), we had complete visibility of every order at risk, as well all material inbound and outbound. We could essentially manage the factory virtually. The team was able to access what materials and which orders we should re-route, and significantly minimize the impact on customers and Cisco’s operating plan. While some competitors reported significant impact, Cisco was able to recover within a few weeks.

Leveraging IoT to reduce energy consumption and carbon emissions. A team of supply chain engineers came up with an idea to leverage analytics to understand energy consumption within our factories. Even though we manage almost every penny in our supply chain, energy use and its cost were not understood. The idea was to deploy several thousand sensors in one of our factories in Penang, Malaysia to gather energy consumption data from every piece of equipment. Using advanced analytics, they were able to identify inconsistencies and inefficiencies and take corrective actions. A small investment paid off in less than a year, with a 30% reduction in energy consumption and cost. Programs like this enabled Cisco to meet our goal to reduce supply chain carbon emissions by 1 million metric tons in 2020, ahead of schedule.

Digitization starts with culture and data

During our digital journey, we’ve learned that culture is just as vital as technology. We had to convince our teams that digital technologies, analytics, and automation are not paths to job elimination. Rather they are paths to deliver higher value work, increase empowerment and improve performance. We had to embrace a culture of experimentation, creating room for our teams to try things without going through an exhaustive ROI exercise and fighting for dollars. We had to shift our transformation model and mindset to an agile methodology, we had to create a self-service platform, and we had to embrace IT.

We also had to get serious about establishing a strong data foundation across functions and services. We had to partner with IT to embrace the ever-changing technology landscape and be willing to rapidly change. We had to work with our central Operations Data and Analytics team to partner across many functions to link data and enable an end-to-end perspective. We also learned the value of business architecture. Today, our supply chain data lake connects all sorts of data – structured and unstructured, supply chain and engineering, from customers and suppliers – and from it, we’re able to deliver real business value. For us, many years ago, what started as a traditional ERP upgrade led to building the foundation for exploiting digital technologies.

Bottom line: Go Digital, get smarter, respond faster, improve results

Whether it’s the increasing volatility of markets, the shift to consumption-based models, or the ever changing geo-political landscape, supply chains have to be equipped to effectively react and respond in real time. Establishing an effective data foundation, investing in culture change, and investing in digital technologies is essential in this rapidly changing environment. We have seen significant improvement in our ability to respond faster and increase results. But the fun part is that we have just started.

What are the opportunities you see in embracing digital technologies? Please type in the comment box.