The Intersection of Predictive Analytics, Predictive Repair and Reverse Supply Chain
In the mid-1900s, information was trapped in books, and the growth rate of libraries indicated there would be hundreds of millions of books sitting on shelves by the mid-2000s. However, incredible technological advancements and new types of media have made information easy to capture, store and communicate anywhere, anytime. Now that systems and devices include features to capture data, the race is on to figure out how to use this data in innovative and transformative ways across every industry. CIOs understand the impact data can have on business transformation, making today’s CIO one of the most important change agents of the digital age.
Today, there is power in data-based predictions. One way I’ve seen this put into practice to benefit customers is with predictive repair. For example, engineers collect and structure data from built-in diagnostic tools, tech support interactions, hands-on diagnostics, defective part evaluations and more. They use this data to determine the best fix for a system issue after it presents itself and, in some cases, identify potential future issues and prevent them from occurring. Customers benefit from less downtime and stress, and better customer experiences, while vendors use fewer parts for repairs, reduces turnaround time and lower repair costs. Additionally, product engineers can use the findings to design next-generation platforms, resulting in better products.
Using predictive analytics to tell us about tomorrow so we can better address customer needs today is how we need to leverage our data
Building and deploying an optimized, predictive repair-driven reverse supply chain can be complex. Multiple groups need to collaborate to develop an ecosystem, revamp operations and develop work flow models. Using predictive analytics to tell us about tomorrow so we can better address customer needs today is how we need to leverage our data. For this article, let’s focus on the application of predictive analytics to repair customer devices.
Reverse logistics drive billions of dollars in costs and can be a headache for many customers. From the customer’s perspective, losing access to their device while it is in the repair process is becoming synonymous with a life-altering event, because so many of our daily tasks are enabled by technology. From a CIO’s point of view, even though massive amounts of data may be collected during the repair phase, it is not necessarily used to unlock insight into customer experiences, product performance, or to drive technological advances.
Predictive repair combines all relevant data to identify patterns and predict outcomes, such as what will need repair before the unit arrives at the repair depot or even before the customer is aware that a fault exists. The adoption of big data mining and emerging data science technologies, including machine learning, deep learning and AI are shaping new ways to better serve customers and drive better outcomes. The CIO is at the forefront of this evolution and adoption of new strategies.
Many customer devices continue to be sent to a repair shop for deeper diagnostics and issue resolution. Replaced parts are sent for failure analysis to determine what, if anything, malfunctioned. During each step, data on each part is gathered, entered into a system and stored. Later, certain data may be accessed and used as part of a targeted analysis.
CIOs may understand the concept of data-based predictions, but building an ecosystem that connects, interprets and learns from disparate data, as well as creating a model and operational infrastructure to make predictive analytics work for repair, are two separate but equally important tasks.
First, there’s the need for common language. With thousands of technicians, systems and parts moving through the reverse supply chain at any given time, it is critical that common issues are described using a uniformly accepted vernacular. Requiring technicians to align on symptom descriptions provides deeper insights into repetitive issues and enables better grouping for more thorough analyses.
Secondly, traditional operations process flows must evolve to fit an answer-first model. When a system arrives, it is diagnosed and then repaired. In the future, a repair solution is predicted with a part identified, prepared and paired with the system upon arrival, and then the repair takes place. Given the predictive repair engine will not always be correct, the process flow needs to accommodate repair paths to optimize current and future state models. This presents major investment and training challenges for existing operations. While there are many ways to address this, we’ve found piloting variations of these lines on small product sets lessens the impact to larger repair depots and mitigates risks to customers.
Lastly, the repair engine should continue to learn from the confirmation of accurate predictions and incorrectly predicted recommendations. Put simply, it will get smarter over time. Our initial pilot resulted in 80 percent accuracy and has quickly increased as the machine learning algorithms drive smarter predictions. Each successful prediction reduces the total time to repair by 20 minutes. This is resulting in a more efficient workforce and the ability to allocate inventory more accurately reducing parts movements by 15 percent. We reduce downtime for our customers and drive a better overall customer experience.
As consumers, we’ve come to expect that the next generation of products will be smaller, smarter and more advanced. However, producers of these gadgets are not necessarily working against those same expectations. As a result, businesses and customers alike rely on CIOs to expand thinking on how data coupled with advanced applications can turn time-consuming customer experiences into low-stress affairs, while driving business value.
Predictive repair represents a way CIOs can reimagine the use of data to not only resolve issues more quickly and accurately, but also to prevent issues and spur innovations throughout the enterprise. By using data in creative ways, CIOs drive operational improvements and workforce transformation, shifting how they marry technology with process to improve delivery of goods and services.