Design, deliver, and run enterprise blockchain workloads quickly and easily.
Breaking down data silos with closed-loop manufacturing
Imagine you work for one of the world’s leading auto manufacturers and you are assigned to observe a test drive of a new car model.
The trial will look something like this: A professional driver will be assigned to operate the vehicle in a closed circuit, with three engineers sitting in the passenger seats with their laptops. While the driver takes the car through its paces, data streaming from the vehicle’s systems will be transferred to the laptops using traditional RS-232 serial connections. The engineers will monitor the live streams and store more complete datasets for detailed analysis back at the office.
Each will gather a specific set of metrics. The design engineer will want to measure vibration and sound levels. The developer working on the electronic braking system may be looking at the data for a single ABS component. The HVAC engineer will want to measure how long it takes for the air conditioner to drop the interior temperature 5 degrees in different weather conditions and passenger configurations.
Silos vs. the virtuous cycle of closed-loop manufacturing
What’s wrong with this picture? On the surface, nothing. The engineers are doing exactly what they need to do for their respective areas of responsibility. Thousands of similar test drives will take place before the design, engineering, and production specifications are locked down and manufacturing can begin. For decades, this has been the way the auto industry has prepared new vehicles for market.
But there is a problem with this approach and, for that matter, similar approaches used in other industrials sectors. It relates to data. While each one of the engineers will return to their desks with terabytes of performance metrics and diagnostics, there is no common framework for gathering and structuring the data, much less sharing it with other departments. In effect, processes that govern the design, manufacture, sales, and service of the vehicles are dependent upon a collection of data silos. This leads to all kinds of negative outcomes, including:
- Missed market opportunities: Production and sales forecasts are lower than actual market size or market saturation.
- Design flaws: Products don’t meet customers’ needs or preferences, or they include features that are not needed or are not being used as intended.
- Production inefficiencies: Opportunities for lowering production costs are missed.
- Service and maintenance deficiencies: There are delays in responding to defects and flawed customer service.
- Unidentified correlations: Important connections cannot be made due to missing information or siloed datasets that prevent insights from being shared across departments.
These issues represent wasted opportunities to improve efficiency, make better products, and compete more effectively.
An emerging framework called closed-loop manufacturing provides an alternative. It leverages the vast amounts of data that are already available to industrial companies but are often locked away in silos and, therefore, underutilized. The general idea is to take all the data generated by different departments and stages in the process and, instead of keeping it in data silos, make it available for participants across the organization.
The data is not shared haphazardly, however. For an auto manufacturer, closed-loop manufacturing uses structured processes and continuous data flow that starts with R&D, design and prototyping, production, aftermarket services, and circles back to R&D, creating a virtuous cycle.
Figure: Illustration of the closed-loop manufacturing concept
Image credit: HPE
Data silos in the auto industry
Data silos are common across companies and industries. Employees in a team see themselves as being responsible for a small process or task and will gather data to serve the goals of their own departments or to meet a certain specification. Seldom will they be thinking across organizational boundaries, let alone other functions.
Returning to the auto manufacturer example, the new car model may be plagued by a defect once it’s out in the field and customers start driving it. As owners become aware of the defect, they will continuously show up at dealerships or service centers to get their cars fixed under warranty.
But it will take a long time for the manufacturer to actually address the defect, precisely because of this silo mentality, which ends up slowing the transfer of information from the customers, sales, and aftermarket services back to the design, engineering, and production departments. Feedback loops will be further hampered by inconsistent data types and the limitations of the company’s IT infrastructure.
The data feedback loop can take too long, having a clear impact on the bottom line. If the defect involves a safety issue, the actual production fix may not be available for 18 months. For less critical defects, it may take two or more years to make the necessary changes in production. In the meantime, there’s a high likelihood that a new car rolling off the assembly line will carry that known flaw. A U.S. study based on 30 years of National Highway Traffic Safety Administration recall data found that the industry average recall rate was 1,115 per 1,000 cars, reflecting the fact that many models are recalled more than once.
Data silos can also lead to sales problems, which leads to impact on revenue. In a given year, a major car manufacturer may produce tens or even hundreds of thousands of cars that cannot be sold because they are the wrong color, wrong transmission, or wrong equipment customization. This is not an engineering problem or manufacturing mishap, but rather a data and insights disconnect between the sales organization and production management. In essence, production management produces what they know makes sense based on their own datasets and calculations, rather than using data or insights from sales staff who interact with customers every day and have a much better grasp of market needs and preferences.
How to get started with closed-loop manufacturing
The goal of closed-loop manufacturing is to capture data about products—how they are being used, their performance, and in what context—and extract relevant insights that can improve processes across the value chain within the organization. While it won’t immediately eliminate defects or excess inventory of certain auto models, it can help jump-start the auto manufacturer to more quickly identify and respond to problems, which will also benefit its supply chain partners.
It all starts with connecting data silos. While a large manufacturer will typically be well-equipped with IT systems, it may not be interconnected to the degree that allows closed-loop manufacturing. There will be CRM, ERP, and PLM systems supporting departments like procurement, sales, and distribution, as well as MES driving core manufacturing processes. But if something goes wrong in, say, the production stage, a lack of data and insights feedback loops may prevent the teams from understanding what’s happening and developing a timely response.
Most companies considering a closed-loop manufacturing model of sharing data and insights are starting small but thinking big. Unless it’s a greenfield project, no company can transform everything at once without putting entire processes and production at risk. So, companies start with pilot projects that serve as proofs of concept, targeting smaller use cases that can deliver quick results and be easily scaled up or across. And they're using the results as part of the “internal sales” job: convincing others within the organization that the framework will have a positive business impact and offer superior outcomes to the bottom line.
As with any changes in large organizations, it’s critical to build support from the people that are going to be directly and indirectly affected, across levels and organizations. You can buy all kinds of disrupting technologies and come up with plans to connect siloed systems and data according to framework, but the success will hinge on the understanding of the vision and benefits, as well as the support of employees on the production floor. Many of them have been performing the same set of tasks for decades and will be skeptical of or even feel threatened by massive change.
With all of these considerations and many more, the task of defining the scope of a closed-loop manufacturing project and getting started can seem daunting. This is when partnerships and consultation with the right ecosystem of partners—both from IT and the operations side—is essential. You can leverage their knowledge and experience, and they can help you get started with tools like an ideation workshop. Again, the idea is to think big but start small, delivering successful outcomes and encouraging colleagues to gain an interest in joining digitization efforts.
To learn more about closed-loop manufacturing, be sure to stop by the HPE booth at Hannover Messe, which takes place April 23-27. There will be examples that follow the lifecycle of a product and demonstrate how the framework can create real value with data from across the enterprise.
Closed-loop manufacturing: Lessons for leaders
- A new manufacturing framework called closed-loop manufacturing leverages the massive amounts of data generated by industrial systems to strengthen feedback loops and streamline production.
- Data silos can slow a manufacturer's response to defects. Closed-loop manufacturing can potentially break down these barriers and help companies respond to production problems more quickly.
- When it comes to breaking down data silos and improving feedback loops across a car manufacturer, communicating the vision and building support is key. Many employees will be skeptical of or even feel threatened by technology-driven change.
- The closed-loop manufacturing framework means starting small but thinking big. Unless it’s a greenfield project, no company can transform everything at once without putting production at risk
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.