Every step of the production process is just as essential as the next and as such, each procedure leaves behind increasing amounts of data all throughout manufacturing. The value of product data grows when it’s captured across the product’s life, from early design through development and onto its life with the end user. It can be gathered from suppliers and distributors with the aim of capturing enough data to identify and correct problems at the earliest possible moment.

EEIOL 2016JUN09 MPU NT 06 01 Figure 1: *Big data flow: This diagram shows the pathways of the data collected on a product. *

The ability to capture and manipulate data is changing the way products are designed, manufactured, and maintained when it’s with the customer. “Big data offers insight into the true product performance, from the time it’s designed, through manufacturing, out into distribution, and into the hands of the consumer,” said Bill Boswell, senior director for cloud services at Siemens PLM. “Complexity just gets deeper and deeper. You have suppliers, maybe contract manufacturers, your suppliers have suppliers, and then the product gets shipped out to stores and to millions of consumers.

Big data prompts new questions: How do you get a big picture view of the product across the whole lifecycle? How does quality affect the development of the product and its use by customers? “PLM doesn’t stop at the loading dock door,” said Boswell. “We need to know what’s happening with the product in the field.”

The data is stored in silos across companies and suppliers as well as with the people doing service and support. Many companies are not even aware of their own data. “Gartner has a term for the data that companies have but don’t know they have: It’s called dark data,” said Boswell. “It’s data that was collected and used for one purpose, and after it’s been used, it goes on the shelf.”

Boswell noted that companies need access to their dark data in order to manage their products effectively. “There’s all this data in the different systems that was collected for one purpose,” he said. “If you can bring that data in from IoT, CRM, warrantee, and repair, and if you can contextualise it and relate back to the product, you can start to answer questions you didn’t think to ask, such as: What is my product doing in the field? What is this product costing me in recalls?”

Tracking down product flaws

A bump in recalls and repairs can indicate a problem in the testing and manufacturing of the product. Having access to the data can help solve the problem. “When a company in electronics is manufacturing boards, there is all kinds of test data that indicate things are working properly,” said Boswell. “The test data is delivered with the end batch of materials. It’s valuable validation data.”

EEIOL 2016JUN09 MPU NT 06 02 Figure 2: Big data explosion: Over the next few years the amount of data collected on products will expand dramatically.

Siemens was working with Dell when a high volume of application crashes were being reported. Dell had to find out why the applications were failing. Was it the firmware, the computer, or the application software? They had to find out the source of the problem. “We discovered that memory lines were the problem,” said Boswell. “We pulled the validation from that supplier, and we could see that the memory lines were not put in correctly in the first place.”

Part of the process of building the big-data picture is to capture data from suppliers. This is becoming an issue in automotive as carmakers begin to require production data from their suppliers. “You want the supplier to put the info in the system,” said Boswell. “That varies by industry. Big data can change the conversation with the suppliers. If they’re putting their data into the system, you can see when there are problems. Takes a lot of the guesswork out when you’re looking at data that is based on fact.”