Your guide to Big Analytics and Big Data in manufacturing

Welcome back, Troubleshooters! For the past few weeks, we’ve been writing about how manufacturers can manage the massive amounts of manufacturing data that are quickly becoming available to them and find the meaningful signals in the noise. Mainly, that has to involve a data strategy for the next decade or more. It also involves turning raw data into useful information that can help manufacturers take advantage of efficiencies and increase profits. Today, in the final part of this series, we’re taking a deeper dive into how Big Analytics plays a major role in extracting the value from Big Data.

What is Big Data?

Big data refers to the growing streams of data available to manufacturers—vast data sets that may be analyzed using advanced computing capabilities to uncover useful associations, patterns, and trends.

The data itself often already exists but is not usable because it is stored in separate business units (silos) within an enterprise, often in incompatible formats. The streams may come from machine sensors, routers, processes, firewalls, applications, historical databases, or other sources. It takes Big Analytics to take it all in, convert it to useable formats, analyze it, and reveal valuable insights.

Big Analytics: the data management platform

Big Analytics simply refers to an overarching data management platform that can tap into all of a business’s operations, and gain access to the silos of data stored in various shared file systems, databases, IoT data aggregators, and historical, financial, and geospatial data feeds. A data management platform can help managers find the root causes of problems by ingesting large volumes of information and examining operational processes in real time to discover, for example, why a shipment is missing or why a product is defective, so the problem can be solved right away.

A data management platform can correlate current information with historical databases, and apply machine learning to find patterns in the data and understand the causes and effects. For instance, on the production line, they can find equipment or process failures; in the quality assurance arena, they can determine the cause of component failures; in operations, they can discover conflicts in new processes or components. Executives using these systems report fewer equipment breakdowns and unscheduled downtime, fewer instances of unplanned maintenance, drops in supply chain management issues, safety incidents, and more.

Sometimes these platforms have the ability to hold a year’s worth of data at the ready for instant comparisons and analysis. Such a platform must be cloud-based because it requires enormous computing resources.

Big Analytics in action

Manufacturers, wrap your heads around this: Big Analytics can predict product failures before customers experience them. By analyzing and comparing a number of factors (e.g., the particular production line, the batch size, the day and month the product was made, the number of engineering changes it went through, consumer usage patterns and demographics, etc.) it can tell manufacturers which lot is most likely to fail and where the problem is, even if it’s half a world away.

This capability has huge implications for big problems such as warranty costs. It’s well known that auto manufacturers spend 2% to 3% of annual revenues paying warranty claims. In 2017, GM paid out a net $3.13 billion, and Ford paid nearly $3.46 billion. If you consider the intangibles such as a tarnished brand image, disintegrating customer loyalty, and legal liability, the costs might be even higher.

Traditionally, when a defective model hits the market and the warranty claims start flooding in, it has taken between 90 to 120 days for the claims to work their way through the system and alert the automaker to the problem. That’s three or four months of the defective car continuing to roll off the assembly line and into consumers’ garages, and countless more warranty claims to pay out.

To address this need, a number of business analytics firms are offering warranty analytics. For example, Ernst & Young, SAP, and Hortonworks combined EY’s analytics, SAP’s technology, and Hortonworks’ open source platform to deliver advanced analysis of warranty data that can anticipate product failures so they can be addressed proactively during the design phase and reduce warranty claims. It also monitors warranty claims in order to detect fraud. Hewlett Packard Enterprise (which also offers warranty analytics) reports that in at least one instance, they helped an automaker proactively resolve an issue and avoid a potential $1 billion recall. Elder Research, a data science and predictive analytics firm, was able to save a client $67 million over a 5-year period by using warranty analytics to detect service provider and warranty fraud.

And there you go, Troubleshooters! That’s wrap on our series about managing Big Data and leveraging it to find the efficiencies that will make your enterprise more competitive as the digital age rolls on.

Follow Troubleshooting Thursdays

Tune in to Troubleshooting Thursdays for reliable tips, general troubleshooting process, and industry insights. Stay up to date with Simutech Multimedia:

Have a subject you would like Troubleshooting Thursdays to cover? Send Simutech Multimedia an email at [email protected].

Looking to give simulation learning a try? Get started with our award-winning first solution here: Get Demo.

Get started with your demo

Share This