Only a select few automakers currently use cognitive predictive analytics to prevent recalls. This could be costing them millions of dollars per year.
Manufacturers of all types of products are prone to recalls. No company is safe, from toy manufacturers—1 million Hasbro Easy Bake Ovens pulled off the shelves in 2007 after several hundred children suffered burns using them—to consumer electronics manufacturers—Samsung Galaxy Note 7 with its battery fires costing the company about $2 billion.
But no one beats the automobile companies when it comes to recalls. Just last week in fact, Federal regulators inquired as to why it took Hyundai and Kia 18 months to recall nearly 1.2 million vehicles that may have the same engine defect that resulted in an earlier recall of 470,000 sedans.
It’s just the latest in a relentless pattern of high visibility incidents that touch virtually every automaker, resulting in a record 62 million vehicles being recalled in 2014 at a cost of about $9 billion. Such rampant failures pose a number of questions that beg to be answered:
- Why was nothing done to pre-empt these issues?
- Why are recalls becoming more and more commonplace in the industry?
- Can owners really trust that their vehicles are safe and secure considering the amount of safety features and technology loaded into each automobile?
In most recalls, the finger can be pointed at human error. While manufacturing processes are well monitored and regularly maintained, and undergo strict quality checks, “unnoticed manufacturing defects” lead to these massive recalls. There is simply no amount of manual analysis that can address this. People are adept at detecting macro patterns, but not micro anomalies or anti-patterns (tiny changes that lead to future defects that go undetected in the quality check process). This is where the need for a cognitive predictive maintenance system arises—one which takes a machine-first approach to predictive analytics for detecting issues much ahead of time so they can be prevented.
The Value of Predictive Analytics
According to Deloitte, only 8% of auto executives use predictive analytics to help prevent, prepare for and manage recalls. Yet analytics and predictive maintenance can be deployed in two distinct points in a vehicle’s lifecycle that could dramatically impact the recall trajectory. Such technology deployed during the manufacturing process and while the vehicle is out in the field could by some estimates save the industry as much as $50 billion per year.
While still in the factory, cognitive predictive maintenance could help identify in-line defects, correct them before the vehicle is delivered to the customer and provide information about defects to industrial engineers who can fix the processes that caused the defects in the first place.
The results: reduction of warranty costs and risk of recalls. For example, if a defect costs $100 to fix, and it occurs 500 times a day on a line that runs 340 days per year, the annual cost of that defect is $17 million. Even a 25% reduction would save the manufacturer $4.25 million per year. And this is just for one defect!
And once in the field, the cognitive predictive maintenance model can be derived based on multiple data sources, like data collected from connected vehicles and service records, test data on parts that have been replaced and even social media. This data can be used to identify and solve problems faster,to either avoid a recall altogether or at least initiate it sooner.
The result: according to a leading original equipment manufacturer, implementing a recall sooner can be worth more than $1 million per day.
According to McKinsey, predictive maintenance will help companies save $630 billion by 2025. To see how you can get in on some of these savings, visit www.datarpm.com.To get a more in depth look at the impact predictive analytics and predictive maintenance could have on the automobile industry, read the DataRPM report: Recalls in the Automotive Industry: Creating a Million Dollar Dent.
Ruban is the Co-Founder and Chief Product & Analytics officer at DataRPM (acquired by Progress) where he leads product and the data science for the flagship Cognitive Predictive Maintenance product which solves the complex business problems of minimizing asset-failures / unplanned-downtimes and maximizing yield/efficiency/quality in Industrial IoT. Ruban is a serial entrepreneur and technologist with rich and diverse experience in machine learning, natural language question answering, data science, product, technology and business. He holds multiple patents.