Sooner or later, whatever brand of car you drive, you’ll probably get a recall. According to Automotive Recalls for Consumers, there were 14 cars on the list as of 10 May 2010 … some higher profile than others.
Not surprising given the complexity of cars these days – there can be over 14,000 parts in a single car. And we all know you need a computer technician as much as a mechanic to keep them in tune. The same can be said for washing machines and a whole host of other appliances we use day in, day out.
How a manufacturer responds to the need for a recall, and how fast has a direct impact on profit, reputation and future earnings. In the case of one of those higher profile recalls, the sum of $2bn has been quoted merely for Q1.
So how do manufacturers diagnose problems?
There are 2 basic approaches:
The first analyzes the production line using statistical process controls – effectively testing quality to ensure that products fall within acceptable pre-determined levels. The idea here is to put quality first – stopping ‘poor’ products from ever leaving the factory. Companies like W.L. Gore and Associates take this even further by designing and analyzing a series of experiments to improve not only the product, but the processes involved during manufacture.
The second analyzes field repair information. Honda for example, continually analyze warranty claims, tech data, customer feedback and other indicators that could alert them to potential problems. When something gets flagged, they can alert engineering, manufacturing or even the dealers’ repair shops immediately. If they can fix issues before more customers experience them, satisfaction improves and with it, the probability of repeat purchases.
Analyzing warranty information can be time consuming. Sub-Zero and Wolf appliances note that it can take, on average 6 months to a year to identify and address issues. Data acquisition, manipulation and classification takes up a large portion of that time – particularly when you have to go through lots of text captured in call reports. By leveraging the latest text mining technology in this area, Sub-Zero and Wolf have cut that time in half, reduced resource requirements by nearly 4,000 work hours per year and improved accuracy.
What other approaches have you seen to improve quality?
