If I must date myself, I cut my teeth as a product development engineer in the automotive industry in the bad old days when warranty terms were shockingly poor. Let’s examine this through the lens of my first new car which I purchased in 1984. Why was my warranty coverage so bad, what drove the significant improvements seen in the industry since that time, and how do we now validate our products to keep it that way?
Bathtub curve
Before we get to that, let’s take a look at the bathtub curve, which is familiar to all validation engineers and many others. It relates failure rate to time in service for a product line.
Early in their life product lines incur a high rate of failure. These infant mortalities tend to be manufacturing and supplier quality problems or material issues that make it to the customer. Eventually, these manifest as issues and we move on to a period called the useful life of the product. In this period, the customer experiences a lower but constant rate of failure that is often the result of the customer using the product more severely than the product was designed to handle. For example, in automotive we design and validate to survive a certain depth of water splash but not to survive a plunge into a lake.
Then eventually, as expected, the product will fail with a wear-out failure mode such as thermal or mechanical fatigue, corrosion, or insulation breakdown. This is shown on the right end of the bathtub curve as an ever-increasing rate of failure. This should happen after the part has reached the design life of the product. If this never happens or happens far beyond the design life, the product is likely over-designed making it difficult to compete in your market.
The bad old days
Now let’s look at how this curve might have been drawn for my sparkling new 1984 car. In those days, there were many quality issues requiring many visits to the dealer (trust me). The warranty period for my car was 12 months/12,000 miles. Indeed! I show this period in green on the curve. The failure rate at the end of this period was still very high. So of course the warranty period was short – no auto manufacturer could afford to warrant for longer. Who paid the price for failures after 12,000 miles? Yes, the consumer (me). It’s a blur I admit, but I had many early quality issues such as hoses, sensors and other components, electrical problems, etc.
On the right end of the curve, observe that my car’s wear-out failures began well before I reached the design life of the car, which was a whopping 100,000 miles. I recall multiple too-early wear-out failures including alternator and transmission. In 1984 if your car made it to the design life without these failures, you would consider yourself very lucky.
I was a frequent customer of the repair shop at my local dealer. I was also on a first name basis with the guy working the parts counter. Did this horror story lead me and countless others to complain and look elsewhere for our cars? Not really. This paradigm was the norm in the industry and consumers often just accepted this.
What came after: The new and improved bathtub curve
Then came the quality revolution which hit the industry in full force in the late eighties driven in large part by competitive pressures from companies such as Toyota and guidance from Dr. Deming and other quality gurus. The many engineering rigors and tools that came along with this push spilled over into vast improvements in quality and product reliability. It was an exciting time to be a product and test engineer!
An updated bathtub curve reflecting these improvements would look something like this:
Notice that the steeper declines in infant mortalities have allowed for a much longer warranty coverage. With failure rates dropping so quickly, the incremental cost to cover a longer warranty term is more affordable. Today, an automotive warranty period might be 3 years/ 36,000 miles or more. Note also a much wider useful life period with very low failure rates and a longer vehicle design life, often 15 years/ 150,000 miles or more.
What came after: Improved validation test plans
Since 1984, in addition to all the manufacturing/quality improvements that have led to extended warranty coverage, overall durability/reliability gains have been realized. Advances in design and validation practices and the reliability sciences have led to longer design lives and happier customers.
So where have we landed with our validation test plans?
- They are rigorously derived from customer requirements and specs, risk assessments, and field application data including environmental and usage profiles.
- Design verification testing is utilized early in development to reveal failure modes. This testing uses highly accelerated testing that might involve HALT, step-stress, proportional overstress testing, etc. The best verification plans give early prototypes a chance to fail quickly.
- Design/Process Validation (DV/PV) test plans meet internal and customer test specifications and reliability requirements:
- Tests are “meaningful,” that is they are rigorously correlated to field use.
- Tests are quantitatively accelerated and lead to failure modes that align with what is observed or expected in the field.
- Tests demonstrate R/C to one design life using a mix of testing to failure and success testing.
- Legacy test plans are challenged.
- During and after initial product release, much attention is given to parts in customers hands. Part returns are aggressively harvested; early warranty data is analyzed for issues and trends.
Of course, validation test strategies and plans vary significantly depending on the industry. I’d be willing to bet though that most have matured greatly over the past two or three decades alongside or as a result of the great quality and reliability revolution of the 80’s and 90’s. Since I began helping clients in other industries with their test and validation processes, I can attest there is a lot of variation in how test and validation is implemented. Some industries are just learning the latest practices and others are setting the standards for the future.
So, what is your experience? I’d love to hear about it.
Abdulrahman Alkhowaiter says
Thank you, a great article. What is missing in the analysis is an important motivator of change: Strong Competitors. Up until 1984, the American auto industry gave out a 12 month warranty for US vehicles; The same with imported European cars, and Japanese cars. That kind of short term gaurantee is not going to win many reliability awards…..
Somebody at the Japanese factories in Japan around year 1983 or 1984 decided that: “Our cars are doing great in terms of reliability, so how can we capture more US and worldwide market share by emphasizing that?”
They calculated the potential losses of a two-year 24 month warranty and found its no problem at all. This new warranty greatly increased japanese vehicle sales in the US. American manufacturers were soon forced, within one year to reply with a similar 2-year warranty. Then the warranty claims starting piling in and major losses occurred with the American vehicles; consumers loved it, but manufacturers hated this situation. The huge losses forced these manufacturers to install reliability teams to study why the H** their cars failed too early. Result was a gradual improvement in US and European automotive relaibility. This was also the time when Italian cars lost a large market share in the US market.
Just when they started breathing again, and treading water, the Japanese hit back somewhere around 1988 with a *** Three Year *** vehicle warranty complete vehicle and drivetrain to 75,000 Miles! this again caused a reduction in US vehicle production-sales, and the clever Japanese captured an even larger segment of the total sales.
This time, when US manufacturers applied a three year warranty, they lost much more money as the vehicles in general were just not designed around the concept of reliability first, unlike Honda, Toyota. More warranty losses and claims were raised and new reliability programs were initiated. Since year 2010, the American car industry went far forward in reliability and competes well in this catgeory till this day.
John E Kreucher says
Hello Mr. Alkhowaiter,
I really enjoyed reading your comment and thank you for the detailed insight you’ve added regarding the competitive pressures applied by Japanese auto makers in the mid 80’s and the effect that had on the US marketplace. Surely the 2-year and 3-year warranty escalations implemented by the Japanese were made possible (i.e. economically feasible) by the improved quality and overall reduced infant mortalities they achieved as a result of the quality/reliability initiatives they implemented in the years leading up to this period.
Your final remark about the roughly equivalent level of achieved reliability is also quite true. This is evident each year when you look at the latest JD Power survey results. You’ll find OEM’s from many regions represented with high scores in most of the categories.
Abdulrahman Alkhowaiter says
Your article is suited with some tweaks, to be in some of the top mass media magazines, a segment of their readers are interested in the topic.
As a consumer of vehicles since 1982, have been quite happy with the competition in automotive reliability over the years. At the present time own a Model 2013 3.6 L Jeep Wrangler which gets a B-grade in reliability with quality care by owner, and a KIA Mohave 2009 model six cylinder gas engine powered which has been an A+ reliability vehicle.
This is an interesting discussion on Japanese vehicle reliability:
https://www.reddit.com/r/whatcarshouldIbuy/comments/13l39i0/why_are_japanese_cars_so_much_more_reliable_and/?rdt=35951