Last winter I was half way through a bike ride and went over a small bump in the road. Looking down, on my handlebars the (not-inexpensive) bike computer that I use to track my progress suddenly cut out. I rode home in a grump. 13 months after purchasing and one month after the warranty period was over, my computer was broken. Typical!
Being an engineer I took it apart. Immediately it seemed clear that the design team had made some compromises in order to make a more compact product and no doubt save some pennies in the process. It turned out that plugging the charging cable in had weakened the connection between the battery and the main board. This meant a slight bump would turn the computer off, and spoil my ride!
I reasoned that it had been cleverly designed so that the connection only broke after 13 months and forced the consumer to buy a new product. A quick google search only confirmed my suspicions, with plenty of loud voices shouting their dismay that their expensive computer had broken just after the warranty, how devious …
But was it a deliberate design decision? I asked around fellow cyclists, many had the same model, and many had examples much older than mine. No one had suffered the same problem. My experience, whilst not unique, was far from universal. In a large group of friends’ I was the only one who suffered a broken unit. Whilst the device had a weak spot, there was no malicious plan from the manufacturer to make consumers buy another product after 13 months.
My slightly blinkered review of information online had, at that point, confirmed my gut feeling that the product was designed to fail at the end of the warranty period. In truth, for any popular consumer device, there will undoubtedly be failures. It’s a case of luck and your usage of the product. Perhaps I had been slightly rougher with my bike computer than my Lycra clad friends, or maybe I’d trained harder and gone on more rides! Or maybe I had just been unlucky.
So, what does this mean for engineers looking at failure data? Well, when a product fails, it’s tempting to complain long and loudly. Products that go on working faultlessly, on the other hand, generate little attention. Additionally, when presented with a failure it’s tempting to try and fit patterns to data and make sense of this.
In my case I came across a failure, came to a conclusion that the part was deliberately designed weakly, and then searched out data which was likely to confirm my bias. However, when it comes to fairly evaluating the reliability of a product, it’s vitally important to take into account the quiet, un-showy units that are just getting on with the job!
Jeremy Monk, Engineer & Cyclist