Jeff Immelt’s memoir ‘Hot Seat‘ about his time at GE, where he finally rose to become its CEO, is an interesting book for importers. In one chapter, he details a disastrous product recall of refrigerators with faulty compressors and if you’re making new products, there are learnings to take from this episode.
Let’s look at the example from the book, and then I’ll give you some thoughts on what can be learned….
GE’s horrendous product recall
If you’re faced with a product recall something has gone seriously wrong during product development and/or production. With luck, it’ll only be a single batch of products that are affected, but either way, this will can be a costly and embarrassing exercise.
In 1989, GE was having all kinds of problems with a vast number of their refrigerators where their compressors would fail in a very strange way. They would basically fail as soon as the weather warmed up so, say for example in Florida, where the weather is warm in Wintertime so they’d start failing there and then as the weather started warming up across the USA towards summer the failures would gradually creep northwards all the way to New York and beyond:
After a little investigation, we determined that every single compresor inside 3.3 million refridgerators was going to fail, one by one, in a wave that was going to roll across the country…Each repair would cost us $210 – more than half what the customers had paid in the first place. It was a disaster.
It was a really strange situation where they could almost predict when and where the failures would happen, but they didn’t know what to do because there was a national shortage of compressors and every time they repaired and replaced a faulty unit with a new one it would still fail again, it just was a matter of time.
It got so bad that they got to the point where they didn’t even have any more parts to replace:
…for a while we replaced broken compressors with other faulty compressors that we knew would soon break. How’s that for demoralizing? Every unit we fixed we knew we would soon have to fix again. Not only did that anger our customers, it also put even more stress on our overworked service techs.
The cost to fix the issue was immense (remember this was 1989):
…given that there were millions of affected fridges, the company would have to take a $500 million charge – the the biggest write-off in the history of GE.
And, to make matters worse, Immelt also found himself called in front of then-Senator Al Gore for a serious dressing down, as he’d been informed that the repairs were causing the leakage of CFCs, which we now know are a particularly harmful greenhouse gas, into the atmosphere.
What went wrong?
My guess is that what probably happened is that whoever from GE sourced the compressor was working to an estimated typical maximum operating temperature for the fridge, let’s say 50 degrees C, for example. So maybe they chose a compressor that would work up to 50 degrees. That’s logical. However, if this is indeed the case what they maybe didn’t take into account is that the real operating temperature of the fridge would be higher as heat can build up inside. Add this to a warm outside temperature, and they were asking the compressor to work in conditions well outside of its specified capability, perhaps 55 or 60 degrees.
This unfortunate situation could have been avoided if GE had done several things during the design and development process.
1. They should have chosen a very well qualified and reliable component supplier
When sourcing suppliers, they should have done a thorough supplier audit to evaluate and assess whether the supplier of the compressors was able to produce them to the standard required. Not only did GE miss that the compressors were defective and likely to fail in a common scenario (when the weather was warmer), but so did the supplier. This brings us on to…
2. They should have tested this CTQ component more thoroughly
The compressor, as a critical component for GE fridges, should certainly have been tested for quality and reliability by default, especially if it was a custom part developed especially for them. Even if the supplier did their best to make the best compressor possible, GE should have taken some samples of those new components and done what they call ‘CTQ testing’ (critical to quality) putting all of those compressors through temperature and humidity testing. By doing so they would have failed during the product development process and the engineers would have warned the design team that this component is problematic. This clearly didn’t happen.
3. They didn’t do satisfactory ORT
One more test GE could have done that would have caught the defective compressors is ongoing reliability testing or ORT after the first manufacturing run. If they missed it during the development, as actually happened, then ORT which is more of a screening test could have caught it after production and a course correction could have been implemented before even more fridges were manufactured that were fated to suffer from the same problem. However, if they had done any kind of product level testing they would have still caught the issue during product development which is always preferable to finding it after production has started.
So, by undertaking fairly common testing on the compressor like HALT and temperature and humidity testing on the component and complete product, GE could have saved themselves $500 million (which must be worth several billion dollars in today’s money) and a lot of embarrassment. A similar modern example of a disaster caused by deficient quality and reliability testing would be the Samsung Galaxy Note 7 scandal where the smartphones caught fire and posed a risk to users. This cost Samsung billions of dollars and severely damaged the business despite its size.
What can we learn, then?
The lesson here is that it’s essential that you investigate and test what sort of catastrophic failures could occur in order to make a reliable product. In GE’s case, reliability would be choosing a CTQ component that has a reliability margin that takes into account varying environmental factors that affect the operating temperature of the part.
In reliability engineering, we have a saying:
“Pay me now or pay me later.”
Pay me now means give me the testing budget and provide me with the samples that I need to test this product or we will have a high return rate.
Pay me later means that if we don’t do the reliability testing in advance you’ll be spending a lot of money, maybe double or triple on the returns after the product has hit the market.
How the testing works in practice
Let’s look at the example of when we develop and manufacture products for our customers at Agilian.
When we design a product we normally have several stages of testing going on. First of all, during the design and development, we do component testing and supplier audits to make sure all the components are going to be highly qualified, reliable, and of high quality. Components and suppliers must adhere to our customers’ requirements and testing requirements.
The next thing we do is make sure that all product-level designs go through the end-user requirement reliability testing as a minimum. That’s how the product is going to be used in the field by end-users.
Then at the same time, we also make sure that we design and build the product and test to the customer requirements. For example, some customers want their products to have a 10-year life, meaning that the product should operate without failure for 10 years. So we must design our products to meet this requirement.
Western companies like us tend to follow a similar EVT, DVT, PVT process during product development during which you are likely to have at least two or three reliability testing revisions (perhaps around one per phase).
To give you a clearer idea of where product testing fits into the new product introduction process, take a look at this video from our parent company Sofeast:
This testing is really critical as it allows you to actually see that the product is maturing from one build to the next. When the product is maturing that means that during the first build you’re going to have the most failures, typically critical failures.
Once those are fixed you move on to the next build and are taking a larger number of samples for testing. You’re still seeing some failures, but hopefully, if they have done a good job of fixing the issues found there shouldn’t be any major failures, rather you’re going to see smaller less critical failures.
Then by the time you get to the third pre-production build you’re going to see few product and component failures. It’s more likely that you’ll find problems like firmware failures and issues with painting or coating, but nothing as serious as the defective compressor that caused GE’s massive recall.
Conclusion
No matter what kind of a part you use for whichever operation, you must always have a margin of reliability so that if there is an environmental change from cold to warm and humid, or even if there is an electrical change such as a voltage spike, or there is a current increase, you should have a margin of reliability within the component so it can handle those variations.
When you design a product make sure that there is ample tolerance for variations so the product can still operate even if the environment or temperature changes. GE possibly didn’t do this on this occasion.
The takeaway for future entrepreneurs and whoever’s in the middle of designing a brand new product is to make sure to plan for reliability testing. Do you have some kind of product development life cycle where you can measure product reliability from one build to the next? Also make sure you have a reliability goal, such as how long you want a specific component to last. Will it be one year, two years, or just for a warranty limit?
Even the largest companies like GE have been caught out by not designing for reliability, and if you don’t design for reliability there is a very good chance that you’re going to incur production failures which, if not caught, will result in field failures and possible recall later on.
Get help to develop and manufacture your new product
Having read this are you feeling like you might need help to source suppliers and components, set up your testing plan, and bring your product to market with the minimum of risks?
Well, we do all of this in-house at Agilian with our engineers and designers who ‘become a part of your team.’ Get in touch with us for a free discussion about your project. There are no strings attached and we keep any information about your product confidential and are happy to sign an NDA.
