AI Didn't Kill Your Business! COPQ Did!

The trends and advancements in the flow of information, and the enhancements in the fields of Artificial Intelligence, 3D Printing and Automation require the quality of company products to be above four sigma if the company is to survive. You have heard it before. Investing in the quality of your product is important, but it is never clear when to collect the rewards from all of that investment, or when improvements begin to yield diminishing returns.

The answer is you collect the rewards as you improve, and there is a time when the benefits from improvement wane. I'm going to show you how that works, and why being above four sigma is critical to your survival. Don't worry it isn't complicated at all, but it is really important to understand.

Before we get too far I want to make sure we have a few concepts down. We will be combining these ideas into a single model and then interpreting what it tells us.

The first concept is "The Cost of Quality". It is made up of two components.

1. The cost to produce good quality. The cost to produce good quality is made of:

2. Appraisal costs (think testing and inspection).

3. Prevention costs (think process capability & Failure Mode Prevention).

4. The cost of poor quality. The cost to deal with poor quality is associated with a company's reaction when errors occur. These costs are classified as:

5. External Failures. Failure that occur outside of the company. (think failures in the field with the customer or dealer)

6. Internal Failures. Failures that occur inside of the company walls. (Think inside of manufacturing)

For more information on the Cost of Quality follow this link. We have a fantastic graphic for you.

The second concept is "Preventing a problem is less expensive than reacting to a discovered problem". This is akin to the old saying "pay me now or pay me later". The earlier we find and deal with a problem the less expensive it is. In the product design and manufacturing space there is a multiple in cost whenever a mistake is missed and passed on to the next project phase. Catching the problem in the design phase is less expensive than finding it in the verification/validation phase, which is less expensive than finding it in the manufacturing phase, which in turn is less expensive than having the customer find it.

So the idea is by spending the money to prevent a problem in a earlier project phase, both time and money will be saved later in the process. Although every industry and product differ in cost multiples, a good rule of thumb is for each phase there is a multiple of 10 in the cost to implement failure mode counter measures. You can read more on Failure costs here.

The third concept is how failures move up-stream when a company is forced to address failures. When customer complaints become too common a company is forced to react to them. Reacting creates a wave that travels through the company. It starts with beefing up product appraisal through inspections before shipping to the customer. Once the product is stopped from reaching the customer the next step is to find who is responsible for allowing the failure to reach the end of the line. To stop the failure from being created the process is investigated and eventually a root cause is found. This allows the process to be managed in order to avoid problem creation. Soon it is realized the ongoing expense of closely managing the process could be avoided if a change was made to the design to make it more robust to the variation being controlled within the process.

This demonstrates how a problem at the customer is pushed up stream from the customer, to the end of the manufacturing line, to a specific process, to being managed within the process and eventually back to the design for correction. For a more detailed review of this concept and their respective quality cultures, you can click here.

The fourth concept is about defects. Naturally defects are something we are trying to avoid. Because we are trying to avoid them, we intend for them to happen infrequently, or better yet not at all. If they happen infrequently then we have to have a large sample of products in order to find the defects. So we talk about defects in terms of defects per million opportunities or DPMO. This DPMO value can be translated into a "Sigma" value. We won't get into all of that here. Just know there is a direct translation from one to the other. I make this translation so it will be easier to read. For this example I will be using only simple math and it is all laid out in a table and graphs for easy visualization.

Our New Product

Okay now that the background concepts are out of the way here is the single model that combines all of these concepts.

We make a light for our customers. If everything is perfect it should only costs $3 to produce. We sell it for $10. This gives us a $7 profit per light. We sell 1,000,000 lights giving the company $7,000,000 of profit. We call this our entitlement. As business owners it is what we strive for and what we expect.

In this thought experiment we will start with the Cost of Poor Quality in its extreme form and we will calculate the company profit using the concept of DPMO and Sigma Level. A few assumptions: Our company can not be sued, there is no liability to our company for poor product performance. However when we have an external failure we expect the customer will return the defective item, and we will return their money in full. With each iteration we will assume we can meet the demand of 1,000,000 lights by repairing the defects. This will eliminate the need for more inventory build up to account for the parts that failed internally. These are all huge assumptions which only reduce our true costs of poor quality and bring us to profitability much more rapidly. They simplify the calculations and explanations tremendously for this article, and allow us to visualize what each iteration of improved product quality does for profitability based on defects only. We will move through the different cultures of quality as described in the third concept above. As failures move up-stream we will apply the factor of 10 at each phase as described in the second concept. As we transition through quality cultures we will calculate the company profit with each iteration. Each iteration will assume the 80/20 rule. Only capturing 80% of the previous failures while 20% remain unsolved until the next iteration of improvements.

Starting with a very poor quality of product, lots of defects and nothing to prevent them from getting to the customer, is a company ignorant of their product. We call this a culture of ignorance. When the customer finds a defect they return the product for a full refund ($10). Once returns are high enough the company begins to react to the failures and enters the protective quality culture. Starting with a sigma value of 0.5 which is horrid! A sigma value of 0.5 means there will be 841,345 lights returned out of the 1,000,000 sold. At $10 per light the company returns $8,413,450 to dissatisfied customers. But it only made $7,000,000 in profit. So there is a loss of -$1,413,450 resulting from the culture of Ignorance. Unsustainable the company will fail.

So in the second year and iteration, reacting to the problem the company puts inspection at the end of the line before the lights ship. Inventory at the end of the line due to inspection builds. Each light is repaired and the inspections are pushed further up-stream closer to where the defects occur. Action is being taken before the customer receives the product. This protects the company and the customers. This establishes a culture of protection.

In the second iteration, inspection will only catch 80% of the problems. Twenty percent still slips through to the customer. This means 168,269 defective lights will reach the customer and be returned ($1,682,690 of warranty). For this defect rate the sigma value is 2.46 sigma. But it also costs the company money to find all of those defective lights, and repair all of the inventory before it is shipped. All 1,000,000 lights are sold but with a 10% added manufacturing cost. So... (.8 x 841345) x $1 = $673,076 in rework plus $1,682,690 in warranty = $2,355,766. Subtracting that from the $7,000,000 entitlement, there is only $4,644,234 of profit. Better! The company is now surviving but still only about 67% of the entitlement.

Let's continue to push the defects up-stream with similar iterations. I won't describe the calculations to avoid being tedious. Instead I will just reference the table and graphs to give a better picture of the costs and profit of each iteration.

Reviewing the table and graph above, the first observation is the further the defects are moved away from the customer the more profitable the company becomes. This culminates in the prevention of most defects through designs which are robust to the processes used in their manufacture. The more robust the product is the more closely it mirrors entitlement. Although I can't show you data from this thought experiment, I can tell you my experience has shown, the act of making the design robust to the processes requires the designers act creatively while remaining in tune with the customer requirements. This in turn sparks innovative ideas regarding how to better satisfy customer requirements. With fewer design fires to put out on current products, designers are focused on making the future design better which results in innovation.

The second observation is that as the company moves from 4 to 4.5 sigma and beyond there are only a few thousand dollars of profit left on the table. The profit on the product is so close to mirroring entitlement, the few failures that are left are very difficult to capture and it is easy to spend more money trying to capture the defects than the improvement projects will yield. Hence the point of diminishing returns has been reached.

Does this mean the company can't improve or that it shouldn't try? No! The answer again lies in innovation. There is still efficiency out there to capture, but not with the current process. So around the four sigma level, management needs to begin changing focus away from continuous improvement of the current process and start focusing resources on new and innovative processes that break out of the shackles of the old process. It's time to change not only the way you think, but how everyone in the company thinks, and maybe even the industry.

If the company is operating above the 4 sigma level it will have the profits to go with that reputation of quality. It will also be better prepared for industry changes or shifts.

Now in that light I would like you to consider the changes that are upon us all right now. Consider how you will leverage Artificial Intelligence, Constant Flow of Everything Digital, Mass Customization, Additive Manufacturing, Group Sourcing, Crowdfunding, Machine Assisted Decisions and Robotics everywhere? If your company is below 4 sigma, it will probably struggle to find the investment dollars to readily deal with the changes. The tendency will be to delay while those competitors who are prepared will leverage the efficiency of newer technologies, and eat into your company profits and market share.

If a companies quality levels are not above 4 sigma, I suggest they are ill prepared for the decisions and changes that will soon be thrust upon them. It's time for them to get their house in order by eliminating the defects and waste present in their systems. They should learn everything they can about their customers, product, processes, suppliers and employees so they can take advantage of the new opportunities about to be presented.

I wish you great speed in your preparation for change, and offer my services. If there is anything I can do to help, please call.

Like it? Leave a comment & feel free to share it!