top of page
Search
Chris Butterworth

Problem Solving in Manufacturing

In manufacturing, much of the profits are eaten up by quality problems.





The Cost of Poor Quality (COPQ), defects found on-site as well as defects found by the customer, consumes far too great a share of a company's profit. These costs are larger than one would consider at first glance. Defective parts have a unit cost associated (e.g. bill of materials value), but their true cost is a lot higher. These parts have to be identified, segregated, reported, moved around a few times and eventually scrapped. Having missed a few items on the production order may also require additional production time to replace those defective units and likely overtime hours.


If the defective units escape and end up in the customer's facility, the costs will include a LOT of your management time and attention. Documented costs will include costs of replacement, overnight courier, management travel to customer location, and more. Undocumented costs will include the hard-to-measure effect on your reputation.


The customer will bare some costs that you may not see. It could be that your product (e.g. small electronic assembly) caused their product (e.g. consumer product) to fail and their costs could be much higher than what you see. You probably won't have to absorb all those costs but you can be sure that the customer will start looking at your competitors more favourably. They might drop you altogether.


When you consider all these costs associated with producing poor quality, you see the value in problem solving. There is no better return on investment than developing problem-solving skills among your workforce.


It would be great if we could prevent the failures from happening in the first place. We do try that with some tools but they aren't completely effective. There are thousands of things that can go wrong but only a handful of them end up causing defects. For the problems we weren't able to prevent, we need to resolve them after they first appear.


Consider this scenario: a company has observed a few defective items and this triggered a Nonconformance and Corrective Action process. They start by collecting a team of people to begin brainstorming for root causes. But there's a problem right out of the gate here. There is a lot of information contained within the company's current data set that has to be reviewed first. That data needs to be collected and many charts have to be generated.


As obvious as it sounds, it doesn't happen that way often enough. The team starts brainstorming as though the problem was completely random, which it seems like at first. When something occurs at random, it's pretty tough to diagnose. With many ideas generated during brainstorming, the workload to investigate each idea on the chalkboard is overwhelming. From that point, the problem is seen as much more complex than it actually is.


During initial data collection, and after the brainstorming has generated over a dozen candidates, an important piece of data shows up that changes everything. The important data point is that the defect was not random. Here are three examples of mine that exhibit the same problem solving tip: collect all the data you currently have and look for structure in that data set.


- in an assembly, the defective component was always the first one in the assembly, never one of 96 other parts. That was news to the team.


- in a cutting process, it was always the first or last piece, never the 2nd, 3rd or 4th. This information was not known for weeks.

- products failed at final test but it was later learned that, at an upstream process, all the failures came from just two out of the 12 stations. This was very insightful and quickly led to the root cause.


These three examples show the common error made when investigating a root cause. Like a detective searching for clues, you have to look everywhere. There's information that you don't currently possess and it is likely not far away. The three examples above changed the focus of the investigations. They also did something even more valuable. They generated knowledge about their processes that the team did not possess before. New knowledge is incremental innovation.


In the first example, the team learned that, although every component in the assembly is presented/delivered in the same manner, there must be something different about the first item. That fact provided needed focus for improved brainstorming.


The second example was a cutting process where a frozen food product was presented to a cutting knife. The items were weighed later in the process and the order of cutting was unknown. As a random defect, root causes focused on variability in the food preparation and the distribution of food items. But this new information that showed only the first and last were over/under weight changed the focus to the cutting process.



In the third example, the company learned that their 12 workstations exhibited significant differences from one another with two stations producing unacceptable product that failed a subsequent test.


In these three cases it was existing data from the factory floor that provided the key insights to focus of our efforts. Not brainstorming.


Collect more data, plot everything you have and look for patterns that stand out. If a chart shows an unexpected pattern (trend, shift, cluster, etc), now you have something to brainstorm over. Dig deeper and find out if this pattern helps to connect the dots between the defects and root cause.

5 views0 comments

Recent Posts

See All

Comments


bottom of page