Quality Engineering: a Data Science Analysis.
It’s nothing new that large industries invest lots and lots of money in innovation and quality to have increasingly competitive products, at the lowest cost and that meet their consumers’ expectations.
It is important to clarify that investing in quality does not mean increasing investment in the quality sector of an industry. Investing in quality means looking at the entire production system from the perspective of what will or will not add value to the final product to meet the expectations of the target audience. And this ranges from paying your employees well so that they work happily, carrying out customer satisfaction surveys or even investing in more technological machinery, in short, everything impacts quality.
All these factors that make up the production system have a monetary cost x benefit ratio for quality. There is a science behind the ideal way to direct capital in the production system to have a product with maximum quality and the lowest cost, and we can use a little data analysis and regression to reach this optimal point.
Case study.
A general collection of quarterly data was carried out in an industry about the amounts spent on all its costs over the course of a year. The industry had 22 different costs, ranging from legal costs, machine maintenance costs, employee training, to labor costs, raw materials, etc.
Current scenario
Working with this vast number of variables would be extremely complicated for statistical analysis. So, initially, all these costs must be categorized into four types of quality costs: Prevention Costs, Appraisal Costs, Internal Failure Costs and External Failure Costs.
Figure 1 shows the scatter plot that gives us an overview of the collected and categorized data, where on the abscissa there are descriptions of the costs of the production system and on the ordinate the respective monetary costs spent. It is noted that, even though the company has more prevention costs, there are 2 outliers — Warranty Claims and Reworks — which shows us that perhaps the company is not managing its quality costs in the best possible way.
Figure 2 shows us the monetary amount spent on the 4 types of quality costs throughout the collected period. It can be noted that the company directs 76% of the total monetary amount to costs of poor quality (Internal Failure Costs and External Failure Costs — COPQ) and 24% to costs of quality (Prevention Costs and Appraisal Costs — COQ). This shows an inadequacy in the targeting of quality costs.
Analysis
One has to think that perhaps there is a correlation between the variables cost of poor quality (COPQ) and cost of quality (COQ). In fact, figure 3 shows us the mathematical relationship between these two variables, according to authors in the field. These are two exponential curves, due precisely to the natural variability of quality response to investment in COQ and COPQ, that is, the marginal cost is increasingly higher the more you want to increase product quality. In other words, the more you invest in quality, the more difficult it will be to increase quality.
The failure cost curve (COPQ) has a greater curvature than the assessment and prevention cost curve (COQ) due precisely to the fact that, when there is a small increase in investment in COQ, there will be a much more pronounced fall at COPQ. Some authors measure that by investing $1 in COQ, the company saves $10 in COPQ. In other words, if you take precautions, you won’t have to deal with the consequences.
This is the key to increasing quality and decreasing total costs: investing more in COQ. However, investment in COQ should not be indiscriminate, because as mentioned, increasing quality becomes increasingly difficult the more quality you have, so you must, in fact, find the optimal balance between quality and total costs.
Projection and results
To achieve this objective, exponential regression was used to estimate the COQ and COPQ curves for the collected data, where it was conjectured that the company operated with a quality level of 60%. The result, as well as the current scenario and the projected scenario, can be seen in figure 4.
The projection shows that, once the company allocated $863,000.00 to COQ (instead of $530,000.00), its total cost would be $1,248,000.00 (instead of $1,782,000.00), due to the drop in COPQ from $1,251,000.00 to $385,000.00. In summary, as compared in figure 5, there would be a 40% drop in the company’s total costs and an increase in quality of 17 percentage points.
Conclusion
The increase in quality generated would have a positive ripple effect for the company: higher quality products would reach the hands of customers, thus resulting in a low warranty claim rate and greater customer satisfaction, which would result in a positive perception regarding the brand image and consequently an increase in the number of sales of the company’s products.
I hope this article has somehow provided you with some value. Feel free to leave suggestions or comments below or find me on Linkedin. The code used to generate the analyzes discussed here can be accessed through my Github.
Emerson Santos
Do you identify as Latinx and are working in artificial intelligence or know someone who is Latinx and is working in artificial intelligence?
- Get listed on our directory and become a member of our member’s forum: https://forum.latinxinai.org/
- Become a writer for the LatinX in AI Publication by emailing us at publication@latinxinai.org
- Learn more on our website: http://www.latinxinai.org/
Don’t forget to hit the 👏 below to help support our community — it means a lot!
