What are the best data analysis techniques for reducing manufacturing waste and defects?

Pareto Analysis is one of the most powerful tools in all of business -- not just for Quality and Process Improvement. The reason is simple. It allows you to identify the areas where you can focus your limited resources and get the "biggest bang for the buck." A common mistake is to assign so many activities to the "workers," that they can't spend adequate time on any of them. As a result, completion time for all issues increases. By identifying the most important issues and focusing resources on those, you get a more immediate payback. And, it doesn't mean sacrificing activity on the lower priority items -- just sequencing the lower priority start times to a later date (while reaping the benefits of fixing the important items first).

Ask the operators that own this task. You may get some off the wall answers but as you drill down into them you will likely find very useful information that is not clearly evident to a casual observer.

Statistical Process Control (SPC) is crucial in the food industry, ensuring consistent product quality through a systematic, data-driven approach. Using control charts, SPC monitors key parameters, distinguishing normal fluctuations from special causes that might impact food quality. Vital for compliance with safety standards, SPC enables manufacturers to identify and address issues, reducing defects and maintaining high product quality in a dynamic industry.

I like to use SMED (Single Minute Exchange of Dies), by video recording the process whether it being the problem area for defect concern or just an area that was earmarked for any type of improvement. Then using a Multi Disciplined Cross Functional Team we access the the Video, through which all give their ideas or comments. Allot of times we tend to get "Factory Blind" and might overlook the simple things and through this method I have had many of success, another person might notice the small thing that might make the biggest change and might just be that Break Through needed in that area or Process.

I agree with John Bell's article. However, there may be times where the number of factors are too numerous. In these cases group factors according to the major process groupings such as procedure based, equipment issues, raw materials, training, etc. From there a secondary pareto Analysis may help to focus the corrective actions.

The tools mentioned here all work and work well!! Fishbone's are the easiest to teach and coach. I use a version that asks at the bottom for the most likely 3 to continue analysis, proof. Fault trees and 5 why work well, a little more complicted. I learned at General Motors the 3X5 Why. You do 5 why;s 3 times. One for the process that produced the non-conformance. Then 5 why's for the measurement or detection that missed it in the process. AND then, 5 why the management process that missed a process control or measure/detection.

Root cause analysis seeks to understand the cause of a failure with a view to preventing a repeat or reoccurrence. The Ishikawa/fishbone diagram is one of the very useful tools for analysis. The use of the 5 “WH” questions is also popular. Asking the right questions helps understand the source of a problem and ultimately show to correct or prevent such.

The use of 5 whys is very useful in the search of the root cause of a problem / opportunity. Tools such as KT Analysis help as well, now in my experience these tools will give you a brought idea of where the opportunity / problem is. After you identify them it is helpful to leverage experimentation and test the hypotheses to make sure we can turn on and off the problem. By doing this we can then work on the corrective and preventive actions / systems to eliminate the problem.

Root cause analysis is a strong tool. My recommendation is that you make it an iterative process. The solutions are dependent on the participants the facilitator. The results are not an absolute but a step forward.

In my point of view, through more than 20 years working on process engineering, Statistical Process Control is extremely useful for waste reduction, it allows to identify and address quality issues early on, helping minimize production of defective products, saving time, materials, and money. Improved efficiency: Also by monitoring process behavior, SPC allows for proactive adjustments, preventing disruptions and ensuring smooth operation.

I have seen and found SPC to be useful in tracking real-time data in manufacturing. If set up properly, you can actually use the data tracking to prevent rejects from occurring if certain results start drifting towards the control limits.

For me, this is a tremendous tool to the process support. SPC helps me understand many variables at the same time, and not only to focus on the result, but also to be able to train and establish debate about what should be done and what should be improved.

For process, mathematical modelling and predictions of output helped us to increase significantly our performance. When you have sufficient data to work on and a knowledge of how your process woks, it is possible to make a simulation of it using an algorithm. It help you predict some values based on machine learning. You can test this values on the process and adjust them to have an optimal point of functioning.

O controle estatístico de processo, proporciona uma série de benefícios, incluindo identifica??o eficiente de problemas, melhoria contínua da qualidade, aumento da eficiência operacional, redu??o de custos e aumento da satisfa??o do cliente.

Crafting a robust framework for industrial experiences on the shop floor involves a blend of structured learning, hands-on exposure, and mentorship. Designing rotational programs, on-the-job training, and cross-functional projects allows employees to gain comprehensive insights into various processes. Pairing novices with seasoned professionals fosters knowledge transfer and skill development. Moreover, encouraging a culture that values curiosity, experimentation, and continuous learning ensures that every shop floor experience becomes a stepping stone towards expertise and innovation.

DoE is not exactly a data analysis tool to reduce the losses. It is a technique to find the best parameters for your production system to improve quality or scrap rates.

Planejamento de experimentos oferece benefícios significativos, como a identifica??o precisa de fatores de influência, otimiza??o de processos, redu??o de variabilidade, economia de tempo e recursos, e aprimoramento contínuo da qualidade do produto.

In the description it states Lean focuses non eliminating any activities that do not add value to the Customer. This is not entirely true. Lean categorizes work activities as Value added, Auxiliary and Waste. Lean actual focus is to eliminate Waste, minimize Auxiliary (steps necessary to complete Value added) Ultimately to increase value added to replace waste and excess auxiliary that was eliminated. Auxiliary examples could be walking too far or picking up a tool multiple times. Seeing Auxiliary opportunities requires time spent in the Gemba observing the process.

The lean Six Sigma technique is a tool to enhance efficiency and effectiveness. Lean focuses on efficiency as an attempt to remove redundancies in the process or activities that do not significantly add value to the process. In other words, using the least necessary inputs in generating the desired outputs. Six Sigma on the other hand focuses on effectiveness and the ability to accurately replicate the process and quality- Effectiveness.

One tool I’ve found helpful is the SIPOC tool combined with the three C’s. Following up on Suppliers, Inputs, Processes, Outputs, and Customers with the questions in mind around does it Change the thing? Does the paying customer Care? Is it correct the first time? Then ea h of the SIPOC elements can be evaluated for value add or drop or reduce the cost.

Choosing the right data analysis technique for reducing waste and defects in manufacturing depends on various factors such as the nature of the manufacturing process, the type of waste or defects being addressed, the complexity of the processes involved, and the resources available for analysis.

For reducing manufacturing waste and defects, a combination of techniques is often effective. Lean Six Sigma integrates several methods, including Pareto analysis, root cause analysis, and statistical process control. Design of experiments can be beneficial for optimizing processes. So, a holistic approach combining these methods could yield the best results.

To reduce manufacturing waste and defects follow three work tracks. 1: Bottom up E2E cost and loss approach to bring all cost and loss visible in the system followed by a zero-loss journey. 2: Benchmarking exercise to compare performance with best-in-class players to have a clear visibility of gap and decide from-to level and glide path. 3: Envision perfect scenario what I call ideal state and have a journey from current state (#1) to future state (#2) to ideal state (#3)

In my point of view the first to understand are the nature of the manufacturing defects/waste. Is it sporadic or chronic? If sporadic, we need to know the Root Causes, so using RCA will be effective. If chronic, and have sufficient data to analyze, PARETO Analysis is the starting point to focus your resources in solving the main contributors. The most important is to validate all the identified issues/root causes in the shopfloor and understand what are the main activities (e.g. cleaning, change-over, start-up, maintenance, etc) of the operators, that might possibly contributing to the issues, check if there are standards available(cleaning, inspection, etc)to ensure the consistency of execution and they are following the standards

I'm with Dan Reid. We used to say that your first actions should be to "plot the data, plot the data, and plot the data". Scatter plots, histograms, run charts. Depending upon the issues being addressed, go from there applying the tools appropriate for the problem. "Deep end of the pool" statistics and Six Sigma are often overkill ... But sometimes the only way to truly solve it. Regardless of approach, the other key is to involve all who may be able to contribute including "outside eyes". Don't try to be the hero and operate solo.

Ishikawa said most quality problems can be solved using 7 basic quality tools: pareto, check sheets, graphs, histograms, cause-effect diagrams, scatter diagrams and control charts. The complexity of the problem will dictate what tools are needed. Complex problems may require more advanced tools, e. g. DoE, Six Sigma, ANOVA.

The integration of new Quality Control Circle (QCC) tools is revolutionizing the industry's quality landscape. These innovative tools offer precision, efficiency, and real-time insights, elevating quality standards to unprecedented levels. From AI-driven analytics to IoT-enabled monitoring, they streamline processes, identify issues early, and facilitate data-driven decision-making. Embracing these tools isn't just about enhancement; it's about staying ahead in an ever-evolving industry, ensuring excellence at every step.