How to Use the DMAIC Framework for Lean Manufacturing
Many manufacturing companies are facing pressures to advance their manufacturing processes to increase productivity outputs and adapt to global challenges. Lean manufacturing is one of the most effective ways for companies to address these challenges as it allows for businesses to operate with smaller workforces and maintain higher efficiencies.
Ultimately, lean aims to apply a systematic approach for long-term business improvement.
This blog is part two of a series that aims to address the challenges facing the Canadian manufacturing sector and will provide an in-depth analysis of the DMAIC methodology, a crucial scientific process for lean manufacturing. If you are unfamiliar with the concept of lean manufacturing, please read “Lean Manufacturing heading into 2023: A General Review”.
This article will dive into the DMAIC methodology and how it can be leveraged to address critical challenges facing the manufacturing sector today:
Introduction to DMAIC Methodology
DMAIC principles are largely based on the scientific process, including developing a hypothesis, running tests, analyzing results, and making conclusions based on sound data. When using a change management process following these concepts, businesses can drive greater certainty that their changes will lead to sustained benefits.
The DMAIC methodology uses five stages to deliver successful outcomes:
- D – Define: Define the problem preventing success, a project plan, and goals of the activity.
- M – Measure: Define the project scope, performance specifications, and starting measurements.
- A – Analyze: Define performance objectives and compare to the measured results to understand processes where inefficiencies exist.
- I – Improve: Using methods most appropriate to the organization, implement changes and keep only those leading to repeatable improvements.
- C – Control: Establish standards and process controls to ensure quality is maintained.
While it’s important to understand how each stage of the process acts independently, it’s even more critical to use them continuously to ensure operational improvements become part of the culture.
Consider how to adapt these steps to your organization:
Leadership teams should start lean/Six Sigma projects by determining the types of projects that will be most beneficial. During the early stages of any Six Sigma implementation, management teams should consider asking questions such as:
- Who are the business’ customers?
- What do customers want and expect from the business?
- How do customers evaluate quality?
- What is a defect?
- How should progress be measured?
There are a few ways a company can begin to answer these questions. Some of the methods include:
- Voice of the Customer: Involves asking or observing clients to identify important service qualities. Common methods include feedback discussions or interviews, surveys, focus groups, or complaint logs.
- Critical to Quality Elements: Any characteristic of a product, service, or process that a typical customer would consider important in their purchase decision. For some businesses, customers would prioritize a ‘fast turnaround time’ as critical to the quality of their relationship.
- Pareto Priority Assessments: Takes a profitability-based approach to identify variables where change may occur. This is typically calculated by dividing potential savings by potential costs. Projects with a higher Pareto Priority Index (PPI) suggest the project would have a more optimal outcome.
- Process and Value Stream Mapping: A visual construction of steps, events, operations, and value-adds that constitute a process. This can help teams to identify unidentified complexities, problem areas, and redundancies that may have been overlooked in the past without a formalized overview.
Once processes have been observed and defined through frameworks like those above, it’s easier to understand the scope of the project ahead. Taking the time to comprehend factors limiting success will always support improved and more profitable operations.
Once areas for improvement have been discovered, management teams should develop a baseline measurement for how their organization is currently operating. This is one of the most critical aspects of the DMAIC process because without baseline measurements, it would be impossible to determine if the project was successful or not.
What types of performance variables should processes be measured by? Some of the most common measurements include:
- Number of Total Operations: The number of steps in a process to lead from the initial state to final state. A high number of operations would suggest efficiencies could be attained if total steps were reduced.
- Work-in-Process: The value of goods being worked on at any given time. This represents purchased materials and labour currently dedicated to production, where revenue for these activities hasn’t been collected yet. This could help to identify inefficient processes and production bottlenecks.
- Scrap and Rework: The amount of work that has fully completed the process and cannot be sold due to defect. In this case, management must decide whether to scrap the work and start again or modify the defective product/service and remedy the situation.
- Cycle Time: The amount of time elapsed from the beginning of a process to the end. This is typically measured in minutes or seconds, and if unknown, will result in unreliable workflow planning.
- Change-Over Time: The amount of time required to prepare a piece of equipment, device, or process so that it performs a new function or is ready to develop a new batch.
- Lead Time: Comprised of all cycle times, queue times, change-over times, and idle times between processes. This is a wider view of the entire value-added service and seeks to determine efficiencies/inefficiencies between individual stages.
While these measurements provide deeper insight into operational performance, variability in processes is where lean Six Sigma activities are primarily used. Many of the factors above (especially scrap and rework) can be improved if an organization uses statistical analysis to understand how processes work together, and most importantly, meet customer requirements.
Process Capability Index
Customers have an expectation for products and processes; if these aren’t met, customers will begin searching for solutions that better meet their specifications. However, if a company is committed to surpassing the quality requirements of customers, processes could take longer or cost more to deploy. The key is finding an acceptable range where processes can be performed at a healthy margin while also satisfying customer requests.
One way to ensure processes meet customer requirements is to understand the lower specification limit (LSL) and upper specification limit (USL), with an acceptable range between the two. For example, if a steel manufacturer receives an order for 10,000 6-foot-long pipes, and the client expects between 5 and 7 defects, the manufacturer must have a process in place that can guarantee this repeatability.
Measuring the Capabilities of Processes
Measuring performance serves two important functions: to understand if processes meet client standards and to stimulate continuous improvement. Making sure organizations can measure accurately is essential to lean Six Sigma success, therefore, standardized gauges should be used whenever possible.
Measurement Quality: Accuracy vs Precision
To produce quality results, gauges should be both accurate and precise. These two terms do not indicate the same thing, rather:
- Accuracy: The gauge’s ability to measure the variable it’s supposed to without being mislead by other elements of the process.
- Precision: The gauge’s ability to gather the same result, without variation.
Measurement Quality: Repeatability vs. Reproducibility
Likewise, it’s important to look at how human and machine errors can lead to inaccurate measurements. When measuring the quality of processes, there should be a system in place that is both repeatable and reproducible:
- Repeatability: Whether the same appraiser can receive the same value multiple times, using the same measurement device on the same process. This seeks to identify equipment error during measurement.
- Reproducibility: Whether different appraisers can receive the same value multiple times, using the same measurement device on the same process. This seeks to identify user error during measurement.
Once a company is confident in the measurements it’s received from certain gauges and analyses, it’s possible to measure the findings and develop a roadmap for organizational improvement.
The ‘Analyze’ phase of the DMAIC methodology seeks to review process maps and data gained through previous stages and determine where inefficiencies exist. This stems from a root cause analysis and can expand to address opportunities for improvement.
‘Analyze’ is the phase at which businesses truly begin to implement Six Sigma tools and techniques; until this point, many activities performed still fall under the ‘lean’ umbrella. ‘Analyze’ is where organizations implement a systematic approach to process improvement.
Generally, businesses will strive for performance data that meets customer demand and is normally distributed. Processes that do not meet expectations or are distributed unevenly generally signal that there are inefficiencies which must be addressed. There are many possible sources for this variation, including:
- Positional Variation: Processes being completed differently from region to region, or worker to worker.
- Temporal Variation: Processes being completed differently given a different time of day, week, or shift.
- Sequential Variation: A series of processes being completed differently from activity to activity, or lot to lot.
Root Cause Analysis
Once data variations have been identified, it becomes more important to find the source, or root cause, for the inefficiency. Using data to concentrate on success factors will help focus efforts during the implementation and lead towards a successful result. Some of the most popular ways for determining a root cause include:
- 5 Why Analysis: Initially perceived issues aren’t always the root cause of variability. By asking why routine processes are completed, it’s easier to strip away non-issues and focus only on real problems.
- Fishbone Diagram: Uses six branches to critically analyze a process and determine the factors contributing most to variation.
- Failure Mode and Effect Analysis (FMEA): Identifies the ways in which something might fail, and the consequences of a failure happening. This also helps identify potential control measures, reducing the chances of a failure generating a substantial negative effect in the future.
Of course, there are many statistical/graphing tools that can use data to drive insight too. These can help manufacturers and other well-established empirically driven organizations to drive insight into variation and attempt to understand it better. Some graphical analysis tools include:
- Scatter Plots: Allow for the comparison of two variables, such as time and costs to determine if there is a cause-and-effect relationship
- Histograms: Display the frequency and distribution of data in an easy-to-understand bar-style illustration. Other descriptive statistics, such as standard deviation, can be discovered with this data.
- Box Plots: Provide a quick look at data by easily showing outliers and the distribution of values compared to the whole data set.
- Flow Charts: Use a system of annotated boxes to identify trends or major processes. This can help to organize a visual representation of what the process looks like and enables the inclusion of qualitative and quantitative data.
- Analysis of Variance (ANOVA): Test the average values gathered from a product or process. Can be either a one-way ANOVA (single input factor variance) or balanced ANOVA (multiple input factors) to understand how deviations can occur.
Distribution of Quantitative Data
In most processes, random variations will conform to normal distribution. This is based on the Central Limit Theorem which states that regardless of the shape of the distribution curve, data points will typically focus around the middle, or median number.
There are many statistical analyses an organization can perform, such as:
Overall Equipment Effectiveness (OEE)
The performance of a specific piece of equipment – or organization as a whole – can often be determined by focusing on the culmination of three factors:
- Availability: The amount of time a machine is being used for production.
- Performance: The operating speed at which a piece of equipment is used, compared to its design speed.
- Quality: The amount of high-quality product produced.
Organizations can discover their Overall Equipment Effectiveness by using the following calculation:
OEE% = (Availability %) X (Performance %) X (Quality %)
Reduced productivity can usually be tied to one of the three equipment effectiveness criteria outlined above. When this happens, it’s important to find and remedy the problem before it worsens. To do this, project managers should determine what kind of loss is occurring and decide whether the issue is chronic or sporadic.
- Chronic Losses: Are present during normal operating conditions and typically arise because of hidden problems with machinery and production methods. They occur anytime a defect is produced and is not tied directly to a sporadic loss; the key is adopting more innovative practices to reduce variability.
- Sporadic Losses: Typically occur as a result of a sudden breakdown of equipment due to unforeseen circumstances. These are easy to detect since they deviate from standard operating procedures; if sporadic losses occur, management should seek to restore normal operations.
Once inefficiencies have been fully analyzed and understood, it becomes possible to reverse some of the error’s negative impacts on performance. By acting quickly and improving processes, organizations can become more competitive.
It could prove very difficult for some organizations to perfect their processes. Although this can be challenging, it’s the pursuit of quality that is important to remember. Although perfection may never be possible, processes can (and should) be reviewed over time to improve their efficiency.
Zero Defect Quality
Zero defects don’t necessarily equate to zero errors; remember that errors are constantly happening in processes. Defects are caused by one or more errors happening in the process and creates a sub-standard outcome.
Zero Defect Quality is a concept that some errors are unpreventable, but all defects can be prevented with the right control processes. This mindset becomes valuable when developing an operational action plan. There are four methods considered when achieving Zero Defect Quality:
- Point of Origin Inspections: Checking for the cause of errors, not the resulting defect.
- Immediate Feedback or Action: Ceasing operations until the source of defects has been identified and fixed.
- Audit Checks: Developing integrated systems to automatically check for errors and report if they exist.
- Poka-Yoke: Mistake-proofing systems to ensure the chances of errors happening are virtually zero.
Redesigning business processes to eliminate defects will drive long-term benefits, but it can be difficult to do. Luckily, technology is making it easier to design mistake-proof systems and achieve Zero Defect Quality.
Poka-yoke serves three basic functions – predict, prevent, and detect errors to improve responsiveness and reduce the overall defect rate. To do this, poka-yoke systems integrate three defect prevention measures:
- Control: Preventative measures are put in place to make process errors unlikely or impossible.
- Warning: Signals an error may occur soon and that employees or operators should act to prevent it.
- Shutdown: The operation is shut down entirely so that the error or defect can be inspected before more occur.
To do this, sensors and gauges are relied on to provide accurate measurements and integrate with predictive management systems. Poka-yoke measurement methods often include:
- Contact: Functions by detecting an object or part of the process as it contacts the sensor.
- Counting: Relies on a fixed number of operations or parts and counts to ensure all requirements have been included in the final product.
- Motion-Sequence: Sensors understand motion to interpret if a process has been completed properly.
Design of Experiments
Six Sigma uses ‘the scientific process’ to understand value-added processes and discover the root cause of errors. Design of Experiments is where business leaders can plan experiments yielding statistically useful results. Through organized methods, it’s possible to understand the relationship between two or more variables. These experiments help companies to achieve a desired result with the least amount of time and cost.
One-Factor vs. Factorial Experiments
Design of Experiments can be applied in various ways, two of the most common are one-factor and factorial experiments.
- One-Factor Experiments: Seek to identify and understand the effects of changing one variable at a time. This can provide limited insight compared to factorial experiments, though they’re easier to run and analyze.
- Factorial Experiments: Seek to identify the relationship between several different variables within a process. Multiple variables are studied simultaneously to drive insight about the entire process, ultimately leading to more actionable results.
The 5S Methodology
Often seen as the ‘holy grail’ of Six Sigma implementation, 5S refers to a Japanese-based methodology of workplace organization. By implementing each of the five stages, organizations can drive greater productivity and reduce errors. The five elements of 5S include:
- Sort: Separate needed items from unneeded items, keeping only what is immediately necessary on the shop floor. Maintaining a minimalist design and reducing clutter will help develop a sense of order.
- Straighten: Organize the workplace so that needed items can be accessed quickly and easily. This eliminates wasting time to look for tools or other pieces of equipment.
- Shine: Sweep, wash, and clean everything within the immediate work area. This improves safety and helps prolong the life of equipment.
- Standardize: Keep workspace clean for a constant state of readiness. At this point of 5S, standardize seeks to synchronize the previous three elements and reproduce the lean system/process.
- Sustain: Ensure everyone acknowledges, understands, and uses best practices while inside the facility. This involves continuous employee engagement in 5S, including employee training, developing pocket manuals, or creating a visual map of the process to help the team understand where future errors may occur.
Methods of Improvement: Innovation vs. Continuous Improvement
5S is generally thought of as a method for continuous improvement; once implemented, teams within the organization perform tasks continuously to ensure all elements are being upheld. Although changes may happen slowly, over time they lead to large-scale organizational improvements.
Innovation is an alternate approach to increasing workplace efficiency. However, it can have drawbacks.
Although innovative improvements, such as purchasing new machinery, could lead to significant improvements, over time the benefits provided will lessen. Ideally, businesses should integrate innovation and continuous improvement to become more efficient. Small, ongoing modifications can be made to ensure productivity remains high, and then innovative upgrades can lead to transformative growth.
Undoubtedly, one of the most common ways to integrate continuous improvement is by facilitating Kaizens. The general implementation of this strategy involves a group of workers (generally with different roles in the organization) collaboratively going through a set of procedures to identify waste or opportunities for error. Through a systematic approach, it becomes easy to continuously reduce waste.
Kaizens can vary in scope and duration. Small Kaizens are often completed in a day or less; most Kaizens take about a week, and large-scale Kaizens could even take two weeks or longer. This all depends on the sophistication of the process being observed.
Embedding efficiency in the organization’s culture can be a tricky task, but with the right attitude, establishing a process to control these standards can become easier. This is a necessary step, otherwise work done to bring the improvements will be wasted. Process controls can be established through four steps:
- Implement Ongoing Measurement: Tracking performance variables in regular intervals to determine if further action is necessary.
- Standardize Solutions: Make changes to normal operating procedures.
- Quantify Improvements: Check results with the intended goal over time as some solutions may work initially and become less effective.
- Close the Project: Ensure a clear process has emerged and report on the project, clearly stating techniques for controlling it in the future.
Once in the control phase of process improvements, its important to have a toolbox of methods for maintaining a state of high efficiency. Some of the most common include:
- Monitoring: Visual and electronic review of the process to determine if inefficiencies still exist.
- Standardizing: Using best practices at each stage of the process to reduce variability and ensure repeated success.
- Documenting: Recording the process and known issues that have occurred within the process so the impact of future errors can be minimized.
Process Control Plans
For many organizations, documenting processes is the easiest and most effective way to communicate how steps should be performed. Developing a Process Control Plan to identify the most critical pieces of control information will help employees across the company to use best practices and enhance efficiency.
The Process Control Plan for activities should be a comprehensive overview of everything someone needs to know about the process prior to completing it. Developing these plans will increase consistency and accelerate training programs for employees learning the process.