How do you apply statistics to fix process problems?
Statistics are essential tools for Lean Six Sigma practitioners who want to improve processes and eliminate waste and variation. Statistics help you measure, analyze, and control the performance and quality of your processes, and identify the root causes of problems and opportunities for improvement. In this article, you will learn how to apply statistics to fix process problems using the DMAIC framework and some common Lean Six Sigma statistical methods.
The first step in fixing a process problem is to define it clearly and precisely. You need to state what the problem is, where it occurs, when it occurs, how often it occurs, and how it affects your customers and stakeholders. You also need to quantify the problem by setting a baseline and a goal for the process metric that you want to improve. For example, if your problem is high defect rate in a manufacturing process, you need to measure the current defect rate and set a target defect rate that meets your customer's expectations and your business objectives.
-
Prashant Joshi
Vice President - Manufacturing, Sricity Product factory, Daikin Airconditioning India Private Limited
I think the simplest and most powerful statistical method in fixing any problems is the Pareto analysis or the 80~20 rule. Understanding the 20% causes which will address the 80% problems is the most important and effective way in problem solving. And adequate or ample statistical data will help in improving the success or effectiveness of the Pareto analysis.
-
Juliano Anjos
Analista da Qualidade |🏆23x Top Voice | Black Belt LSS | Excelência Operacional | Manufatura | Auditor ISO/IATF/BIQS | Fornecedor | Gestão Projetos e Processos
Begin by clearly stating what the problem is, of this should be a specific description of the issue or challenge that needs to be addressed, and use clear and concise language to articulate the problem statement. Describe where and when the problem occurs, of this provides context and helps identify potential root causes, but understanding when and where the problem occurs can also help determine the scope of the improvement project. Determine how often the problem occurs and the impact it has on customers, stakeholders, and the organization as a whole, then this helps prioritize improvement efforts and allocate resources effectively!
-
Ajay Makwana
QC | QA | SQA | ISO Lead Auditor | Safety coordinator | Carbon footprint | Automation | Product Certification - BIS ASTA KITE Mark | LSSGB | QMS | Process improvement | SAP | PPAP | APQP | PFMEA | Lean Manufacturing |
Statistics are used to fix problems by analyzing data, testing hypotheses, improving quality, predicting outcomes, identifying root causes, optimizing processes, assessing risks, and making data-driven decisions.
-
Anil Chaudhary
Finance Transformation | Certified LSSBB
At a define stage not much of Statistics is used but more of six Sigma tools such as Process maps, SIPOC, CTQ, Gantt chart, RACI etc.You can also use some initial statistics like descriptive statistics to understand your process data.
-
Tasnimul Alam
Production Supervisor || Experienced in Manufacturing & Automotive Engineering || Engineering Design || CFD || Aircraft Structures || Project Management
Statistics are critical in addressing process issues. In manufacturing, for example, if failures arise, statistical analysis might identify root reasons such as machine faults or operator errors. Adjustments, experiments, and process controls can all be carried out using statistical approaches. This data-driven methodology aids in defect reduction, quality improvement, and cost reduction, ultimately resolving process issues and increasing efficiency.
The second step in fixing a process problem is to collect and analyze data that relates to the problem. You need to select the appropriate data sources, methods, and tools that will provide accurate and reliable information about the process and its variation. You also need to use descriptive statistics to summarize and display the data in charts, graphs, and tables. For example, you can use histograms, box plots, run charts, and Pareto charts to show the distribution, variation, trends, and frequency of the data.
-
Fahmi Fisol
Senior Process Engineer at Kaneka (M) Sdn. Bhd. | Project Management, Plant Design, Process Optimization
If you are at a loss, the Ishikawa diagram, combined with the Pareto chart, is a good starting point to help you focus and prioritize which area to improve. But, of course, you need to ensure the data is accurate.
-
Kartick Dey,LSSBB®
Deputy General Manager -Operation Planning @ SUN Mobility | Lean Six Sigma
Measure is the second step in DMAIC. It is one of the most important step in Problem solving. Peter Drucker said, "If you can't measure it, you can't improve it". Data collection, measuring the data and representing data in a structured forms in terms of charts, graphs, visuals form the core of Data measurement. Statistical sampling is used to collect data. Various techniques like Pareto charts, Histogram, Scatter plots are used to represent data. Well presented data with engaging stories can lead to deeper understanding of the underlying issues. Measurement System Analysis consisting various Statistical techniques enable measuring the data in the right context leading to effective Analysis of the data.
-
UJJWAL S.
Lead Marketing | AWMPL | RE Sustainability Ltd | BIT SINDRI | IDTR |
First we have to understand the use of statistics management according to the projects . Now generally what happens is we have to collect and analyze data but many statisticians don't even know which formulas have to apply in it and while reaching to the conclusion(right formula) during the research work they almost waste their time. why? because there is a lack of knowledge of statistics . First we have to focus on learning, we don't have to apply just by watching somewhere or just saw that it was used by someone I can also crack it . We have to be more realistic by using FFF Find, Focus and Fix . Complete the research work first then apply the basics of statistics management and you will feel that you are pro in it.Then go for advance
-
Alex GOH
Leader in Operational Excellence | Regional Program Management | Continuous Improvement | Change Management
Statistics is used to help make better decision or have better clarity of situation and thus contribute to "fix " the process problem. Using DMAIC, Statistics can be used in Measure phase to baseline current performance. For e.g. use of Cp/Cpk process capability, you can use this statistical routine to understand if the process is capable. Statistic is often used in Analyze phase as evidence to a hypothesis. For e.g. 2-sample-t test used to show significant difference between two samples, and validate a cause due to differences . Similarly, statistics can be used in Improve phase to validate success of improvement . Finally, statistics can be used in Control phase like control chart to show process stability after improvement
-
Ravindra Satav
Continuous Improvement |Top Lean Six Sigma & Lean Manufacturing Voice |Lean Practitioner | Operational Excellence | LSSBB | Process Improvement | Operations | Quality | TPM | TQM
Whenever we start the six sigma project, we need to ensure that we collect the data of Y alongside data of X. This way we can drive statistical analysis to validate the potential cause using hypothesis testing.
The third step in fixing a process problem is to identify and verify the root causes of the problem. You need to use inferential statistics to test hypotheses and draw conclusions about the relationship between the process variables and the problem metric. You also need to use tools such as fishbone diagrams, 5 whys, scatter plots, and correlation and regression analysis to explore the potential causes and their impact on the problem. For example, you can use a fishbone diagram to brainstorm possible causes of high defect rate, and then use a scatter plot and regression analysis to quantify the effect of each cause on the defect rate.
-
Kartick Dey,LSSBB®
Deputy General Manager -Operation Planning @ SUN Mobility | Lean Six Sigma
Analyze is the third step in DMAIC. In Analyze phase the causes of the problems are uncovered. In analysis of the problems various Statistical methods are used namely Correlation and Regression analysis, Hypothesis testing etc. FMEA is another methodology often used in the analysis of the problems. This is a very powerful prevention-based risk management tool. The key concepts we should be aware of this tool are Failure Mode, Effect analysis, Severity, Occurance, Detection and Risk Priority Number. RPN= Severity× Occurance× Detection In Analyze phase, additional analysis methods comprise of Gap analysis, Pateto chart. 5 Why's and Cause-and-Effect diagrams are the most common used tools to find out Root causes of the problems.
-
Felipe “Pina”
Anl Engenharia de Processos PL | Kaizen | OFM | Franquia | Lean Expert | LSS Green Belt | Black Belt |Ágil | Transformação | OKR | Projetos e Processos | Itaú Unibanco
As análises estatísticas fornecem insumos e validam o que de fato está precisando tratar/sanar. O uso delas, nos mostra o quão podemos ser tempestivos e identificar oportunidades do negócio.
-
Juan Valdez, CLSSBB
Mechanical Engineering | Program Management |Lean Six Sigma Black Belt | NPI | Continuous Improvement | R&D | Increase Sales & EBIT% | Reduce Costs | Innovator |
For instance, in a manufacturing setting with a high defect rate, use statistical tests and visual tools like scatter plots. Utilize techniques such as fishbone diagrams and 5 Whys to identify root causes, applying regression analysis for quantification. Validate findings through experiments and implement changes based on statistical evidence for sustained improvement.
-
João Henrique da Silva
Manufacturing Manager | TPM | Continuous Improvement Manager | Plant Manager | Project Manager | Manufacturing Excellence Manager
A good practice is to assess the current capability of the process to generate defects and repeat this analysis after the actions are executed. Here's how to do it: - Map the process and its variables. - Highlight the critical steps, inputs, and outputs. - Clearly define the defect modes. - Understand process and customer requirements. - Measure the process over time. - Run a capability analysis, assess Cp and Cpk, and evaluate the findings.
-
Muhamed Ekram
Sr. QA/QC Engineer @ NPPA |MBA(MQM), BPM, CI, Lead Auditor, ASNT NDT LV II, TOT, Operational Excellence, Lean Six Sigma, Process Improvement| |Nuclear Power Plants-Oil & Gas|
During the analysis phase we need to understand the current process (As is) and analyze the previously collected data, so it can be accomplished by different statistical methods: -Run Chart to track the process over time to figure out trends and special causes of variation; -Normality test to measure the p-value >=0.05 to make a decision wheather data is Normally distributed or not and to make sure that MSA is performed properly; -Distribution curve to have an overview of the distribution of data, mean, median, mode, SD, etc; Finally, determine the process capability value of as is a process (CP baseline) and prioritize the improvement actions according to grading criteria (significance of variation/defect)
The fourth step in fixing a process problem is to generate and implement solutions that will address the root causes and improve the process performance and quality. You need to use tools such as brainstorming, affinity diagrams, prioritization matrices, and design of experiments to generate, evaluate, and select the best solutions. You also need to use tools such as pilot tests, control charts, and hypothesis tests to validate and verify the effectiveness of the solutions. For example, you can use a design of experiments to test different combinations of factors and levels that affect the defect rate, and then use a control chart and a hypothesis test to compare the before and after results.
-
Kartick Dey,LSSBB®
Deputy General Manager -Operation Planning @ SUN Mobility | Lean Six Sigma
Improve is the fourth step in DMAIC. In the previous step Analyze, ideally root causes are identified. In this step, Actions for the problems in order to eliminate the root causes are generated. There are generally two types of actions- Corrective actions and Preventive actions. CAPA are generated through brainstorming, running pilots on proposed solutions, validating the solutions, Design of Experiment etc. A few Lean Improvement Tools are Kanban, Pull Systems, 5S, Standard Work, Poka-yoke, Cycle time reduction, Single piece flow, Single Minute Exchage of Die, Production Levelling, Kaizen,Total Productive Maintenance. Quality Circle is a very effective group activity which is often employed in finding and improving the solutions.
-
Mantosh Narayan Gupta
LinkedIn Top Lean Manufacturing Voice| Lean Guide | Program management | Quality | Operational Excellence | Training & Development| Mahindra Institute of Quality(MIQ) |
In Six Sigma, Design of Experiments (DOE) is like a detective's toolkit, helping you crack improvement cases by methodically studying how factors influence your process. Imagine baking the perfect pizza: temperature, dough amount, and cheese type are factors. DOE lets you test various combinations to find the ideal set-up for crispy crust, gooey cheese, and savory perfection. By analyzing the results, you pinpoint the factors creating the best pizza, replicating that success consistently. Similarly, DOE in any industry tackles complex processes, uncovering the optimal settings for improved quality, reduced costs, and happier customers. It's a powerful tool for Six Sigma practitioners, turning intuition into data-driven solutions
-
Prashant Kolambekar
Regional Manager at INTERCERT INC
Certainly, improving the solution involves taking action to make the process better. Here's how you can do it in simple language: 1.Generating Ideas 2.Organizing Thoughts 3. Choosing the Best Solutions 4. Testing Solutions 5. Making It Permanent: If the solution works well in the pilot tests, you can make it a permanent part of the process. It's like deciding to always use the improved recipe because it tastes better.
-
Joris Lootens
Lean Six Sigma Black Belt North-West Europe at PepsiCo
During the Improve phase of your project, you are actively looking to improve or solve the root causes identified in the Analyze phase. The team can think about potential solutions and what effect those will have on the process. A benefit-effort matrix can be used to prioritize which solutions to implement. An FMEA can support the team to asses the risk of the solution Piloting your solutions is the way to go to test the impact on the process. During the pilot, you collect data and asses the impact. Using the relevant statistical tools, you can verify the impact on your process control and process capability, whereas hypothesis test will tell you if the solution has a statistically significant impact on your process vs your baseline data.
-
Percival J.
CLSSYB | LSSBB | Manager, Global Product Management - Product Operations at Monoprice
As a Lean Six Sigma practitioner, one can apply statistical tools such as Hypothesis Testing, Regression Analysis, and Control Charts to identify areas of variation or inefficiency. Applying the correct statistical methods and analysis helps in understanding the root causes of problems, optimizing processes, and ensuring sustained improvements. This method is within the framework of Lean Six Sigma methodologies to make data-driven decisions to enhance efficiency and reduce defects.
The fifth step in fixing a process problem is to monitor and sustain the improved process and prevent recurrence of the problem. You need to use tools such as control charts, process capability analysis, standard operating procedures, and audits to ensure that the process is stable, capable, and compliant with the specifications and standards. You also need to use tools such as dashboards, feedback, and continuous improvement to track and communicate the process performance and quality, and identify and address any new or emerging problems. For example, you can use a control chart and a process capability analysis to measure and report the defect rate over time, and use feedback and continuous improvement to identify and resolve any issues or opportunities.
-
Joris Lootens
Lean Six Sigma Black Belt North-West Europe at PepsiCo
The Control phase is an often neglected or rushed part of LSS-projects. However, failing to properly execute this phase, won't give you a sustainable outcome of the project. During this phase the team needs to ensure all changes in the process are properly documented and trained: are all procedures updated? are all relevant employees informed and trained on the new way of working? are the targets updated? There needs to be a formal close-off of the project and handover from the project team to the process owner. A control plan will help the process owner and the involved employees to understand when the process is deviating from the new standard and how to react if this happens. Last but not least: reward & recognize the team!
-
Mantosh Narayan Gupta
LinkedIn Top Lean Manufacturing Voice| Lean Guide | Program management | Quality | Operational Excellence | Training & Development| Mahindra Institute of Quality(MIQ) |
Here are some of the key tools used in the Control phase of Six Sigma: 1. Control Charts Reveal any unusual patterns or trends Like X-bar and R charts, Individuals and Moving Range (I-MR) charts, and Cumulative Sum (CUSUM) charts. 2. Capability analysis Involves process capability to confirm results 3. Process Documentation Clear and detailed documentation of the project 4. Training Equips employees with the knowledge and skills to maintain the improved process. 5. Visual Management Uses visual cues like charts, graphs, and displays to communicate process performance 6. Control Plans: Define the specific actions to be taken to maintain process control. 7. Audits and Reviews: Regularly evaluate the effectiveness of control measures
-
Prashant Kolambekar
Regional Manager at INTERCERT INC
Let's simplify the concept of controlling the process: 1. Keeping an Eye on Things 2. Using Special Tools 3. Following the Rules 4. Regular Check-Ups 5. Staying Updated and Improving
-
Ronanki Ravi
Supplier Quality Management , Quality Assurance Professional.
specific examples of how statistics can be used to fix process problems: Use statistical process control (SPC) to monitor and control key process variables. SPC is a collection of statistical tools and techniques that can be used to monitor and control key process variables to ensure that the process is operating within specified limits. Use design of experiments (DOE) to optimize the process. DOE is a statistical method that can be used to test different combinations of process variables to identify the combination that produces the desired results.
-
Vinay Dahiya
Senior Consultant - TQM & Lean Six Sigma | Corporate Quality | Business Excellence | Project Management | Continuous Improvement | Operations & Supply Chain Management | Process & Supplier Quality | Ex Honda Cars
1. Through Control charts, we get to the process stability. Herein assignable causes are identified and removed. 2. Through histogram, process shifts are observed. While Cpk determines process capability. Herein actions are taken to make the process capable. Please note: First a process needs to be made stable and then only it can be made capable.
-
Ravindra Satav
Continuous Improvement |Top Lean Six Sigma & Lean Manufacturing Voice |Lean Practitioner | Operational Excellence | LSSBB | Process Improvement | Operations | Quality | TPM | TQM
The main objective of using statistics is problem solving to validate problems and stop the reoccurring. We need to satisfy the transfer function Y=f(x) to resolve the problem as we are going to validate all problems caused to the potential cause with data for the subjected problem (y). 7 qc tools,one way ANOVA, 2 sample T test , regration analysis ,box plot ,SPC, DOE etc these are popular tools used to identify , validate and generate solutions for six Sigma problems.
-
Mantosh Narayan Gupta
LinkedIn Top Lean Manufacturing Voice| Lean Guide | Program management | Quality | Operational Excellence | Training & Development| Mahindra Institute of Quality(MIQ) |
Applying statistics to fix process problems involves utilizing various statistical tools and techniques to analyze data, identify patterns, and derive insights to address inefficiencies or defects in a process. Initially, data related to the process is collected and organized. Statistical methods such as hypothesis testing, control charts, regression analysis, and root cause analysis are then applied to understand the variation within the process, pinpoint potential causes of problems, and prioritize areas for improvement. By analyzing statistical data, identifying outliers, trends, or anomalies, and understanding the root causes of issues, organizations can make informed decisions to implement corrective actions
-
Suryakant Bhavikatty
Next-Gen eQMS Expert | LinkedIn Top Quality Management Voice | 4 million Impressions | Driving Automated End-to-end QMS, EHS, and Integrated Solutions
There are many statistical charts and studies available. As an organisation we need to identify which type of charts/studies are applicable for different types of processes problems. Following studies can be used for variation and waste reduction: 1) SPC and MSA: both collectively contribute to the overall variation. Need to identify specific studies for resolving the manufacturing/measurements variation problems. 2) DOE can be used to understand the causes with different relationships. 3) Correlation or Regression analysis can be used to see the behaviour of any multiple variables. 4) Multilayer of data input will help the users to go to the micro level , track and Analyze the problems faster to take the actions
Rate this article
More relevant reading
-
Lean Six SigmaHow do you coordinate with other Lean Six Sigma experts when using statistical methods?
-
Six SigmaHow do you validate the 80/20 rule in Pareto analysis for Six Sigma?
-
Lean Six SigmaHow do you rank root causes?
-
Six SigmaHow do you update and maintain Pareto charts in Six Sigma as the process changes or improves?