Monday, 24 June 2013

Measure eLearning Course Quality with Six Sigma

Let's focus on how the Measure Phase of Six Sigma DMAIC gets implemented during eLearning Course Development Process improvement initiative. Measure phase is about validating measurement system, analyzing data, gathering root causes, measuring high impact elements and mapping process. After the improvement efforts are scoped in Define phase, the team starts working on measure phase. Instead of looking for new measurements, the team needs to take measurements that already exist in the process and unravel them. The team needs to identify high impact eLearning course development process elements, develop measureable baseline metrics and create detailed process mapping. In Measure phase, the team ensures measurement system’s consistency and verifies if the tool used to collect output variable is flawed or if all operators interpret the tool reading in the same way. The team decides on the purpose of data collection, the type of analysis and the period of data collection. Various tools are used for collecting and communicating baseline metrics and priorities.

Approach to Measure

Process Mapping
Process mapping helps a team understand how the activities take place. It gives a high level overview of a process defining how, when or where a process should be measured. It helps the team investigate where problems might occur and where alterations be made that gives optimum outcomes. The process map is a graphical representation of all process steps (value added and non value added), process inputs (X’s), and process outputs (Y’s). A process map consists of flowchart having symbols such as arrow, circle, diamond, oval or rectangle and details the inputs, activities, decision points, rework loops and outputs of the process. 
process mapping six sigma

Data Gathering
Data collection is an important aspect of Six Sigma projects. Inaccurate data impacts the results of a study and eventually derails the project. In order to collect data, the team needs to identify the purpose of data collection, decide the period of data collection and find out if the necessary data is already on hand. The numerical data that will be measured during Measure phase falls into two categories:
-> Continuous data can take any value within a range. For example time, weight, height etc.
-> Discrete data can only take certain values. For example: number of defects.
Funneling is choosing the appropriate measures to get precise information of the problem. The Six Sigma methodology provides two techniques to identify the “critical few” measures.
-> Prioritization Matrix/ XY Matrix
Prioritization matrix - Prioritization Matrix, also known as criteria matrix identifies the critical few variables that need to be measured and analyzed. It helps to focus data collection effort, formulate theories about causes and effects and improve decision making. Prioritization matrix, a systematic approach is implemented when too many variables have an impact on the output of the process and collecting data on all possible variables would cost too much time and money. In case of team members having different theories about what happens in the process, the prioritization matrix promptly unravels basic disagreements. It allows the team to narrow down the focus and leads to implementation success.

To construct a Prioritization Matrix for input/process variables
-> List the output variables
-> Rank order and weight the output variables
-> List the input variables
-> Evaluate the strength of the relationship between output and input variables (correlation factor)
-> Cross-multiply weight and correlation factor and add
-> Assess variable with every other variable
-> Highlight the critical few variables
prioritization matrix six sigma
FMEA (Failure Mode Effect Analysis) - FMEA allows to discover potential issues in a process, possible impact of each issue, and approach to fix each issue. There are two types of FMEA, Design FMEA and Process FMEA. Design FMEA is used during process or product design and development. The primary objective is to uncover problems that will result in potential failures within the new product or process. With the help of FMEA, the team can collect useful information, capture engineering knowledge, minimize late changes and associated costs and reduce same kind of failures in future. Process FMEAs are used to uncover problems related to manpower, systems, methods, measurements and the environment.

FMEA Advantages
-> Bottom-up look on each criteria
-> Emphasis on problem prevention
-> Improved quality
-> Improved competitiveness
-> Cut down cost
-> Avoid defects, failures, and downtime
-> Reduce the possibility of same kind of failure in future
-> Increased customer satisfaction

FMEA identifies, quantifies and evaluates failure and its goal is to improve quality, competitiveness and customer satisfaction. FMEA drives systematic thinking about a product or process and focuses on the three basic issues:
-> What might cause system/process stoppage? – Failure
-> How bad the effect is, when something goes wrong? – Risk
-> What can be done to avoid things from going wrong? – Corrective Action

FMEA attempts to identify and prioritize potential process or system failures. The failures are rated on three components:
-> Severity – Impact of a failure
-> Occurrence – Frequency of causes of failure
-> Detectibility – How easy it is to detect the failure

Steps of Conducting an FMEA
1. Define the scope of FMEA
2. List process steps of existing process
3. List possible failure modes
4. List potential effects of failures
5. Assign severity of each effect
6. List potential failure causes
7. Assign occurrence rating for each cause
8. List current process controls for detection of failure modes
9. Assign detection levels to each failure mode
10. Calculate the risk priority number (RPN) for each cause
11. Rank or prioritize causes
12. Take action on high risk failure modes
13. Recalculate RPN numbers

Risk Priority Number (RPN)
The team develops a ranked list of potential failure modes by calculating Risk Priority Number (RPN). Then the failures in an FMEA project are prioritized by Risk Priority Number, or RPN values. The RPN values are calculated by multiplying together the Severity, Occurrence, and Detection (SOD) values associated with each cause-and-effect item identified for each failure mode. The higher the RPN value, the higher the priority to work on that specific function. Team now takes action on high risk failure modes. Each cause of high risk needs to be acted upon and closed properly. The reduction in RPN scores results in dramatic improvement in process sigma level.

FMEA six sigma 

Data Collection Plan 
A data collection plan is a detailed document. It explains the technique and the sequence to be executed in gathering the data for the Six Sigma project. This document is very critical and it makes sure that each team member of the Six Sigma project is in agreement with the data plan. Below mentioned action items ensure that the data collection process is stable and the measurement system is reliable. 

1. Identify data collection plan objectives
2. Decide the time frame and data elements to be included in the measurement
3. Create clear, concise and detailed operational definitions and methodology for each variable. Specific and concrete operational definition reduces ambiguity, offers the understanding of variable characteristics and explains the method for measuring the characteristics.
4. Ensure the reliability of data collection plan and reliability of measurement system by making sure that the plan is being implemented precisely and consistently
5. Follow through the data collection process and results to ensure the consistency and accuracy of execution 
data collection plan six sigma
The difference between two processes/products/services for the same characteristic is called variation. In Statistical Process Control, the variation in a process is classified in two ways, Common Cause Variation and Special Cause Variation.

Common Cause Variation – The Common Cause Variation is created by many factors that are also known as traditional 6Ms (Man power, Mother Nature, Materials, Method, Measurement and Machine). 6Ms affect any process and will confirm to a normal distribution. Common Causes are always inherent and always present to some extent in the process and can only be reduced by fundamental changes to the system.
Special Cause Variation – The Special Cause Variation arises from specific factors/causes or non random events that are unusual and not previously observed. It has a considerable effect on the process. Special Cause Variations lead to unexpected change in process and makes the process unpredictable and unstable. Special Causes can be tracked down and removed using Special Cause Problem Solving methodology.

The objective of the eLearning Course development team is to minimize variation in the process and stabilize the process. The team needs to identify and isolate the causes of variation that are influencing the course quality and find solutions to them in order to eliminate these causes on a long term basis.

“The objective in driving Six Sigma performance is to reduce or narrow variation to such a degree that six sigma – or standard deviations – can be squeezed within the limits defined by the customer’s specification.” (Pande)

Data Summarization
Measuring Central Tendency
Mean, Median and Mode are the three ways to measure the central tendency of data. These 3M’s are measures to find the middle point of observations. This middle point of observation helps us find the representative value of entire distribution and enables us compare data.

Mean represents the average of data. The concept of mean is simple and most intuitively understood. However, Mean gets affected by extreme value.
Median represents middle most of central value in a set of ordered data. Extreme values do not influence the Median value as strongly as mean.
Mode represents the value that is repeated most often in a data set. Extreme values do not influence the Mode value as strongly as mean. However, statistical calculations do not support Mode and Median calculations as much as the Mean.

Measuring Dispersion
Measures of dispersions are important for determining the extent of the spread of the data from the mean value. Various methods are used to measure the dispersion of a dataset. These methods are
- Range ( R )
- Standard Deviation ( S )
- Variance (S­­­­2)
- Co-efficient of Variation (CV)

To calculate these measures, the team can use one of the statistical software packages, Minitab.

Histogram graphically represents data frequency distribution in bar form and helps summarize data from process that has been collected over a period of time.
histogram six sigma

Box Plot
Box Plot is the great tool for graphical analysis especially for non normal data. Box plot summarizes the set of data on an interval scale. The spacing between the different parts of the box indicates the degree of dispersion (spread) and skewness in the data and identifies outliers (the most extreme values) in the dataset.

box plot six sigma



In this section we reviewed the concepts of Process Mapping, Data Gathering, Funneling, Data Collection Plan, Variation and Data Summarization. My next blog will elaborate the concepts of Measurement System Analysis, Pattern Analysis, Pareto Chart, and Process Capability.

Share you views and experiences on how a robust Measurement System leads to a successful Six Sigma initiative.

No comments :

Post a Comment