What is Six Sigma?

Introduction to Six Sigma

Introduction to Six Sigma

Six Sigma is usually related to the magic number of 3.4 defects per million opportunities. People often view Six Sigma as yet another rigorous statistical quality control mechanism.

Pioneered at Motorola in the mid-1980s, Six Sigma was initially targeted to quantify the defects occurred during manufacturing processes, and to reduce those defects to a very small level. Motorola claimed to have saved several million dollars. Another very popular success was at GE. Six Sigma contributed over US $ 300 million to GE's 1997 operating income.

Today Six Sigma is delivering business excellence, higher customer satisfaction, and superior profits by dramatically improving every process in an enterprise from financial to operational to production. Six Sigma has become a darling of a wide spectrum of industries, from health care to insurance to telecommunications to software.

What is Six Sigma?

It is important to recall that every customer always values consistent and predicable services and/or products with near zero defects. Therefore they experience the variation and not the mean. Mean is their expectation or our target.

If we can measure process variations that cause defects i.e. unacceptable deviation from the mean or target, we can work towards systematically managing the variation to eliminate defects.

Six Sigma is a methodology focused on creating breakthrough improvements by managing variation and reducing defects in processes across the enterprise.

Sigma is a Greek symbol represented by "σ".

Why do we call Six Sigma as Six Sigma and not Four or Five Sigma or Eight Alpha (another Greek symbol)? Sigma is a statistical term that measures process deviation from process mean or target. Mean is also referred as average in common language. The figure of six was arrived statistically by looking at the current average maturity of most business enterprises. We would like to revise this figure to 8 or may be 9 provided the world becomes a more orderly and predictable (even with increasing entropy or chaos) place to live!

We have a detailed discussion on keywords "breakthrough improvement" and "variation" apart from the "methodology" in later sections.

Example

Let us take an example to bring a breakthrough improvement in our current understanding of the concept of Six Sigma. This requires us to have basic knowledge of statistics. We have a detailed discussion on required statistical concepts later.

Consider a pizza delivery shop that guarantees the order delivery with 30 minutes from the time of accepting an order. In the event of a delivery time miss, the customer is refunded 100% money. The management took a target (read mean) of delivering every pizza order within 15 minutes and aligned its processes to meet this goal.

If we collect data of delivery times over a large number of the delivery made by the pizza shop and make a frequency distribution graph, we discover that it resembles a "bell shaped curve". A frequency distribution graph is constructed from the frequency table; a frequency table lists different time intervals (called classes) like 0 to 2 minute, 2 to 4 minutes, to 14 to 16 minutes to 28 to 30 minutes and the count of the deliveries made in each interval. The mean is found to be 16 minutes and standard deviation (measure of deviation or dispersion in data i.e. σ) is found as 2.5 minutes. A graph drawn from the data of over 5000 deliveries 

Introduction to Six Sigma

Introduction to Six Sigma

Six Sigma is usually related to the magic number of 3.4 defects per million opportunities. People often view Six Sigma as yet another rigorous statistical quality control mechanism.

Pioneered at Motorola in the mid-1980s, Six Sigma was initially targeted to quantify the defects occurred during manufacturing processes, and to reduce those defects to a very small level. Motorola claimed to have saved several million dollars. Another very popular success was at GE. Six Sigma contributed over US $ 300 million to GE's 1997 operating income.

Today Six Sigma is delivering business excellence, higher customer satisfaction, and superior profits by dramatically improving every process in an enterprise from financial to operational to production. Six Sigma has become a darling of a wide spectrum of industries, from health care to insurance to telecommunications to software.

What is Six Sigma?

It is important to recall that every customer always values consistent and predicable services and/or products with near zero defects. Therefore they experience the variation and not the mean. Mean is their expectation or our target.

If we can measure process variations that cause defects i.e. unacceptable deviation from the mean or target, we can work towards systematically managing the variation to eliminate defects.

Six Sigma is a methodology focused on creating breakthrough improvements by managing variation and reducing defects in processes across the enterprise.

Sigma is a Greek symbol represented by "σ".

Why do we call Six Sigma as Six Sigma and not Four or Five Sigma or Eight Alpha (another Greek symbol)? Sigma is a statistical term that measures process deviation from process mean or target. Mean is also referred as average in common language. The figure of six was arrived statistically by looking at the current average maturity of most business enterprises. We would like to revise this figure to 8 or may be 9 provided the world becomes a more orderly and predictable (even with increasing entropy or chaos) place to live!

We have a detailed discussion on keywords "breakthrough improvement" and "variation" apart from the "methodology" in later sections.

Example

Let us take an example to bring a breakthrough improvement in our current understanding of the concept of Six Sigma. This requires us to have basic knowledge of statistics. We have a detailed discussion on required statistical concepts later.

Consider a pizza delivery shop that guarantees the order delivery with 30 minutes from the time of accepting an order. In the event of a delivery time miss, the customer is refunded 100% money. The management took a target (read mean) of delivering every pizza order within 15 minutes and aligned its processes to meet this goal.

If we collect data of delivery times over a large number of the delivery made by the pizza shop and make a frequency distribution graph, we discover that it resembles a "bell shaped curve". A frequency distribution graph is constructed from the frequency table; a frequency table lists different time intervals (called classes) like 0 to 2 minute, 2 to 4 minutes, to 14 to 16 minutes to 28 to 30 minutes and the count of the deliveries made in each interval. The mean is found to be 16 minutes and standard deviation (measure of deviation or dispersion in data i.e. σ) is found as 2.5 minutes. A graph drawn from the data of over 5000 deliveries made is given below. Note, this not a real graph and is used only for illustration purposes.

This bell shape curved is called "normal distribution" in statistical terms. In real life, a lot of frequency distributions follow normal distribution, as is the case in the pizza delivery times. Natural variations cause such a distribution or deviation. One of the characteristics of this distribution is that 68% of area (i.e. the data points) falls within the area of -1σ and +1σ on either side of the mean. Similarly, 2σ on either side will cover approximately 95.5% area. 3σ on either side from mean covers almost 99.7% area. A more peaked curve (e.g. more and more deliveries were made on target) indicates lower variation or more mature and capable process. Whereas a flatter bell curve indicates higher variation or less mature or capable process.

After this statistical detour let us come back to our pizza example. If the pizza shop delivers 68% of pizza orders in time, we call it a "One Sigma shop". Similarly, if the pizza shop makes 95.5% deliveries on time, we call it a "Two Sigma shop". In our example, data suggests that it is almost a "Three Sigma shop".

We should now be able to appreciate why management took a delivery time target of 15 minutes and not 30 minutes. Imagine what would have happened with a 30 minutes delivery time target!

The "delivery time" is a critical-to-quality parameter from the customer perspective and has a significant impact on profits. In addition, it is an entry barrier for the competition. Such a parameter is called a CTQ and its definition in context of our pizza shop is given below:

CTQ Name: Timely Pizza delivery
CTQ Measure: Time in Minutes
CTQ Specification: Delivery with 30 minutes from the order acceptance time

Now we can easily define a defect:

Defect: Delivery that takes longer than 30 minutes
Unit: Order
Opportunity: 1 per order i.e. only "1" defect can occur in "1" order

Technical Note: This discussion on the example did not include 1.5σ process shift during the above analysis. The concept is discussed later. The Six Sigma conversion graph including a 1.5σ shift in process is given below:

This graph is on a logarithmic scale. Notice the increasing rate of improvement. For example, 1 sigma to 3 sigma is only 10 times improvement; 3 sigma to 4 sigma is a big 10 times improvement; whereas 5 sigma to 6 sigma is a whooping 1825 times change. That is why we are talking about breakthrough improvements in a journey to Six Sigma.

How does Six Sigma work?

The driving force behind any Six Sigma project comes from its primary focus - "bringing breakthrough improvements in a systematic manner by managing variation and reducing defects". This requires us to ask tougher questions, raise the bar significantly, and force people to think out of the box and to be innovative. The objective is to stretch, stretch mentally and not physically. To make this journey successful there is a methodology(s) to support Six Sigma implementations.

There are 2 potential scenarios - (a) there is already an existing process(s) that is working "reasonably" well and (b) there is no process at all. A bad process is as good as no process.

Scenario (a) focuses on significant process improvements and requires use of DMAIC:

1. Define process goals in terms of key critical parameters (i.e. critical to quality or critical to production) on the basis of customer requirements or Voice Of Customer (VOC)
2. Measure the current process performance in context of goals
3. Analyze the current scenario in terms of causes of variations and defects
4. Improve the process by systematically reducing variation and eliminating defects
5. Control future performance of the process

Scenario (b) focuses on process design using Design For Six Sigma (DFSS) approach. DFSS typically requires IDOV:

1. Identify process goals in terms of critical parameters, industry & competitor benchmarks, VOC
2. Design involves enumeration of potential solutions and selection of the best
3. Optimize performance by using advanced statistical modeling and simulation techniques and design refinements
4. Validate that design works in accordance to the process goals

Note, sometimes a DMAIC project may turn into a DFSS project because the process in question requires complete re-design to bring about the desired degree of improvement. Such a discovery usually occurs during improvement phase of DMAIC.

In addition to the methodology, what counts in this journey is being smart and innovative

made is given below. Note, this not a real graph and is used only for illustration purposes.

This bell shape curved is called "normal distribution" in statistical terms. In real life, a lot of frequency distributions follow normal distribution, as is the case in the pizza delivery times. Natural variations cause such a distribution or deviation. One of the characteristics of this distribution is that 68% of area (i.e. the data points) falls within the area of -1σ and +1σ on either side of the mean. Similarly, 2σ on either side will cover approximately 95.5% area. 3σ on either side from mean covers almost 99.7% area. A more peaked curve (e.g. more and more deliveries were made on target) indicates lower variation or more mature and capable process. Whereas a flatter bell curve indicates higher variation or less mature or capable process.

After this statistical detour let us come back to our pizza example. If the pizza shop delivers 68% of pizza orders in time, we call it a "One Sigma shop". Similarly, if the pizza shop makes 95.5% deliveries on time, we call it a "Two Sigma shop". In our example, data suggests that it is almost a "Three Sigma shop".

We should now be able to appreciate why management took a delivery time target of 15 minutes and not 30 minutes. Imagine what would have happened with a 30 minutes delivery time target!

The "delivery time" is a critical-to-quality parameter from the customer perspective and has a significant impact on profits. In addition, it is an entry barrier for the competition. Such a parameter is called a CTQ and its definition in context of our pizza shop is given below:

CTQ Name: Timely Pizza delivery
CTQ Measure: Time in Minutes
CTQ Specification: Delivery with 30 minutes from the order acceptance time

Now we can easily define a defect:

Defect: Delivery that takes longer than 30 minutes
Unit: Order
Opportunity: 1 per order i.e. only "1" defect can occur in "1" order

Technical Note: This discussion on the example did not include 1.5σ process shift during the above analysis. The concept is discussed later. The Six Sigma conversion graph including a 1.5σ shift in process is given below: 

Introduction to Six Sigma

Introduction to Six Sigma

Six Sigma is usually related to the magic number of 3.4 defects per million opportunities. People often view Six Sigma as yet another rigorous statistical quality control mechanism.

Pioneered at Motorola in the mid-1980s, Six Sigma was initially targeted to quantify the defects occurred during manufacturing processes, and to reduce those defects to a very small level. Motorola claimed to have saved several million dollars. Another very popular success was at GE. Six Sigma contributed over US $ 300 million to GE's 1997 operating income.

Today Six Sigma is delivering business excellence, higher customer satisfaction, and superior profits by dramatically improving every process in an enterprise from financial to operational to production. Six Sigma has become a darling of a wide spectrum of industries, from health care to insurance to telecommunications to software.

What is Six Sigma?

It is important to recall that every customer always values consistent and predicable services and/or products with near zero defects. Therefore they experience the variation and not the mean. Mean is their expectation or our target.

If we can measure process variations that cause defects i.e. unacceptable deviation from the mean or target, we can work towards systematically managing the variation to eliminate defects.

Six Sigma is a methodology focused on creating breakthrough improvements by managing variation and reducing defects in processes across the enterprise.

Sigma is a Greek symbol represented by "σ".

Why do we call Six Sigma as Six Sigma and not Four or Five Sigma or Eight Alpha (another Greek symbol)? Sigma is a statistical term that measures process deviation from process mean or target. Mean is also referred as average in common language. The figure of six was arrived statistically by looking at the current average maturity of most business enterprises. We would like to revise this figure to 8 or may be 9 provided the world becomes a more orderly and predictable (even with increasing entropy or chaos) place to live!

We have a detailed discussion on keywords "breakthrough improvement" and "variation" apart from the "methodology" in later sections.

Example

Let us take an example to bring a breakthrough improvement in our current understanding of the concept of Six Sigma. This requires us to have basic knowledge of statistics. We have a detailed discussion on required statistical concepts later.

Consider a pizza delivery shop that guarantees the order delivery with 30 minutes from the time of accepting an order. In the event of a delivery time miss, the customer is refunded 100% money. The management took a target (read mean) of delivering every pizza order within 15 minutes and aligned its processes to meet this goal.

If we collect data of delivery times over a large number of the delivery made by the pizza shop and make a frequency distribution graph, we discover that it resembles a "bell shaped curve". A frequency distribution graph is constructed from the frequency table; a frequency table lists different time intervals (called classes) like 0 to 2 minute, 2 to 4 minutes, to 14 to 16 minutes to 28 to 30 minutes and the count of the deliveries made in each interval. The mean is found to be 16 minutes and standard deviation (measure of deviation or dispersion in data i.e. σ) is found as 2.5 minutes. A graph drawn from the data of over 5000 deliveries made is given below. Note, this not a real graph and is used only for illustration purposes.

This bell shape curved is called "normal distribution" in statistical terms. In real life, a lot of frequency distributions follow normal distribution, as is the case in the pizza delivery times. Natural variations cause such a distribution or deviation. One of the characteristics of this distribution is that 68% of area (i.e. the data points) falls within the area of -1σ and +1σ on either side of the mean. Similarly, 2σ on either side will cover approximately 95.5% area. 3σ on either side from mean covers almost 99.7% area. A more peaked curve (e.g. more and more deliveries were made on target) indicates lower variation or more mature and capable process. Whereas a flatter bell curve indicates higher variation or less mature or capable process.

After this statistical detour let us come back to our pizza example. If the pizza shop delivers 68% of pizza orders in time, we call it a "One Sigma shop". Similarly, if the pizza shop makes 95.5% deliveries on time, we call it a "Two Sigma shop". In our example, data suggests that it is almost a "Three Sigma shop".

We should now be able to appreciate why management took a delivery time target of 15 minutes and not 30 minutes. Imagine what would have happened with a 30 minutes delivery time target!

The "delivery time" is a critical-to-quality parameter from the customer perspective and has a significant impact on profits. In addition, it is an entry barrier for the competition. Such a parameter is called a CTQ and its definition in context of our pizza shop is given below:

CTQ Name: Timely Pizza delivery
CTQ Measure: Time in Minutes
CTQ Specification: Delivery with 30 minutes from the order acceptance time

Now we can easily define a defect:

Defect: Delivery that takes longer than 30 minutes
Unit: Order
Opportunity: 1 per order i.e. only "1" defect can occur in "1" order

Technical Note: This discussion on the example did not include 1.5σ process shift during the above analysis. The concept is discussed later. The Six Sigma conversion graph including a 1.5σ shift in process is given below:

This graph is on a logarithmic scale. Notice the increasing rate of improvement. For example, 1 sigma to 3 sigma is only 10 times improvement; 3 sigma to 4 sigma is a big 10 times improvement; whereas 5 sigma to 6 sigma is a whooping 1825 times change. That is why we are talking about breakthrough improvements in a journey to Six Sigma.

How does Six Sigma work?

The driving force behind any Six Sigma project comes from its primary focus - "bringing breakthrough improvements in a systematic manner by managing variation and reducing defects". This requires us to ask tougher questions, raise the bar significantly, and force people to think out of the box and to be innovative. The objective is to stretch, stretch mentally and not physically. To make this journey successful there is a methodology(s) to support Six Sigma implementations.

There are 2 potential scenarios - (a) there is already an existing process(s) that is working "reasonably" well and (b) there is no process at all. A bad process is as good as no process.

Scenario (a) focuses on significant process improvements and requires use of DMAIC:

1. Define process goals in terms of key critical parameters (i.e. critical to quality or critical to production) on the basis of customer requirements or Voice Of Customer (VOC)
2. Measure the current process performance in context of goals
3. Analyze the current scenario in terms of causes of variations and defects
4. Improve the process by systematically reducing variation and eliminating defects
5. Control future performance of the process

Scenario (b) focuses on process design using Design For Six Sigma (DFSS) approach. DFSS typically requires IDOV:

1. Identify process goals in terms of critical parameters, industry & competitor benchmarks, VOC
2. Design involves enumeration of potential solutions and selection of the best
3. Optimize performance by using advanced statistical modeling and simulation techniques and design refinements
4. Validate that design works in accordance to the process goals

Note, sometimes a DMAIC project may turn into a DFSS project because the process in question requires complete re-design to bring about the desired degree of improvement. Such a discovery usually occurs during improvement phase of DMAIC.

In addition to the methodology, what counts in this journey is being smart and innovative

This graph is on a logarithmic scale. Notice the increasing rate of improvement. For example, 1 sigma to 3 sigma is only 10 times improvement; 3 sigma to 4 sigma is a big 10 times improvement; whereas 5 sigma to 6 sigma is a whooping 1825 times change. That is why we are talking about breakthrough improvements in a journey to Six Sigma.

How does Six Sigma work?

The driving force behind any Six Sigma project comes from its primary focus - "bringing breakthrough improvements in a systematic manner by managing variation and reducing defects". This requires us to ask tougher questions, raise the bar significantly, and force people to think out of the box and to be innovative. The objective is to stretch, stretch mentally and not physically. To make this journey successful there is a methodology(s) to support Six Sigma implementations.

There are 2 potential scenarios - (a) there is already an existing process(s) that is working "reasonably" well and (b) there is no process at all. A bad process is as good as no process.

Scenario (a) focuses on significant process improvements and requires use of DMAIC:

1. Define process goals in terms of key critical parameters (i.e. critical to quality or critical to production) on the basis of customer requirements or Voice Of Customer (VOC)
2. Measure the current process performance in context of goals
3. Analyze the current scenario in terms of causes of variations and defects
4. Improve the process by systematically reducing variation and eliminating defects
5. Control future performance of the process

Scenario (b) focuses on process design using Design For Six Sigma (DFSS) approach. DFSS typically requires IDOV:

1. Identify process goals in terms of critical parameters, industry & competitor benchmarks, VOC
2. Design involves enumeration of potential solutions and selection of the best
3. Optimize performance by using advanced statistical modeling and simulation techniques and design refinements
4. Validate that design works in accordance to the process goals

Note, sometimes a DMAIC project may turn into a DFSS project because the process in question requires complete re-design to bring about the desired degree of improvement. Such a discovery usually occurs during improvement phase of DMAIC.

In addition to the methodology, what counts in this journey is being smart and innovative