Six Sigma:   introductory article
Reducing variation improves quality
Six Sigma has a white-collar image, because statistics play a very important role in it. Its hierarchical management style, which has its roots in the US, contributes to this impression. But is this image correct? No. Applying Six Sigma is not as hard as it seems. Even small and medium-sized businesses can successfully use this method.

Who invented Six Sigma? The mathematician Carl Frederick Gauss (1777-1855)?

Gauss introduced the concept of a normal or standard distribution. He expressed the average spread around a target or median value as the standard deviation sigma (σ).

This kind of maths, particularly the theory of probability, is crucial to Six Sigma. After all, its goal is to reduce variation in the outcome of business processes and, consequently, to maximize the chance that products or services will meet customer expectations.

Gauss laid the solid statistical base that made it possible to measure variation in processes. Credit for the introduction of Six Sigma as an ongoing improvement tool however goes to an American engineer at Motorola, Bill Smith. Later on, Six Sigma grew to maturity within General Electrics, led at the time by Jack Welsh. He achieved huge savings using Six Sigma.

Measurable
Six Sigma is not about the reduction of variation in general. It is clearly focused on the aspects that customers find important, attractive or distinctive. Market research should determine what these critical to quality (CTQ) aspects are.

Six Sigma starts off from the viewpoint that critical quality should be made measurable first. Once you’ve got that, you can constantly improve your processes, striving for perfection. What the customer regards as CTQ should always come first. The chance that a given business process will meet customer expectations should be increased to 99.9997% at least. In mathematical terms, this means that the standard deviation of a particular process should be within Six Sigma (6σ). Once this is achieved, customer satisfaction will increase and, as a result, so will operating profit.

Project management
Not only products can meet customer specifications, but also things like delivery times or service levels.

This means that Six Sigma can be used to improve almost any activity, with as only boundary condition that you can measure the quality of the process involved!  According to Six Sigma consultant Wim Hoogbergen, this method is theoretically suitable for any company and any business process, from production to invoicing!

'The first Six Sigma project took place in the 80s, supervised by Mikel Harry, chief designer at Motorola', Hoogbergen explains. 'Before that, Six Sigma was only a mathematical method to make quality measurable. Harry linked that idea to an existing project management method called DMAIC, which stands for the steps Define, Measure, Analyze, Improve and Control.'

Mikel Harry used Six Sigma and this DMAIC project cycle to set up a production line for pagers. 'At the time, these were new products, that had to meet high quality standards. Thanks to Six Sigma, Harry not only met those requirements, he was also able to deliver the pagers in only two days.'
 
Judo
After this breakthrough Six Sigma spread rapidly to multinationals such as Honeywell and General Electrics (GE). Together with Motorola, GE was responsible for the introduction of judo terms like Green Belt and Black Belt, which identify the level of knowledge and experience of a Six Sigma project leader.

Strictly speaking, Black Belts are people who work fulltime on improvement projects. This suggests that Six Sigma is only suitable for large companies, because smaller ones often can’t afford to free staff to work solely on improvement projects.

There is however no consensus on what level of knowledge Black Belts should have, or what tasks they should carry out. At Philips, for example, a Black Belt is not the same as a Black Belt at GE or Shell. 'I think it’s a pity if small businesses are put off by this American-flavored jargon. Terms like Green Belt and Black Belt just mean that project managers should at least know the basics of Six Sigma. They don’t have to spend all their time working on DMAIC projects.’

Even the term “Six Sigma” itself should be nuanced. 'The goal is not to achieve a variation of less than 6σ, but to improve quality. The name Six Sigma was chosen because it sounds so good. In addition, it emphasizes that this method strives for perfection in the eyes of the customer.'
 
Start small
Novices are advised to start with a small DMAIC project, since that is the best way of acquiring the essential knowledge needed to apply Six Sigma. 'Currently, we’re consulting for a company in the semiconductor industry. Together with their management we’ve begun to evaluate all potential improvement projects. Then we’ll investigate if and how these projects connect with the corporate mission, and if they offer enough return on investment. A third issue is that there should be sufficient human resources available to execute the projects.'
 
Following the selection of a suitable start-up project (Define), phase two of the DMIAC cycle begins: To Measure what the customer perceives as lacking in quality. 'That’s not easy. For instance, it makes a big difference how you measure and who does this. Let me give an example. For a welding spot, we determined that a diameter of less than 3 mm was critical to quality (CTQ). Then we found out that the night shift was rejecting far more welds than the day shift. It turned out that they were including a light-colored edge around the welding spot in their measurement.'
 
Not everything that might lead to customer complaints is measured. 'You should be selective. We start with a Pareto analysis. This is the principle that 80% of the problems are commonly related to only a few quality aspects. We focus on improving those key aspects.'

Analyze
After the Measure phase of the DMAIC cycle comes the Analyze phase, see also the box Six Sigma Jargon.  'The goal at this stage is to identify the root causes of the variation. This is the hardest part, which usually takes up a lot of time.'

The aim is to track down why a quality aspect sometimes falls short of expectations. 'Often people on the shop floor know the answer themselves. In that case, it’s enough to bring this knowledge to the surface.'

If that isn’t enough, you can use one of the many Six Sigma analyzing tools. Sometimes a technique called Design of Experiments (DOE) is used. 'We use DOE to gain insight into the sensitivity of quality aspects to process parameters, such as temperatures. As a test, we conduct experiments to approach – within safe limits – the boundaries of a process. Sometimes we’ll deliberately make faulty products, to observe when a deviation happens. The goal is to find a robust set of parameter values, for which the process becomes insensitive to environmental change.'
 
Once the process improvements are implemented (Improve), you mustn’t forget to ensure that a return to the old situation becomes impossible, this is the Control-phase. 'Otherwise, you risk dropping back to the old quality level'.
 
Senior management
A key factor for the success of a Six Sigma project is the presence of a "champion”, someone supportive from senior management. 'In addition, a cost-benefit analysis is crucial. Time and again the question should be asked: Will improving a certain quality aspect yield enough profit? I’m a strong believer in financial validation, which should be handed over to a financial controller. It’s also a good idea to attract a lot of attention with the first projects, by putting photos and graphs on notice boards, for example. That is a good way of spreading enthusiasm for Six Sigma throughout your organization.'

Ronald Does is not only professor at the University of Amsterdam, but also director of the consultancy firm Institute for Business and Industrial Statistics (IBIS UvA). He stresses the importance of having a well-run business organization, before you start with Six Sigma. 'Applying Six Sigma means making changes. People on the shop floor must want to make improvements. Only tolerating Six Sigma, for example, to give underperformers a second chance by leading DMAIC projects, will certainly lead to failure. People who can successfully run Six Sigma projects are often high potentials. They often have other important tasks, so it is important to free them from these.’

Process Control
Professor Does established his consultancy firm in 1994. Initially this firm was specialized in statistical process control. He recalls: ‘I was contacted by General Electrics in 1996. They were looking for someone with statistical knowledge, for input to their Six Sigma program. Being a professor in industrial statistics, I accepted their request with pleasure. I’m not only a scientist, I like to keep in touch with practices in real business life as well.'
 
'I was impressed by Six Sigma. Therefore we began to initiate Six Sigma projects ourselves at companies such as Sara Lee and DAF Trucks. Today, the thematic improvement methods Six Sigma and Lean Six Sigma are the core business of our consultancy. We are now carrying out an increasing number of projects in the service industry, the Dutch university hospital UMCG is a good example. Our focus on the service industry simply has to do with market demand, since 90% of all companies in the Netherlands operate in this sector.'
 
'Our Black Belt training consists of four weeks of theoretical study, interspersed with three weeks of practical exercises. Candidates must bring along a real-life Six Sigma project, in order to attend our Six Sigma class. After the training, we assess the candidate’s theoretical knowledge, while the management of their company assesses their practical results. Optionally, the training course can be rounded off with a theory exam at the University of Amsterdam.'
 
Ronald Does sees Black Belts as fulltime process-improvers. 'Each year, they should complete four to five projects and generate savings of fifty to a hundred thousand Euro. For people who want to do improvement projects part-time, we have a Six Sigma Green Belt course. This course is also appropriate for small and medium-sized companies.'
 
Conclusion
Six Sigma was initially adopted by companies who wanted to distinguish themselves in the market by superior product quality. Nowadays the application of Six Sigma, and especially of DMAIC, has broadened to a wide spectrum of organizations, ranging from the chemical industry to banks and hospitals.

Relatively new is the growing interest in Lean Six Sigma, a mix of Six Sigma and Lean manufacturing. Lean is then applied to reduce waste in time and materials, while Six Sigma adds ways to achieve a good and constant product quality. The result is a production or business process which is more efficient ánd better.

For an example of a company which applies Six Sigma, see the case Draka Comteq below.

case Draka Comteq
Fewer rejects in glass fiber production

Roland van Laere is head process support at Draka Comteq. This is a Dutch medium-sized firm, located in Eindhoven, that produces fiberglass cables. Van Laere says, 'Our high quality standards bring along that we have fairly high levels of product losses, due to quality issues. This makes Six Sigma extremely suitable for us. It’s an ideal method for the processing industry, where product specifications have to lie between certain limits.'

      
       ^ A Six Sigma projectteam at Draka Comteq

Lining
The production process of Draka begins by lining the inside of a hollow glass tube with about a thousand very thin layers of glass. The refraction index of each deposited layer is just that tiny bit different. To achieve this, the composition of the syngas is varied, which is a mixture of silicon tetrachloride, germanium tetrachloride and oxygen.

When this plasma-activated chemical vapor deposition process is completed, the glass tube is heated until it collapses and becomes a solid glass rod. A preform analyzer is used to verify that the rod meets all the specifications. If the rod passes control, it is shipped to a sister company in France where it is coated with an outer layer of glass.  Back in the Netherlands the glass rod, now about 8 cm thick, is oven-heated until the glass fiber can be drawn. Van Laere explains, 'First you wait until a glass droplet forms. Then you start to pull increasingly harder, until the spinning rate of the glass fiber is about 750 meters per minute.’

The end product is a glass fiber of 125 micrometers thick, with a core between 8 and 62.5 micrometers. During the pulling process, an extra coating is applied around it, for protection.

Structured way
Draka Comteq was already using statistical methods to master product quality. 'But,' Van Laere adds, 'We missed something, namely a structured way to do that. You could say that we were on the way to invent Six Sigma ourselves', he smiles. 'I’d heard of the term but didn’t quite know what it meant.'

At the same time, Ralph van Lankveld, a senior process engineer at Draka, heard about Six Sigma from a friend at Philips. Subsequently he decided to follow a training course. Meanwhile Van Laere attended a lecture given by General Electrics and became convinced of the usefulness of Six Sigma.

Van Laere says, 'Then things went really fast. Ralph started his first DMAIC project while he was still in training. That’s the recommended way to go, because it gives you the opportunity to ask the course instructors targeted questions.'

From January to June 2004 Van Lankveld was working fulltime on optimizing a production line for the multimode glass fiber “MaxCap”. This type of glass fiber is used in local area networks, which need a very large bandwidth. Van Lankveld points out, 'We chose MaxCap as our first Six Sigma project, because we were suffering fairly high levels of product losses due to inferior quality. So, we had much to gain with process optimization.'

Black Belt
During the MaxCap project, the DMAIC system bore fruit. 'During the Analyze phase it is emphasized that you should eliminate potential impact of the measuring equipment. We found that the preform analyzer measurements varied along the length of the glass tube. Therefore we saw deviations that actually were not present.'
 
The solution was to carry out the preform analysis at two set positions, and average those measurements. 'We use this mean number to adjust the glass fiber production process,' adds Van Lankveld, 'but only if needed. Now that our measurements have become more precise, we don’t make so many adjustments these days. This is an important improvement because it takes time to adjust things and you run the risk of over steering production. All in all we’ve reduced product rejects by 8%.'
 
The mean deviation of the production process still exceeds 6σ. 'It remains questionable if we can reach that target, but that isn’t our goal. We just want quality improvement.'

Blindly devoted
Because of the success of the pilot project, Draka Compaq implements Six Sigma as the standard improvement method throughout production and in the R&D department. 'Theoretically,' says Van Lankveld, 'we could also apply Six Sigma in procurement and sales, but that’s not our top priority. We don’t want to be blindly devoted to Six Sigma, because then you risk running into a dead end. It’s important to keep using common sense, besides all the statistical analyses.'

Considering the amount of Six Sigma knowledge he has gained, project leader Van Lankveld would be entitled to call himself a Black Belt. However, he prefers not to use this title.