The Hindu Business Line : Six sigma demystified

SIX sigma is a word that you have most probably heard. Though GE, during Jack Welsh era, was credited with popularising six sigma, it was Motorola which started the six sigma problem solving methodology in the early 1980s and saved billions of dollars by its application in manufacturing.

What is six sigma?

Sigma is a Greek word that stands for standard deviation, which measures the amount of variation in a process. A process is any set of activities that works on some inputs to give an output. For instance, the process of delivering a document by a courier agency works on an input — the document received by the agency — to provide an output, which is delivering the document.

The critical parameter for the courier delivery process is the time taken to deliver the document. The variation in time taken to deliver is a function of how consistent the delivery process is.

Applying six sigma in the process of delivering documents will result in just 3.4 defects per million. The defect here being defined as any delivery that happens after the customer defined time. For instance, the customer requirement could be that the delivery needs to happen within 24 hours. This means that if the delivery process was six sigma then only four out of a million will be delivered beyond 24 hours.

Similarly a five sigma process will result in 233 defects per million, a four sigma process would result in 6,000 defects and so on.

The document delivery process can be replaced by any other process whose output is measurable, and the same techniques applied.

For example, the number of defects per car produced, the number of error-free transactions in a BPO company, the time taken for recruitment in an organisation and so on.

The key word for the six sigma as a tool or a goal is data. If something can be measured it can be improved upon using statistics.

Now that we have understood what six sigma means as a goal, let us try and understand it as a problem solving methodology.

The six sigma toolkit at a very broad level consists of two methodologies. They are known as DMAIC (Define, Measure, Analyse, Improve, Control) and DMADV (Define, Measure, Analyse, Design, Verify). DMAIC is used for existing processes which need improvement. For instance, if the courier agency was delivering only 80 per cent of the documents within 24 hours, then the DMAIC methodology can be applied to improve the 80 per cent to a higher number.

DMADV is used to designing new processes or products such that they produce very little or no defects when they actually are put into use. This can also be used on current processes if it is required to re-design the process completely.

How does it actually work?

First, the problem is defined. This is followed by goal statement. Then it is the measure phase, which typically talks of how good the measurement system is for measuring the output.

For instance, is the start and end time for the document delivery consistently understood by all? Is the start time, the time when the customer hands over the document or is it the time when he calls in to say he has a document to be picked up for delivery, and so on? In this phase the variation due to the measurement system is eliminated. Then data is collected over time to find out exactly what the sigma level of the process is.

Next, the analyse phase looks at the different variables which impact the time of delivery. For instance, the time taken for collection centres to hand over documents to the hub, the number of collection centres, the number of times a document get delayed because of address errors, and so on.

Typically, processes will have variables ranging from five to 100. Statistical tools are then applied to these tens of variables and those that are critical — that is, those which have a major impact on the output of the process — are identifiied.

Once the critical variables are identified, the improve phase tries to find the specifications for those variables such that it gives the desired output. Optimum solutions are found and alternatives considered.

The most feasible solution is then piloted among a small group and the output is verified to see if it meets the goals. Once the goals are verified, the solution is then implemented across the whole process. The control phase puts control mechanisms which ensure that the solution and the improvement are sustained over a long period of time.

These improvement projects are done by Master Black Belts, Black belts and Green Belts. The Master Black Belt has the highest level of technical and process proficiency. They are the people who train and mentor other belts.

Black Belts are those who completely understand the application of statistics to business problems. Green Belts are six sigma leaders who form teams and work on improvement projects in their own areas.

Other than these, companies create their own titles to describe the work done by people.

In six sigma terminology the top management people or functional leaders are known as champions. These are the people who ensure smooth progress of six sigma projects.

Six sigma as a culture

There are a few companies in the world where six sigma is a culture rather than a problem-solving tool. These organisations have moved away from instinct-based decision-making to data-based decision-making.

When six sigma became a way of life at many multinational giants such as GE, Motorola, Dow Chemicals and Johnson Controls, profits amounting to billions of dollars were raked in.

Six sigma as an initiative can be integrated into any other quality initiative such as ISO, TPM, TQM, QS and so on.

Six sigma as a business-strategy-enabler far outweighs the benefits of any of the existing models of quality or process improvement. With the Indian business sector proposed to grow at a steady pace it will be up to every entrepreneur to look inwards, spruce up the performance and meet competition head-on.

(The author, with considerable experience in six sigma implementation, is Founder, Synagoge Knowledge Services. He can be contacted at gopalkrishnakulkarni@rediffmail.com)