Soft Skills

Business improvement with Six Sigma

Application of Six Sigma methodology improves efficiency and effectiveness of processes resulting in customer satisfaction, writes Dr Anirban Basu.

A software development organisation in Banga-lore found that schedule slippages in the 12 large projects executed in 2004 varied from 5 to 60 percent. Any schedule slippage beyond 20 percent was unacceptable as the management kept a provision for 20 percent time and cost overrun. There was pressure on project teams to keep the difference between actual and estimated time and cost as low as possible and within budget.

Another IT company was involved in providing tier two support for an ERP package developed and marketed by their US client. The organisation, a CMMI-level 5 company found that although the SLA (Service Level Agreement) with its US client specified that the support time for most critical (level one) complaints should not exceed 6 hours, on the average, it was taking 8 hours with variation from 2 to 14 hours. Further, its US client told them that the support time needed to be reduced to 4 hours because of strong demand from end-users for better service. The organisation was at a loss to find out how to improve the quality of their services and achieve the target specified by their client. Retaining this contract was extremely important for them as 30 engineers were involved in providing tier two support for this product.

Yet another software development company in Bangalore developing accou-nting software was receiving complaints from their customers that the number of bugs being detected in the field during usage of their product was much higher than expected. The management decided to analyse the reasons for defects after verification and validation phases. The company would lose their market share if the situation was allowed to continue.

These are the typical problems reported by companies, who in spite of having received certifications like ISO 9001:2000 and SEI CMMI for adhering to process standards, are finding it difficult to meet the increasing demands of business. They realise that they need to do lot more than implementing process standards. Most software companies even after implementing all the practices prescribed in SEI CMMI V1.1 levels are not able to meet the increasing demands of the business. Problems arise :

• in requirements development process needing rework to incorporate customer expectations
• in project planning phase due to time and cost overruns because of fallacy in the estimation process
• in the field (after release) when defects which escaped the testing process are detected by the end-user and so on.

Software Six Sigma comes to the rescue in all these cases where improvement by qualitative means do not suffice. In an organisation, pursuing implementation of CMMI Staged Representation, Six Sigma methodology supplements the specific goals and practices prescribed in levels 4 and 5 of SEI CMMI (Staged Representation), namely application of statistical techniques to reduce variation, organisational innovation and deployment, causal analysis and resolution.

This is exactly what the three IT companies did. They engaged a consultant who trained them on statistical techniques and Six Sigma methodology. After diligent implementation of Six Sigma methodology, the software companies could improve the performance of the relevant processes and meet client expectations. They could not only retain their clients but also go in for expansions!

Six Sigma methodology results in reduction of rework, avoidance of schedule slippages and cost overruns. Proper application can prevent losses and generate substantial revenue. It is therefore more a business initiative than a quality one.

Application of Six Sigma methodology generally follows the DMAIC (Define, Measure, Analyse, Improve and Control) approach. It is based on a closed loop control system, where data is analysed and processes are refined depending upon the output of the process so as to reduce the variation to a negligible level.

In DMAIC approach, the first phase is define which defines the problem, identifies the customer, the parameters of interest i.e., the Critical to Quality (CTQ) factors. The word, process, means a series of steps involved in transforming inputs to outputs and all other factors affecting the transformation namely people, tools, etc. The process to be improved is defined by developing a SIPOC which is an acronym for Supplier-Input-Process-Output-Customer.

The next phase measure involves specifying the measure of the process in terms of defects per million opportunities (which gives the sigma rating of the present process). In this phase FMEA (Failure Mode and Effect Analysis) is done and high RPN causes are identified. Histogram is made for variable data and DPMO (defects per million opportunities), long-term and short-term sigma ratings are calculated using descriptive statistical measures.

In the analyse phase, the data collected and the process map developed in the previous phases are analysed to determine the root causes of defects, identify opportunities for improvement, find gaps between current and target performance, identify root causes of defects and sources for variation. The process is analysed to remove non-value added activities, if any.

In the improve phase, the target process (i.e. the process under consideration) is improved by designing creative solutions to problems taking into account the result of the analysis performed in the previous phase. Inno-vative solutions are found to improve performance and implementation of pilot solutions are planned.

The last phase is called the control phase, and involves controlling improvements on the new process suggested in the improve phase. It requires documenting and monitoring the improvements planned so that the process does not revert to the “old process”, and institutionalise the suggested improvements.

Application of Six Sigma methodology improves the efficiency and effectiveness of the processes resulting in customer satisfaction. Satis-fying customers by delivering a better product and services without time and cost overrun contributes to increase in volume of business. Although software organisations emp-hasise on data collection, more importance needs to be put on data analysis using quantitative techniques in order to implement Six Sigma methodology. This requires guidance from consultants who have expertise in statistical techniques and have an excellent understanding of software engineering and knowledge of software metrics.

The author heads "Quality+" based in Bangalore and offers consultancy on software quality standards and software engineering.
E-mail: anirbanbasu@qualityplusindia.com