Computer Services Group Logicacmg

The purpose of this report is to identify and describe the key issues of a business problem of Computer services group LogicaCMG and subsequently employ quantitative or statistical decision tools to provide useful insights for the decision making process.

Significance of the Business Problem

The business problem identified for LogicaCMG and the Information Technology Industry was that end users were complaining that their software products and systems are of poor quality or not up to standard. The chief reason is that many software applications and systems did not undergo sufficient testing before they were launched due to time pressure and tight budget that constrained the testers. As a result of insufficient testing, bugs are prominent in many software releases and systems. In some of the more serious cases, these bugs may even cause a system failure in companies. This adversely affects those bigger corporations, as their operations will be completely retarded due to this breakdown. This will cause the corporations to incur huge amount of losses or put their reputation at stake due to failure in meeting deadlines and customers' requirements.

The end-users (individuals or corporations) of the software applications and systems have reiterated that they will no longer tolerate poor quality products. This creates tremendous stress for the software testers who have to ensure the quality of products are up to standard working on a tight budget and under strict guidelines. Moreover, if the quality of the software products does not improve by a specified time, the IT industry will be faced with a new regulation imposed by the government. This will further increase the burden on IT staffs.

In order to tackle the issue of software quality in the IT industry, LogicaCMG have to adopt the best practice approach in software testing and development. With improved product quality, the end users will be more satisfied and lesser problems will surface from the corporations. This will help enable LogicaCMG to enjoy a growing reputation that allows the company to gain a larger market share.

Aspects of the decision problem for which support could be provided through quantitative/statistical analysis

With respect to the decision problem, there are two areas in which support can be provided through statistical analysis. The first area is relating to the testing of the software products and applications while the second area is relating to the testing of the software systems that is to be installed in those bigger corporations. The appropriate statistical method used will be explained in the next section.

Statistical methodology deemed most appropriate for the decision problem

Brief overview of LogicaCMG IT testing process

The occurrence of bugs in software products and systems is a random process, the event cannot be predicted at all. Hence, inspecting every component of the software will be just be a waste of time and money. The lack of availability of testing resources and capability for them to do so further compounded the problems faced by LogicaCMG.

Statistical methodology

The statistical method suitable for the testing of the software products and systems is random sampling. This random sampling process will ensure that the interferences made from the sample data are not distorted due to selection bias. Selection biases exist whenever there is a systematic tendency to over or under represent some part of the population. Therefore, we propose that LogicaCMG should adopt random sampling to test out their software products and systems in order to levitate the time and cost constraints.

Reasons why the statistical methodology is suitable

Using random sampling, the number of bugs defect occurrence in software products and systems will be obtained and recorded. With this information, companies will ensure that IT products with excessive amount of bugs will not to be launched to the market and systems with an excessive number of bugs will be improved on. Hence, the quality of products and systems will be maintained to a greater extent.

Assumptions made for the testing process

In order to simplify the whole testing process of IT products and systems, we made the following assumptions:

Software products

1. Bugs or defectiveness are not present in all software products

2. The batch which contains at least 2 defaults/bug will be rejected without further improvements to the condition

Software systems

1. The systems are being run 1000 times to determine the number of bugs that occurred during each test run

2. Improvements will only be made if the mean (sample) is above the acceptable mean number of bugs

3. The acceptable mean number of bugs is accurate to certain extent

Data necessary for the analysis

The data needed to perform the analysis are as follow:

1. Sample size

2. Number of bug occurrences for acceptable system

3. Number of defective products in a batch in order to reject it (Tolerance level)

Data gathering process

The following describes the data gathering process:

1. Sample size

The number of samples to be taken has to be sufficiently large enough in order for it to represent the whole batch. According to the Central Limit Theorem, the minimum sample size has to be thirty. This theorem states that for any population (with finite mean Ð'µ and standard deviation σ), the sampling distribution of the sample mean is approximately normal if the sample size n is sufficiently large i.e. 30.

With this theory in mind, we decided that the number of samples to be taken from the software products to be 50 and the number of observations taken from the system to be 30.

2. Number of bug occurrences for acceptable system

The mean number of bug occurrences can be obtained from the database of LogicaCMG. The database contains the information on the number of bugs being detected in each