ENBIS: European Network for Business and Industrial Statistics

Session organised by the Statistics in Practice Special Interest Group

21 September 2009, 14:25 – 15:25

Abstract

Submitted by

Shirley Coleman

Authors

Shirley Coleman, Titanne de Frenne, Irena Ograjensek

Affiliation

ENBIS

Abstract

This hour long session addresses three of the major interest areas of our group.
We first review a paper on Statistical Efficiency, then we have an opportunity to consider the prize winning submission to “The JMP Challenge” and finally have a panel discussion on the gap between academic training and the application of statistics in the real world in practice. The session will be informal and inclusive.
An Analysis of Statistical Efficiency
Chris McCollin, Shirley Coleman, Irena Ograjenšek and Ron Kenett
We have found that businesses tend to work to minimum requirements if the need is not seen as essential. The best known software for Six Sigma (ITIL) covers many processes that should be seen in a continuous improvement situation because it was designed around engineering design practices. It only recently included risk assessment. However, the lack of knowledge of users (among other things) means that these processes are not carried through.
Practical statistical efficiency is a method of summarizing the usefulness and relevance of a statistical intervention. The PSE model can also address the usefulness and relevance of quality improvement techniques. A scoring regime is proposed and exemplified using Taguchi analysis.
The methodology is explored in the context of the “seven deadly sins of quality management” and mapped to the Six Sigma DMAIC phases of project work. The importance of implementing a structured approach to work processes is illustrated with evidence from the Challenger disaster. It is shown how the 7 components of the PSE score relate to key methods of quality improvement, such as EVOP, QFD, SPC and DoE.
The objective is to find literature which covers each part of the DMAIC cycle with respect to the techniques and if where it does not exist, then this implies a low PSE score.
The work then proceeds to explore the ITIL systematic and professional approach to the management of IT service provision. ITIL provides consistent and comprehensive documentation of best practice for IT Service Management and claims to be widely used and to bring quality improvement. The paper aims to show the shared currency in these various approaches and to show how PSE can summarise projects dealt with in any of these ways.
The JMP Challenge

Acell assembles and sells laptops in the Far East, and, in the UK, one of its primary retail outlets is PC Universe who it started to do business with in 2008. A large number of laptop configurations are assembled and sold, involving different screen sizes, hard disk sizes, memory sizes and so on. New, improved components are being introduced almost weekly, giving the chance to build new “better” configurations. Acell’s assembly and distribution costs are more or less fixed, so key determinants of profitability are the costs of the components it sources, and the price it can charge for the different configurations it markets whilst still maintaining sales volumes. Acell is also concerned with brand recognition and is subject to the usual competitive pressures. In the past it has, within some imposed guidelines, allowed retailers to discount at their own discretion in order to move product and secure market share.

This case history looks at some point of sale data from the London area in 2008, where a significant fraction of Acell’s UK sales occur.

The challenge is to undertake an exploratory analysis to see what people have been buying, where, and how much they have been paying.

Using JMP, or other software of your choice, unearth answers to the questions above that are relevant to Acell’s 2009 business objectives. Once you start to work with the data, you may also find you need to pose and answer new questions.

Submit your findings in the way you feel to be most appropriate. We are interested not just in the findings, but the route that you took to reach them.

There will be a prize for the submission adjudged to be the best.

Panel Discussion

One of the problems facing statistical practitioners is that training does not always fit us for practice. In this final part of the session there will be a panel discussion about the gap between the academic training of statisticians and the demands the real world imposes on consultants.

Return to programme