Knowledge: the key to organisational survival

Citation

Raeside, R. and Walker, J. (2001), "Knowledge: the key to organisational survival", Measuring Business Excellence, Vol. 5 No. 3. https://doi.org/10.1108/mbe.2001.26705cab.006

Knowledge: the key to organisational survival

In the developed world since the Second World War, collation of information on customers and potential markets has grown and become a pervasive business force. For this to occur there has had to be a major change in the leadership philosophy of many companies from an inward-focused vision on the product, to a more holistic view of how the company, its employees, its rivals and its customers operate in the market place. The use of business information has grown with the increasing power of computing and more routine gathering of information in the so-called "information society".

Seminal in the move to incorporate information and use it strategically, has been the quality movement as promoted by Deming (1986) and Box (1994). Key to understanding quality issues is data collection and measurement. Thus statistics has emerged as a major problem-solving tool, although this in itself caused many problems and in many companies statistical information on quality has been used in a very half-hearted manner and rarely discussed in the boardroom (Caulcutt, 1987). There appears to be little awareness of statistical techniques: see, for example, Masson and Raeside (1999). This also appears to be the case in planning and forecasting techniques. Sparkes and McHugh (1983), Dalrymple (1987) and Watson (1996) found little awareness of statistical forecasting techniques, with managers preferring to rely on judgmental methods, little reviewing of forecasting performance and integration within organisations.

In one case this resulted in a major disruption to production for a manufacturer of records. The production manager thought the sales forecast too optimistic – having no supporting evidence that he order fewer parts from stores, who had insufficient parts as they had made their own forecasts. The result was no production for almost two weeks. Barriers were raised when statistical information was presented; the only quantitative information reaching senior levels in the organisation tended to come from accountants (profits and losses) and from sales and marketing, when percentage market share information and trends were discussed. A senior manager in an electronics company stated in 1992 "we did it this way in the past – it worked, so what is wrong with that". He saw no need to do any more statistical analysis than to use defect rates as a club for the production manager. The manager was oblivious to the huge dynamism of change that was occurring – since then his company has been taken over twice and is now less than half its size.

Why is use not made of statistical information?

• •

  It is not well understood just what statistics is – even by those who teach it! It is not realised that statistics is an application of the scientific method and is a problem solving method (see Rugtagi and Wolfe, 1982).

• •

  There is little awareness of many of the techniques (Weil and Vardeman, 1992).

• •

  In Western culture there is a preference amongst students for subjects which are not based on scientific rigour – "trendy subjects" are creative arts, people-centred subjects, business studies and IT. In the East, concepts of probability can (in some societies) be viewed with suspicion.

• •

  Senior managers do not use statistics and avoid using statistical information, preferring to rely more on their own "expert" judgement, or bias.

• •

  Statistics is badly taught – often by mathematicians who respond to market demand reluctantly. A common flaw in statistical teaching is to concentrate too much on detail and mathematical proofs, at the expense of the concepts and the design and implement solutions (see Hoerl et al., 1993).

• •

  Statisticians have often over-complicated either the problem, the solution, or both. There are many instances where a statistician has completely ignored the problem requiring solution to pursue a more "interesting" one. If data are available, then the most sophisticated technique possible tends to be used, with simpler methods tending to be downgraded in importance. In our experience simple descriptive statistics, and good design of tables and graphs, can effectively deal with, and increase knowledge on, the vast majority of cases.

• •

  Statisticians have earned a reputation of being too negative and too inflexible – preferring to live in an idealised data-rich normally distributed world.

• •

  Many managers just do not understand the statistical techniques and more training is required (see Dale et al., 1993).

Thus there are faults with both how statistics are perceived by society, and how statisticians view society. To embrace the "information society", and to ensure businesses prosper, education has to improve. All business, management and engineering programmes designed for leaders should have modules on statistics taught by statisticians. Statisticians must be made more aware of what is required of them by business, and spend time becoming aware of the nature of business problems. Senior management must be more committed to the use of statistical information and demonstrate this commitment by making decisions based on this information.

Fortunately, there is evidence that this change of culture is beginning to occur, and can be demonstrated in three case studies.

Electronic contract manufacture

This is a very competitive industry competing in a global market. Cost and quality are key drivers in this market place. In analysing how these companies operate Walker (1999) found that Porter's five competitive forces (ease of market entry, customer power, amount of substitutes, supplier power and extent of rivalry), apply as external influences on the company, but are insufficient alone to inform the company in the market place. To Porter's model the following five internal forces should be added: customer focus; communication; core competencies; complexity; and quality.

These all need to be managed, and to manage them they need to be measured. Hence the need for knowledge acquisition and statistics. This extends Quintas et al.'s (1997) view that knowledge management is the main competitive force. To collate this information, Walker (1999) quickly found that there were structural differences in information inputs and that different divisions in the company worked independently with respect to knowledge management. The most sophisticated of the divisions was production, in which a complex data gathering tool, capable of real-time problem identification and reporting, was developed. This tool also produced headline performance indicators for reporting to the board. To integrate the various information sources Walker used the European Business Excellence Model (EFQM, 1999) as displayed schematically in Figure 1.

Walker found this to be an excellent knowledge collection, analysis and management model and a once data-sparse company in the areas of its key performance drivers is now gradually moving to a data-rich position in which information is produced and managed in an integrated way.

Figure 1 The European quality model

Other electronics companies are also making greater use of information, and in the micro-electronics industry huge time and process stratified databases of several gigabytes are being created from their process control information. Multivariate statistical methods are being used for analysis and informing on processes. This, combined with designed experiments, has led to vastly increased rates of knowledge discovery. This discovery process is limited by shortages of statistical skills in the micro-electronics industry, who typically do not employ statisticians. Support is provided by various consultancy services, and industry bodies such as SEMITECH. A business need has been identified and will be addressed in the near future. One way of addressing this is to use automated methods such as data mining. Data mining is discussed in the next case study.

Financial services

This is a burgeoning sector where great profits are to be made, but also subject to company take-overs and acquisitions. Competition is fierce, and along with product quality and product innovation, a major competitive driver is knowledge of the customer. This has led to a concentration on customer focus which has led to a conjunction between quality and marketing to customer relations management (CRM) (see Brown, 2000). An example of this is given by Smith et al. (2000). All financial institutions now have large data warehouses of customer information; their demographics, their buying patterns, their behaviour as customers and lifestyle variables can be bought. There is a market for customer information which is frequently traded; companies are even acquired for their customer information.

CRM is "trying to identify a customer's demand before they even think of it". To do this the information is derived from statistical models of the customer. Thus employment opportunities for statistical analysts have soared, prompting one senior banking manager to state "Career-wise, banking qualifications are becoming increasingly irrelevant, a qualification in statistics is more useful."

Statistical methods of classifying customers, such as K-means clustering, discriminant analysis and logistic regression are being used to segment databases of millions of customers and hundreds of variables. Identifying the characteristics of segments allows an institution to stratify the management of customers based on principles of quality (satisfaction), cost/profit, and future profitability potential.

A solution to the skills shortage is to make more use of automated tools – notably various data mining packages such as the SPSS product Clementine, and the SAS product Enterprise Miner. Data mining has been described by Aaron Zornes of the META group as a "knowledge discovery process of extracting previously-unknown, actionable information from very large databases". For more information see the SPSS Web pages on data mining (SPSS, 2000). The use of these applications has produced encouraging results but is not a substitute for conventional statistical analysis. Data mining approaches often suffer from the following:

• •

  Naïvety – not really allowing business knowledge to be incorporated and sometimes allowing absurd relationships to be formed. Issues such as causality are often not well handled. Spurious patterns and relationships can be generated (see Chung and Grey, 2000).

• •

  A tendency to over-react to non-representative groups.

• •

  A need for extensive data cleaning, which is not often done and brings with it issues of subjective decision making.

• •

  Diagnosis of efficacy and reliability is not easy in some implementations.

• •

  Customers change behaviour, and the value of some information input is questionable, especially if it is more than a few years old.

• •

  Data mining relies heavily on neural networks – why things happen is often not clear.

• •

  Approaches are often model-free, and only large effects tend to be found (see Jorgensen and Gentleman, 1998).

There is a need for an integrated approach (see Figure 2).

Information is currently being used strategically in financial institutions, but the rapid growth in the analysis of it will, we fear, lead to some disasters, especially if there continues to be more reliance on data mining at the expense of the rigour of the scientific method.

Figure 2 Customer relationship management

A senior analyst for a major UK financial institution stated "CRM and DM promise much but often deliver little". He went on to point out, however, that with the need for speedy solutions, an 80 percent correct model is good enough, and certainly a lot better than relying on some expert's judgement.

Information and statistics has also caused a revolution in the areas of credit and behavioural scoring (see Thomas, 2000).

Education

In the UK there currently appears to be an over-provision of higher education, and UK universities are having to become aware of the need for rationalisation and performance assessment. This has created the need for more informed decision making by universities. Statistical forecasting techniques are now being used to identify areas to withdraw from and areas to enter. Market analysis is increasingly being used to inform student recruitment campaigns. Students as customers are becoming the focus of attention. Their views and satisfaction are assessed and most universities now carry out student satisfaction surveys. Charts like the one in Figure 3 are increasingly becoming the subjects of meeting of governing bodies. This has been in operation at Napier University and has led to an increase from 70 percent to 80 percent of students who would recommend Napier to their friends.

Figure 3 Student satisfaction

To conclude, the strategic acquisition of data and its analysis and management, are vital in the knowledge discovery process. Knowledge will become, in our opinion, the key differentiator between successful enterprises and those that fail in the early part of the twenty-first century. Statisticians should have a key role in this and to be seen as agents of change in promoting the "learning organisation". For this to happen there has to be an improvement in the education of statisticians (see Strickland, 1996), and a greater appreciation by society of the worth of statisticians as wealth generators.

Florence Nightingale wrote in the 1860s:

… the most important science in the whole world for upon it depends the application of every other science, and of every art: the one science essential to all political and social administration, all education and organisation based on experience, for it only gives exact results of our experience.

Robert RaesideSchool of Mathematical and Physical Sciences, Napier University, Edinburgh, Scotland, UK

John Walker Director of Quality at Solectron Scotland Ltd, Dunfermline, Scotland, UK

Action points

• •

  To embrace the "information society" and ensure business prosperity there is a need for improved education.

• •

  All business, management and engineering programmes designed for leaders should have modules on statistics taught by statisticians.

• •

  Statisticians need a greater awareness of what business requires from them.

• •

  They need to focus more closely on the nature of business problems.

• •

  There needs to be greater commitment from senior management with regards to statistical information.

• •

  This commitment can best be demonstrated by their making decisions based on such information rather than on their own judgmental methods.