I'M-Europe Home Page IMPACT Programme Home Page Back to QA Home Page

First published in Managing Information, September 1995, pp 35-37 and reproduced with the permission of the Editor.

Managing the Quality of Information Products

Norman Swindells, Ferroday Ltd

Introduction

Many of the professional customers of the information industry are expressing strong concerns about the quality of the products and services that they purchase and some suppliers are responding to these complaints with positive actions to improve customer satisfaction. Over the industry as a whole however there appears to be widespread confusion about the topic of 'quality' and this confusion will increase as the growth of the information society creates new products and services for a wider customer base. Manufacturing industries have developed effective procedures for managing the quality of their products to ensure customer satisfaction and the information sector could learn a great deal from their success.

What is Quality?

The word 'Quality' itself is the source of much confusion because it is commonly used in everyday speech to imply 'excellence', determined in some way by the personal judgement of an individual or small group. In this sense the word has a subjective value which is difficult to describe but which most people believe that they understand and can recognise. In manufacturing industry, and now increasingly elsewhere, quality has a more specific meaning where it is defined as the establishing the characteristics of a product or service so that the requirements of the customer are satisfied at an acceptable cost. In this sense quality is an objective value which is capable of being measured. The level of this value is determined by agreement between the customer and the supplier, either implicitly or explicitly. It is this objective meaning of quality which is used throughout this article.

Quality, defined as meeting agreed customer requirements, has two main attributes:

Manufacturing industry has learned that, in order to achieve customer satisfaction with its products, it has to take actions on two fronts:

These general concepts apply to information products just as well as to all other outputs of human endeavour.

Managing Customer's Expectations

The management of customer's expectations is achieved by the use of a specification for the product which describes what the customer can expect to receive[1]. A specification is a description of a product using terms that the customer can understand and which are relevant to the intended use of the product. This relationship of the specification to the relevance is very important because it manages the risk to the supplier that the product will be used for a different purpose and fail as a consequence. The specifications of some manufactured products are standardised by either regional or international agreement to give further protection for the purchaser. Conformance to a standard can be used as a defence by the supplier to reduce the effect of claims for legal liability for defective products.

The existence of a specification also requires that there should be tests which can determine if the product conforms to the specification. If the tests of conformance are satisfactory then the product may carry a label or other identifier to certify this result. The use of labels to certify conformance to a specification is therefore the tip of a very large accumulation of knowledge and experience and in particular requires the development of standardised test procedures. Labels to certify conformance are therefore the end-point of the quality assurance procedures and not the start[2].

If the product does not conform to the specification then it will contain defects which are deviations in the quality characteristics from the values set in the specification. A product with one or more defects is a defective product but allowances may be made for some defects to be less serious then others. An example of a product from the information industry would be a record in a data base of postal addresses and telephone numbers. A wrong character in the postal address would be a defect but if the address could still be understood by a human interpreter then it may still be acceptable if it would not cause a catastrophic failure in communication. An error in one digit in the telephone number would be a catastrophic defect because the defective number could not be interpreted correctly by a machine and an attempt at a connection would fail. A specification for this record could therefore give more importance to errors in a telephone number than to wrong characters in a postal address.

The expectations of customers for a product will also depend on the use they intend for it. So, attempts to rank information products by using lists of quality characteristics such as: accuracy, timeliness, accessability, cost, completeness, etc [3,4] and to use the ranking as a general measure of the quality of the product will not be helpful to the supplier because everyone could have a different order of importance and different values for the characteristics. Averaging responses to questionnaires based on such lists of characteristics would be meaningless for the same reason. The relative importance of quality characteristics reflect the personal requirements of an individual purchaser.

Managing Levels of Defects

All human actions are subject to variability and mistakes and it is inhuman to expect otherwise. The creation of products with no defects is therefore unattainable. Manufacturing industry has learned instead to manage the level of defects within controlled limits so that the product still satisfies the customer's requirements and conforms to a specification but can be supplied at an acceptable cost. These tolerance limits tend to reduce over time, with improved knowledge and experience of the production process, so that the level of defects becomes less and the quality improves for the same specification. Nobody is content to produce work for scrap and everyone wants to feel that what they do has some worth. The procedures for continuously checking for conformance to a specification are therefore applied to the process of production and support the natural human instinct to want to do good work, while recognising that people will occasionally make mistakes.

The procedure for checking for conformance to a specification during production is called statistical process control (SPC). Statistical process control has four component actions:

Inspecting 100% of products will not ensure the detection of all defects because mistakes will also be made in the inspection process. Procedures have therefore been devised which use statistical methods to establish the optimum size of a sample which should be inspected in order to determine the probability that a batch of the product should be accepted or rejected at a given AQL. These procedures of statistical process control may safely replace 100% inspection of the result of any process, including data input, and are fully described in international standards[5]. The procedures are simple to apply by using tables from the standards without any need to understand the underlying mathematics. The procedures are applied at the point of production so that the operatives are responsible for checking the quality of their own work. The standards also allow for different classes of defects and for different actions to be taken as a consequence of their detection.

The standardised procedures for process control are applicable directly and without modification to the sampling and inspection of database records for errors and are already used by some companies for this purpose. The main benefits are: reduced costs of inspection, a known probability of success, a reduction in costs from re-work.

The Quality of Management and the Management of Quality

The extensive publicity and official support for the adoption of the ISO 9000 series of standards has probably also contributed to the confusion over what quality is and how it is established. The ISO 9000 series of standards provide guidance on some aspects of the quality of an organisation. Conformance to the ISO 9000 standards by the organisation does not ensure that a customer will be satisfied with the products of that organisation. The previous sections have shown that a complementary set of actions is needed to manage the quality of the products. The combination of the two sets of procedures forms a quality system which providesquality assurance for the customer that the product conforms to it's specification. Both sets of procedures are necessary to enable this to be achieved.

Some Special Problems in the Information Industry

The whole process of achieving quality assurance for products therefore depends on the existence of adequate specifications for the product or service that the customer is paying for. The first special problem is that the information industry has only limited experience of devising the specifications that it needs to manage customer's expectations and to control the level of defects in its products and services. It follows that there is also limited experience of applying tests for conformance to a specification. A particular problem is that the quality of an information product will depend on satisfaction with both the information content and the system that delivers the content and these two components may be the responsibility of different companies.

Successful attempts to specify and test the combination of the information content and the system function were made in the Materials Data Base Demonstrator Programme of the European Commission between 1984 and 1989[6]. A Code of Practice[7] was devised to provide the framework of a specification for the on-line systems taking part in the Programme. Two classes of quality characteristics were identified: those which could be tested on a Pass-Fail basis and those which needed more complex evaluation. Tests of conformance were then derived for each of these classes and were applied to the systems and their contents by a small panel of experts. This is believed to be the first time that independent testing for the quality assurance of information systems had been done.

The success of this limited experience showed that the methodology works but that more needs to be done to devise methods of describing systems and their contents so that the description can form part of a specification which is testable. However the Code of Practice has since been found to be of more general application than was originally anticipated and could be used for any on-line system.

Conclusions

Manufacturing industry has become very successful in achieving a high level of customer satisfaction with its products by using standardised procedures to minimise defects and ensure conformance to specifications. There is now enough evidence to show that the same procedures can be applied to the products and services of the information industry, although there will always be special requirements to be dealt with. One problem is the lack of experience in devising appropriate specifications in this sector. A major problem is the general lack of awareness in the information industry that these procedures exist.

The conclusions for the information services industry are therefore:

References

  1. Specifying and Measuring the Quality of Information Products and Services, Proceedings of a Workshop, 8 June 1994, European Commission DGXIII/E/1, Luxembourg.
  2. Armstrong, C.J. Databases and Quality, Managing Information, 1994, 1, 11/12.
  3. Basch, R. Measuring the Quality of Data, Report of the Fourth Annual SCOUG Retreat, Database Searcher, 1990, 6(8), pp18-23.
  4. Juntunen, R. et al, Quality Requirements for Databases, Proc. 15th Online Information Conference 1991, Learned Information Ltd, Oxford, pp 351-359.
  5. Sampling procedures for inspection by attributes, ISO 2859-1:1989(E).
  6. Swindells, N., Kröckel, H., Waterman, N. (eds), Materials Information for the European Communities, Report EUR 13153 en, European Commission, 1990.
  7. Code of Practice for Use in the Materials Database Demonstrator Programme, doc. ref. XIII/MDP-0S-03, European Commission, DGXIII/E/1, Luxembourg 1986.

Home] [ Help] [ Frequently Asked Questions (FAQs) ] [ Subject Index ] [ Text Search]
Discussion forums] [ Feedback and queries] [ Europa WWW server]

Home - Gate - Back - Top - Qa2 - Relevant


©ECSC-EC-EAEC, Brussels-Luxembourg, 1996, 1997
webmaster@echo.lu