About the Author
Andrew Hiles is a Fellow of the Business Continuity Institute and a Member of the British Computer Society. Andrew was founder Chairman of the influential European Information Market (EURIM) group which supports the UK Parliament's All-Party EURIM Group in handling European legislation. His first book on Service Level Management was published in 1991; his second book in 1993. He contributed to Croner's Guide to IT Purchasing. His software package, SLA FRAMEWORK(tm), has been purchased by leading international companies.
Starting his I.T. career with the Royal Air Force in programming and systems, Andrew moved to London Transport in an operational role. Later, in their Central Productivity Unit and subsequently as I.T. Projects Manager and as Manager of the Business Process Re-engineering function, he led several major technical and organisational reviews involving the reorganisation of key functions of London Transport. From there he moved to the Post Office as their first Business Systems Consultant with responsibilities for major projects. Subsequently as Computer Services Manager at Harwell Laboratory he provided supercomputing, mainframe, midrange and client / server bureau services and operational support of mainframe and midrange installations that Harwell facility managed. He also had Customer Support and Quality Assurance responsibilities for the Datacenter.
Andrew is a Director of the Kingswell Partnership of I.T. Consultants - an international consultancy specialising in delivering service and managing business risk. He has helped hi-tech, financial, transport and government bodies to develop and enhance Customer Support and Service Desk functions and has supported both customers and suppliers in Service Level Agreements, Market Testing, Outsourcing and Facilities Management.
Andrew is a published writer and international speaker on service management. He has presented at Cranfield, Henley, Ashridge and GEC Management Colleges and at numerous conferences in Europe, USA, Southern Africa, the Middle East, Hong Kong, the Philippines and New Zealand and Australia. He has broadcast on IT topics on radio and television.
Excerpt. © Reprinted by permission. All rights reserved.
"Increasingly, computing is being seen as a utility service just like electricity or water. These days, utilities world-wide are increasingly becoming privatised - expected to fund themselves as commercial entities rather than be provided by governments. Similarly, in-house computing services are increasingly expected to be self-sufficient, at least recovering costs from customers, rather than being provided as a corporate service as part of the overhead costs.
"As with the supply of any utility, the end user expects a defect-free service, available 100% of the time, at a reasonable cost. A utility service implies a limited range of standard products - but this is in direct conflict with the increasing technological complexity with which the Computing Service Manager has to cope and the greatly increasing range of services supported especially common user services like electronic mail, office systems and decision support systems. The Computing Service Manager is often trying to cover too much ground with too few resources: he or she needs to standardise and prioritise service offerings.
"All too often, however, this logic is not extended to the provision of the computing service. You know what the tariff rate is for power, gas, fuel oil and all the other utilities. The price may depend upon your negotiating power - but you know what you are getting for your money. You can measure the consumption of these services in terms of efficiency and select the appropriate quality: premium, unleaded or 4 star gasoline for instance. Your organization specifies the quality it requires for any consumable that it uses. This makes sense: buying goods of a higher quality than you need just wastes money.
"Again, this utility logic may not have been applied to computing services. A specification has probably been drawn up on a project-by-project basis to establish the amount of storage and computer power required. Throughput requirements may also have been stated - at least for newer systems and especially for transaction processing systems. But there are many other aspects of the services which frequently are not quantified. Often the service requirement has not been recently checked back with the user so changing user requirements may not have been expressed in the service specification.
"All this can lead to a mismatch of expectations between the end user and the computer service provider. The end user perceives that the service is 'poor". Response may be "slow". Support may be "patchy". All these are unquantified - but they suggest a background noise of end user dissatisfaction with the service - especially vocal if the end user is a paying customer. It is all too easy for the Computing Service Manager to assume the user's perception is wrong: it cannot be wrong! It is the user's perception and will remain so until it is changed.
"The Computing Service Manager, being frequently more analytical than the end user, will doubtless be measuring certain aspects of the service which are perceived as key performance indicators. As long as these are at worst consistent and at best improving, the Computing Service Manager may believe that a good service is being provided.
"This service may not however be what the customer wants! So how do we align the computing service to the needs of its customers?
"Service Level Agreements are normally associated with in-house computing facilities: external computing services, facilities management or bureaux are usually dealt with by contracts which may specify service level requirements. However, many contracts for external computing services are vague in service terms and in some cases a Service Level Agreement may supplement a contract."
