websites and services
multimedia projects
content management systems
online learning
online directories

IT in Retail - Transformation & Globalisation
G. Herman, (c) 1997

Electronic Data Interchange has been defined as:

The exchange of structured data, using message standards, from one computer system to another, by electronic means. (EDI94, NCC-Blackwell, 1994).

This definition covers a lot of ground, but we can narrow the field somewhat by clarifying what is meant in practice by structured data , message standards and electronic means .

Structured data indicates that the data in question is organised, typically into categories such as `date , `product number , `unit price and `customer name which are themselves collected into individual electronic `documents fulfilling the same function as paper invoices, delivery notes, purchase orders, packing lists or acknowledgements. This structuring distinguishes EDI from the more familiar form of communication known as electronic mail, or e-mail, in which a message can contain any information in any order, more akin to a personal letter than a formal trade document. Not surprisingly, EDI is often referred to by the soubriquet of paperless trading , although survey evidence published by the UK s National Computing Centre suggests that a quarter of all EDI traffic is intra-organizational.

Message standards implies that the electronic documents used in EDI will conform to agreed specifications setting out their content and format. All invoices will include date, cutomer name, order number and so on; and invoices will present this data with customer name in the top right hand corner, date below that, item name in a column on the left and, again, so on. The use of such standards simplifies communication between trading partners, reduces the risk of dispute in any transaction, and makes it easier to translate between different EDI systems.

Electronic means simply refers to the fact that EDI is an entirely paper-free system. It is irrelevant to the definition just which electronic means are involved. EDI could mean transporting a computer disk between systems, or it could mean using some form of network to link the systems. The most common forms of EDI involve the use of so-called value added network suppliers or VANS - companies such as GE Information Services (and its subsidiary, INS), AT&T Easylink, IBM and Transpac. A value added network usually does two things: it receives, stores and subsequently transmits an EDI message ( store-and-forward ), and it converts EDI messages from one format to another. This means that two trading partners do not have to have their computer systems permanently connected, because the VANS computer acts as a third party holding messages for downloading at a convenient time. It also means that the EDI partners do not have to have compatible equipment or software, as long as they are all compatible with the equipment and software used by the VANS - and compatibility, of course, is one of the things a value added network supplier specialises in. The advantages are 24-hour availability, affordable access to EDI systems for users with low volumes of trading transactions, and the ability to exchange data or documents using a variety of communications protocols and message formats.

Any form of electronic trading can have business benefits by speeding up the trading cycle, improving cash flow, allowing reductions in stock depth, and eliminating many of the costs associated with paper processing and postage. The Just-In-Tme philosophy which has proved so attractive to manufacturing is increasingly being adopted by large retailers who may eliminate the costs associated with holding stock in depth by employing stock monitoring and ordering systems which are both more responsive and more efficient. Only EDI allows this sort of thing to be done in an orderly and secure fashion, increasing accuracy and reducing the costs associated with queries and errors. The important element of EDI, then, is not computer-to-computer communications - which anybody with a couple of modems and access to a phone line can do - but the organization associated with structured data and message standards.

In most cases, this has meant that EDI has been imposed on a trading community by one large purchasing organization - for example, a motor manufacturer with hundreds of dispersed suppliers, or a retail multiple like Marks & Spencer. This form of EDI is known as hub & spoke and was believed to enhance the competitiveness and reduce the costs of the hub organization. In the UK, Marks & Spencer was one of the critical innovators in EDI, even assisting in the design of the Tradacoms standard. It has enabled us to take critical hours out of a very tight schedule, a company executive told the Financial Times in 1990.

Companies like Marks & Spencer seek to control the supply chain by using EDI to organize the procedures associated with ordering, delivery and settlement. Inevitably, they make significant savings, because hub & spoke EDI is effectively the Taylorization at an inter-company level of the clerical processes involved in trading. But even so the payback period may be as long as 18 months and, according to most informed opinion, until at least a fifth of a company s trade is involved, EDI will remain a cost. The implementation of EDI systems by hub companies also imposes costs on the spoke companies, undermines their a competitiveness by tying them to one customer, and limits the choice available to the hub. Hub & spoke EDI can become a barrier to smaller, innovative suppliers entering the market. This is not just a problem for them - if EDI ties suppliers to a large purchasing organization, it also ties the organization to its suppliers: should they produce inferior goods or fail completely, there may be no ready alternative available.

In retail, the issue may be particularly important, since competitiveness often depends on being early to market with new ideas and lines. In these cases, EDI may place an extra burden on the retailer to undertake market research and promotion, and to direct production. It often encourages an active partnership between retailers and manufacturers, particularly where the manufacturers are significant suppliers - for example, Marks & Spencer and Northern Foods, or Wal-Mart and Procter & Gamble. In one noteworthy development, the UK frozen food supplier, Iceland, seems to have taken am entirely different route, arguing that its larger suppliers resist the transition to EDI precisely because they consider it a large project. Iceland has many smaller suppliers and has negotiated a deal with its EDI service provider, BT, enabling those of its suppliers receiving five or fewer orders a month to hire EDI software and hardware at a reasonable monthly rate.

Hub & spoke EDI tends to give way to the creation of sectoral communities, cohering around EDI standards. The causes are complex: initially hub organizations may gain competitive advantage from their effective control of a proprietary technology, but competition stimulates growth and VANS themselves will inevitably seek new customers for their services. National governments have also pushed for the expansion of EDI beyond closed groups of users, largely for strategic reasons. Historically, the larger EDI communities reflected the operations and business practices of their members, so that global industries like oil, shipping or electronics developed international groupings using international standards, while quintessentially national activities, like retailing, developed EDI around national standards.

In the UK, for example, the Tradacoms standard dominates, while German retailers and those of its major trading partners tend to use the SEDAS standard. This, of course, can ead to problems - particularly when organizations seek to expand across sectors or national boundaries, or when sectoral activities begin to impinge on each other. The answer is usually considered to be to develop an international, multi-sectoral standard which can be interpreted for different sectors and user communities. UN/Edifact is increasingly promoted as that standard worlwide, and is frequently adopted for new national EDI systems. The European Article Numbering Association (EAN), which is responsible for the bar-coding system used by 120,000 organizations in Europe, has developed an international EDI system called EANCOM, described as the retail subset of UN/Edifact . Edifact itself is very similar to the US standard ANSI X12, which has been adapted to different industries in North America and, to some extent, Australia. In retail, the system is called Vics (with UCS and TDCC for grocery and transportation, respectively). Major EDI users in the US include Wal-Mart, K-Mart and Sears, Roebuck. X12 will be phased out from 1997, to be replaced by UN/Edifact.

Despite the move towards internationalizing EDI standards, EDI s legacy is that it has developed as though trading relationships were confined to closely knit communities: the motor industry, retail, or transport, the UK, Europe or the North American Free Trade Area. In fact, organizations are multiply connected within the global economy. As the authors of EDI 94 - The EDI Yearbook (NCC-Blackwell, 1994) point out:

Even the largest multinational companies cannot dictate what standards are used to all their trading partners. When Colgate-Palmolive sells toothpaste to Sainsbury s or Kaufhof, it has to accept the standard laid down by the customer.

But Sainsbury s and Kaufhof subscribe to different standards, which are both national and sectoral in character. Then, of course, Sainsbury s may belong to several different sectors and operate across several different countries. As a distributor, Sainsbury s may be required to use standards relevant to the transport or motor industries; but it is also involved in manufacturing, agriculture, and finance. And, on top of that, Sainsbury s suppliers may all use their own standards. Nestl‚ UK, for example, uses the Tradacoms standard for its local suppliers, but EANCOM for international trade. Sears, the UK store chain with interests in shoe and clothes retailing, and department stores, also uses Tradacoms with its UK suppliers, and EANCOM for internal traffic with its factories in Britain and abroad.

However, even a single international standard would have to reflect different business practices and operational realities. Until every business everywhere in the world trades in exactly the same way, there will not be a universal EDI standard, although there may yet be a single family of EDI standards. There is an argument that there ought not to be a single EDI standard at all, or that if there is it should be a lowest common denominator, only describing the bare minimum that every enterprise can agree to.

The argument goes like this: The most successful EDI implementations are those which have stimulated cooperation between trading partners. Some retailers share sales data with suppliers so that the suppliers may plan production schedules more accurately. Successful EDI means, in the phrase coined in the mid-1980s by US consultant, Michael Hammer, business process re-engineering and, ultimately perhaps, structural readjustment within a sector. Instead of retailers issuing orders to a supplier which may fluctuate for no apparent reason, the retailers provide the supplier with actual sales figures as well as orders. The supplier no longer has to gamble with production levels and scheduling in order to attempt to plan for future levels of demand with no hard information, but has actual figures with which to determine market behaviour. This leads the way for suppliers and retailers to change the nature of their relationship - instead of simply fulfilling orders, the supplier could undertake to maintain a specific service level, in return for which the retailer would undertake to promote, market and distribute the supplir s products. In these circumstances, a universal standard EDI format would be unnecessary - rather retailer and supplier would want a system tailored to their specific situation.

Structural adjustment has other ramifications, however. In a paper at the 1993 conference of the European Commission s Trade EDI Systems (Tedis) programme, Ian Graham of the University of Edinburgh and Dr. Juliet Webster of the University of East London, quoted the US economist Oliver Williamson who has argued that transaction costs are the major determinant of industry structure. Williamson descibes transaction costs as the friction disrupting the interface between organizations , and Graham and Webster argue that EDI s impact on transaction costs will affect industry structure. In particular:

One would expect EDI implementation to be accompanied by activities which have been integrated within a single firm being disaggregated into trading relationships. (Forming EDI Trading Communities, Tedis Conference Paper, 1993)

Such a trend appears to be detectable in retailing, where centralised warehousing and contracted distribution appear alongside EDI, but it is difficult to assess to what extent the trend is a consequence of EDI, to what extent EDI is a consequence of disaggregation (or disintermediation as it has also been called), and to what extent the two are rooted in a common antecedent. For example, some retailers have been disposing of manufacturing interests for many years, often as a result of a strategy of contraction aimed at reducing costs and concentrating effort. The same motivation might drive the adoption of EDI. According to one view, the overall productivity of an organization is reduced as the degree of vertical integration moves outside a desirable band (Paul Turnbull, `EDI: Innovation and organization , in EDI Technology, ed. Mike Giffkins, Blenheim Online, London 1989, p.241), and the introduction of EDI has allowed some companies to focus on the desirable band . But whatever the cause, it s clear that structural adjustment of an industry can take many forms, some of which support the idea of a universal EDI standard and some of which do not.

In fact, standardisation usually results from market dominance or international cooperation, neither of which seem sufficiently well-developed to guarantee a universal EDI standard. As with telecoms or broadcasting, the interfaces between different systems will probably continue to be more significant than the systems themselves for many years.

On balance, EDI provides a more rational approach to computer-to-computer communications within an organization. For companies with a habit of acquiring other companies, or for an organization seeking to impose internal trading accounts, EDI can act as a bridge between IT systems, yet the standards issue remains sufficiently complex to be worriesome.

In Europe, work is progressing on integrating EDI messages within the much more general standard system for handling electronic mail, called X.400. In practice, however, X.400 seems to be losing out in a transatlantic battle with the widely used but inefficient and unreliable Internet standards developed in the US. American political, technological and economic muscle will no doubt ensure that the globe will eventually be circled by a data superhighway rather than an Infobahn , but as yet there has been little discussion of integrating EDI and Internet standards (although some potentially influential steps in this direction have only recently been taken by the three leaders of the US motor indistry, Chrysler, General Motors and Ford). The importance of all this for retailers is that a single global messaging standard will `universalise EDI, making it cheaper to implement and available to any potential trader with a computer, a phone-line and the appropriate software. Cost usually emerges as the main barrier to the adoption or spread of EDI (see table XXX). In the past, the major partners in an EDI trading community have been able to exploit the exclusive nature of the systems to dominate the trading relationship. This may not always be the case.

At the othr end of the standards issue is the problem of the lack of integration between EDI systems and other IT applications. The compatibility issue is another major barrier to the adoption of EDI (see Table XXX), and integration can cost more than all the other aspects of an EDI implementation put together. Bar-coding represents a critical linkage between retail applications, but it remains only a linkage. Ideally, it should be possible to use EDI data directly within, say, spreadsheet or database applications for decision support or executive information systems, and to use spreadsheets or databases to generate data in EDI format for trading. The technical problems involved in this should not be underestimated, although the most important factor in this marked lack of integration has been the failure of software producers to understand the relevance of EDI. This is beginning to change, thanks largely to the efforts of the European Commission s Tedis and Esprit programmes, and the work of the Application Program Interface Association (APIA).

However, EDI standards are complex and may not be readily assimiliable to packaged software, or dedicated accountancy and management information systems. The complexity, of course, is necessary to guarantee the security, completeness and reliability of the data, but it may be pointless if mistakes are introduced when EDI data is printed out and then rekeyed into an accountancy system, or vice versa. Even where custom routines can be written to convert data between EDI and in-house accountancy or data analysis formats, users are often reluctant to abdicate authority to the system. This reluctance is often connected with legal or regulatory requirements to keep paper copies of trade documents which many observers believe have hampered the progress of EDI. If manual rekeying of data can introduce errors, the other alternative - manual vetting combined with printing out the data - will simply introduce delays.

In time, the advantage of being able to produce an invoice and have it immediately sent by EDI to its designated recipient, or of having a purchase ledger system capable of importing delivery notes from an EDI system, will certainly prove irresistible. In practice, however, cost, user resistance, regulations, and the lack of a single standard for EDI messages will delay this outcome.


EDI and supply chain management
B&Q is the UK s biggest do-it-yourself retailer with 284 outlets. It is now a part of Kingfisher plc, which owns Woolworth in the UK, the Comet electrical discount store chain, Superdrug, Chartwell Land and the French electrical retailers, Darty. The company, named after its founders - Block and Quayle - started in Southampton in 1969, and is the most profitable part of Kingfisher, generating profits (in 1993) of more than œ80 million on a turnover of œ1 billion.

The most important distinguishing factor in the operations of food and non-food multiples is the frequency of customer visits. While B&Q boasts around quarter of a million customers a day, 30,000 stock items, and 11 million square feet of selling space (all of which are roughly comparable to similarly-sized grocery supermarkets), the average customer will visit a B&Q store about once a month as opposed to a grocery store once a week. Individual expenditure is likely to vary more in a non-food multiple - particularly one like a DIY store which may stock individual items priced at anything between a few pence and hundreds of pounds. The result, according to B&Q s Lawrence Coppock, is a very volatile demand pattern.... Your fastest selling lines sell four times slower than a food retailer s slowest selling line.

In the 1970s and early 1980s, B&Q s stock replenishment procedures were cumbersome and slow. Sales data would be matched to store inventory overnight, generating a suggested order for each store in the morning. This would be confirmed or modified by the store management, and orders were telephoned through from each store to each supplier. The proces was costly and inefficient, and diverted staff from their essential customer-facing tasks to administrative procedures. In the mid-1980s, the company decided to modernise and centralise the procedures, implementing electronic data interchange (EDI) with suppliers.

Sales data and store inventory is still matched overnight and suggested orders are still produced for individual store management to confirm or modify. The difference is that the process is now managed by a central computer and store managers confirm or modify order suggestions on an in-store terminal. The actual ordering is done from B&Q s central office using EDI, the messaging technology in which so-called structured data , such as order forms, delivery notes or invoices, is sent electronically between the computers of trading partners.

Coppock points out that the company effectively replaced voice technology with data transmission, but the benefits were more substantial than that suggests. They included:

Cost reduction (notably a œ1 million saving on telephone calls);
Improvements to stock availability and ordering procedures;
Fewer errors and increased staff productivity;
Improvements to customer service.

Like many other instances of EDI implementation, B&Q had, in effect, to impose the technology on its suppliers. The same situation has confronted motor manufacturers and other multiple retailers (notably, Marks & Spencer) who are serviced by a relatively large number of relatively small suppliers. Like them, B&Q held a series of meetings with suppliers in different regions of the country. The company worked with INS, the third party supplier whose Tradanet EDI service provided the message-handling facilities for B&Q s EDI system. Within two years, 250 out of the company s 600 suppliers had signed on. For many, it was their first experience of IT.

In this sort of situation, suppliers are typically told to sign on with the system by a certain date or risk losing the business, but B&Q argues that its suppliers have benefited from the move to EDI, citing improved accuracy, reduced costs, the elimination of disputes, enhanced administration of order processing, greater productivity, and the opportunity to dispose of tele-sales staff as advantages.

With almost two-thirds of its suppliers live, B&Q started to look to other EDI opportunities. In particular, the company had also begun to develop a centralised distribution facility so that suppliers would deal directly with head office rather than with each store individually, and it was contemplating the replacement of the existing order-based payment system with a delivery-based system. In other words, B&Q was in the process of completely overhauling its administration. The adoption of EDI now made it possible for the company to move to a completely electronic administrative system.

The company realised that its new system would have an impact on more than simple administration, of course, although the savings in terms of administrative costs and the deployment of human resources are very substantial. (In particular, the company has reduced its data input staff and made substantial savings on paper and postage through the introduction of elecronic invoicing and remittancing). In terms of supply chain management, however, centralisation of ordering and distribution seemed at first to worsen an already difficult situation. When each store dealt with each supplier, the stores at least had a good idea of their own stock situations. Centralisation gave the company as a whole a better picture of stock movements, but it became harder to control them with any degree of precision.

As we ve already observed, sales in the DIY sector can be very erratic (with a number of well-established seasonal exceptions), replenishment lead times vary from supplier to supplier and will accumulate at different points of the supply chain, and demand itself may vary unpredictably from area to area and store to store. A small change in demand by customers at one store may be overlooked completely or may, on the contrary, be magnified throughout the system, depending larely on the availability of information in support of management decision-making. In order to smooth the replenishment process and cater for variability in demand, the centralised ordering and distribution system had buffers built in to it - points in the supply chain at which excess stock was held - but this, of course, was a dramatically inefficient way of proceeding.

In order to reduce this inefficiency, B&Q decided to work with its suppliers on demand forecasting. The idea was to share EDI-derived data with suppliers in order to derive more accurate projections of schedules. Forecasts may be for some months in advance. They are sent to B&Q s larger suppliers who feed them into production schedule models. Each month, the forecasts are refined and model production schedules revised. When B&Q eventually commits to an order, it will have a very good idea of when delivery can be expected. The suppliers also have plenty of advance notice of the likely level of demand and can plan accordingly. The process only works because information can be collated and analysed using EDI data. There must also be a willingness to communicate and routinely share that information with suppliers. While the suppliers get advanced notice of likely order levels, B&Q can get advanced notice of a supplier s capacity. This allows the company to find alternative sources in good time and to allocate scarce stock more effectively.

The benefits that B&Q has noted include:

Reducing lead times and stockholding by B&Q itself and its suppliers;
More effective planning of in-store promotions;
Improving service levels to stores and customers;
Cutting costs along the whole supply chain.

The company s most recent marketing strategy of everyday low pricing on key items has been developed on the basis of supply chain management which can guarantee to maintain adequate levels of a wide range of stock items.

Further down the line, the company looks to a full implementation of electronic trading which will allow the introduction of management by exception in many areas and the sharing of product price files with all suppliers who will then be able to update product descriptions, barcodes, and prices. One objective may be to eliminate invoicing altogether and introduce self-billing, which is only possible to the extent that reliable EDI systems can remove sources of persistent error and confusion. When supplier and reatiler share the same data, as they do in a full EDI system, topics like quantities delivered, delivery dates, and current prices are no longer likely to be sources of contention. Managers can concentrate on the business itself rather than administrative chores and crises.

It should go without saying that two conditions are necessary for this dream to work. Firstly, the EDI system itself must be completely secure and trustworthy and, secondly, there must be a willingness to cooperate - one might almost say a partnership mentality - between retailers and their suppliers. Whether either of these two is a realistic goal remains to be seen, but the adoption of electronic trading itself creates a clear imperative to achieve them.

Computer Telephony Integration
Research predictions suggest that the CTI market is poised to take off. Informal definitions of CTI suppose a telephone system linked to a computer database through a call-centre. Audiotex, voice-processing and digital techniques like touch-tones and caller-line identification (CLI - available on mobile phones and in certain large telephone network implementations) help navigate the system, identify callers, give instructions, and extract data. Voice-processing means the systems can even be used by old-fashioned telephone systems using pulse-coded rotary dialling. Existing applications with relevance to retailing are to be found in direct response TV marketing (for example, QVC), free-phone (800) services, voice-mail messaging, and automated telephone banking, credit-card authorisation or share-dealing systems.

In a typical CTI call-centre implementation, an incoming phone call is routed to the most appropriate destnation, and a human agent with access to relevant databases then deals with the call. Typically, the most appropriate destination would be the one with the shortest queue, but CTI promises much more in the way of customer-service. Already a call can be routed automatically by, for example, identifying which country or area it comes from and diverting it to the right agent. This sort of facility may rely on different phone numbers being allocated to inbound calls. But CLI will allow individual phone numbers to be recognised. Then, for example, calls from those numbers can be dealt with by the same agent every time and the caller s details can be popped automatically onto the agent s VDU screen. In the US, the Federal Communications Committee (FCC) has ruled that long-distance and local telephony carriers must exchange CLI information, and other countries will no doubt follow suit eventually. The result will be that anybody - even the smallest businesses - will be able to collect calling number information to build up a marketing database (although this remains a grey area in European data protection law).

At the other end of the system, CTI is much more than a database technology, since the host computer may be used to run any application and the telephone switch can convey images, video, and conventional data, as well as speech. For example, CTI call-centres will almost certainly be used in practical video-on-demand applications, where videophones or PC-based multimedia terminals linked to the telephone network will allow customers to dial-up a movie, browse a catalogue, enter an order, have their credit checked and payment authorised, and download the movie. At the call-centre end, data on the transaction will be captured and funds transferred. Call-centre agents may subsequently phone the customer with the news that another movie by the same director has just come out on video.

The main benefits of CTI to telemarketing operations at present are increased efficiency and an increased level of customer service. For example, the ability to route a call from a French-speaking customer to a French-speaking agent automatically decreases the amount of time lost in shuttling calls around the system and also leaves a better impression with the customer. It helps if customers can be connected to the people who can help them without delay. CTI advocates suggest that customer retention levels in telemarketing operations may be improved by a significant degree, that customer retention is substantially more effective than cost-cutting in improving performance, and that it costs between five and ten times as much for a telemarketing operation to win new customers than to keep old ones.

CTI systems also enhance the telephone operator s productivity by supporting such facilities as intelligent screen popping (the right window pops-up on the operator s VDU screen automatically), preview, power and predictive dialling (operators can save time on outbound calls because the system can dial ahead or in batches, or can cycle through calls waiting for a connection before identifying who is being called). The systems can also queue outbound calls for operators or agents to make during slack periods.

The call centre itself may be distributed around several locations, and may even be virtual. For example, British Telecom uses home-based teleworkers to run some of its directory enquiries services. With simultaneous access to callers and the BT customer database, these teleworkers can be anywhere in the world. Video links using a window on the teleworker s VDU screen can be used to overcome a sense of isolation and assist performance monitoring.

The development of standards for interfacing PCs with telephone systems will drive the growth of CTI, although as with most innovations in this area, there are contending standards - notably from Microsoft and Intel, on the one hand, and Novell and AT&T, on the other. Novell - which already accounts for around 70% of the LAN software market, and owns WordPerfect and the source code to Unix - is probably the leading player, having launched the Novell Open Telephony Association (Nota) in March, 1994, to promote its Telephone Services Application Programming Interface (TSAPI) which is basd on an international standard. Early members of Nota include Alcatel, AT&T, Dialogic, Ericsson, Global Communications, GPT, Interconnect, Mitel, Philips, Rolm, and SDX. TSAPI is aimed at call-centre applications, while Microsoft-Intel s TAPI (Telephony API) is designed primarily for single users or small businesses who wish to integrate their phone and computer systems.

CTI systems in development promise videophone home-banking from Citibank and Philips, a full range of customer services from Federal Express, and an automated order-status enquiry system from the mail-order PC company, Gateway 2000. One credit card company has already implemented a CTI system to handle retail authorisations at a rate of 10,000 calls an hour. The company has cut relevant staff levels from around 800 to 170, concentrated in one location instead of eight, pushed floor limits down to zero, and still reduced fraud. It promises voice processing so that retailers will be able to speak their own merchant numbers to the system and, eventually, a reduction in merchant fees. Companies like this feel sensitive enough about the competitive advantage offered by CTI to maintain low profiles: In some cases, according to Byte magazine s Jon Udell (July 1994, p.96), computer telephony applications are so strategic that people won t even discuss them.


Designing a retail operation around information systems
QVC s recent merger with Home Shopping Network, Inc (HSN) has created a mammoth home shopping organization with an estimated annual turnover of $2 billion. The organization brings together the promotional and entertainment skills of QVC (the name stands for Quality, Value, and Convenience ) and HSN s tele-marketing operation - said to be the largest in the world. With former head of Fox Broadcasting, Barry Diller, as its chief executive, QVC grew in the 1980s to become the fourth largest television and film network in the US and a player in the battle to dominate the multimedia entertainment market of the future. HSN was also a TV-based operation, and started in 1985 as a subsidiary of cable company, Liberty Media. The 2,000 plus telephone operators at its Clearwater, Florida, headquarters can handle up to 20,000 calls simultaneously. Before the merger, HSN claimed to receive around 200,000 calls a day in the US, about three-quarters of which converted to sales.

There is nothing new about home-shopping - it is simply a development of mail-order operations pioneered by Sears, Roebuck and later exploited by companies like Littlewoods and Great Universal Stores. The history of home-shopping goes back a century, while television-based home-shopping is, naturally, more recent. But the TV itself is no newcomer to retail. In the 1950s, TV shopping magazines were all the rage and, of course, television s reliance on advertising revenue has given it a central place in the consumer market. Three technologies have made the difference and allowed companies like QVC and HSN to grow fat - the telephone, the credit card and IT.

Home-shopping is rapidly becoming a major form of retailing, exhibiting all the key characteristics of large-scale modern retailing as whole: the transformation of shopping into a leisure activity, precise control of the supply chain, electronic funds transfer, the emphasis on marketing and consumer choice. TV home-shopping is in many ways the logical extension of modern multiple retailing, combining extreme centralisation of buying and administration with extreme dispersion and localisation of outlets (typically, people s homes). It combines scale economies with targeted marketing (which is one way of understanding why TV home-shopping invariably features a small range of products - too many goods on display would make targeting too difficult). The real importance of the QVC/HSN merger is the emphasis it now gives to IT and communications in operational, marketing, strategic and administrative roles. At present, mail order and home shopping only account for about 2% of all retail transactions, but there can be little doubt that, in time, a great deal of shopping will be done this way. Developments in communications and computing technology - particularly multimedia, virtual realit, CTI and integrated broadband networks (the much vaunted data superhighways ) - will make it possible, while social changes will make it desirable if not necessary.

While the concept of buying from your armchair is not new, QVC/HSN today is nonetheless a pioneer. Its reliance on information systems puts it at the leading edge, not just of retail but of all enterprises. As Stella Tavilla, HSN s vice-president for management information systems, has observed in discussing the company s brief but spectacular history: It wasn t a question of implementing technology in an existing company. It was a question of building business around technology.

In one sense, home-shopping is another indication of the trend towards discounting, and it remains almost defiantly downmarket. But the systems used by home-shopping channels are, in theory, limitlessly flexible. In other words, the critical success factors for home-shopping in any given market have less to do with price or presentation than with factors like the quality of the delivery system, cultural expectations and the regulatory framework. In the US, for example, mail-order delivery is frequently guaranteed within a few days, while UK mail-order purchasers make do with a few weeks. The spontaneity of buying items from a TV screen is reinforced by rapid delivery. Where that spontaneity is dissipated by a long wait for the goods, TV-based home-shopping may prove to be of limited appeal.

The information systems used by QVC/HSN are the most sensitive and flexible of any in the home-shopping arena and amply demonstrate how data can be harnessed to drive both the day-to-day operations and the longer-term development of a retail business.

The TV viewers see items for sale on the TV channel and may dial a toll-free number as they watch the show. The call is handled by telephone sales staff or by a voice response system known as Tootie that processes over a half of all orders. Tootie or the human operators take customer details, run credit checks and input the sales transaction to a large Unisys mainframe-based order entry system. Credit, charge and debit card authorisation systems are directly linked to banks and card companies, allowing almost instantaneous clearance. Six warehousing and despatch centres pull the relevant information off the system using their own smaller mainframes, orders are shipped within hours, and the system generates repelenishment information to maintain inventory at optimal levels. A core database holds information on around 20 million customers and 75 million transactions. This data provides input to a marketing system running on yet another mainframe

Order processing times, from the order being made to delivery, are a quarter of those associated with catalogue-based mail order. Customers get their goods faster, and the home-shopping company turns over stock and receives its funds quicker. Information and analysis of sales and customer preferences is rapidly available and effectively updated in real time.

At the TV station, monitor screens in the studio give presenters summary details of ongoing sales, updated by the system every ten seconds or so. The presenters can change their sales message and approach accordingly. The system can also provide sales analyses, which can be used to determine offer prices and special discounts to ensure that products always meet sales volume targets. Marketing information generated from sales data and viewing figures also helps the home-shopping channel to schedule programmes targeted at specific audience segments and niche consumers - gardeners, home decorators, housewives, schoolboys, or music fans.

Home-shopping channels use sophisticated audience and consumer profiling and rapid feedback to determine stock levels, sales strategies and target audiences. In theory, they should be able to sell to anybody who watches the TV. Without the overhead of stores or the need to carry a range of stock, the operation can be very lean. It should also translate easily across borders, by the simple device of employing local programme makers to produce the TV shows used as sales vehicles. TV-based home-shopping seems to be a triumph of form over content: the rapid interaction between viewer/consumer and broadaster/retailer drives the sales and marketing processes. As we have observed, a narrow range of goods makes it relatively easy to control both the demand and supply chains wherever they may be. Accordingly, QVC/HSN has launched a joint European venture with the UK s BSkyB and announced a joint Latin American/Spanish language venture with the Mexican Grupa Televisa. These joint ventures offer a combination of broadcasting, marketing and IT expertise, together with local knowledge. It should be unbeatable.

The Spanish language venture is clearly targeted at Latin America and US-based Hispanic communities. This is a sizeable and fairly uniform market which may respond well to a TV-based sales approach. In Europe, QVC/HSN may have more of a problem. Apart from major linguistic and cultural differences, Europe s broadcasting regulations and its controls on privacy are both complex and sometimes quite severe. Companies like BSkyB can avoid some of these problems, if only because nobody is too clear about how the relevant laws and instruments affect satellite broadcasts, but existing land-based shopping channels and programmes in Europe - such as the Swedish-based Kinnevik TV-Shop, France s TF1 Teleshopping and the US-based Quantum International (which collaborates in Europe with Italia 7 TV) - remain localised by language and culture and constrained by regulatory mechanisms which may control the output of TV channels, and the broadcasters ability to maintain databases of consumer profiles, purchases and preferences.

Satellite-based broadcasting will overcome some, if not all, of these problems, and IT can guarantee market sensitivity. Even so, home-shopping may still come up against cultural barriers which hamper its acceptability. The choice of goods to market on European home-shopping programmes will be particularly critical, and if the ability of conventional TV programmes of consumer magazines to cross borders is any indication, the emphasis will be strcitly downmarket. The home-shopping channels do not yet represent a threat to established retail operations or to their near cousins - the catalogue mail-order outlets. Both stores and catalogues carry vastly larger ranges of goods, while shops - which may stock fewer items than a typical catalogue - have the advantage of actually displaying goods and often allowing consumers to touch and try-out items before purchasing them. In time, interactive TV will bring the advantages of multimedia to home-shopping, and this promises to replace catalogues in the foreseeable future. At present, home-shopping - although full of potential - must be considered more of an entertainment than a form of retailing.

An information-based approach to retailing changes the competitive environment, creating pressures from globally expanding competitors and from new types of retailer who combine technology and customer-focus.

Most of the bad investment in IT is due to the fast-changing nature of technology and the overblown claims of IT suppliers creating confusion and inflated expectations (never expect a system to do more than 80% of what s promised).

The goal of distributed computing is at the heart of all significant recent developments in IT - this makes networks and network applications the most important areas to watch.

Distributed computing means that the right information, and the computing power to process it, can be delivered just where it is needed.

The key benefits are flexibility and the ability to manage a complex and dispersed organization, although straightforward cost savings can be made by replacing mainframes.

The single most important technological development in post-war retailing has been the introduction of the intelligent or electronic cash register (ECR); this led to the electronic point-of-sale (EPoS) terminal which is completely transforming retailing.

Networked EPoS systems are seen by many of the largest retailers as strategic weapons in a competitive war.

The combination of EPoS with other technologies like EFT, bar coding, smart cards, electronic shelf-edge labelling, V-SAT and EDI promises to revolutionise store-based retailing, and stimulate the further growth of multiples, although there is significant institutional resistanc to some of the changes involved and the take-up of the technologies complementary to EPoS has been variable.

The technology already exists for a significant extension of teleshopping and the introduction of interactive multimedia and virtual reality shopping, but more immediate benefit will be gained by using multimedia kiosks in-store to supplement existing catalogue sales operations.

Database technology allows back-office systems to provide advanced management information and decision-support facilities to enhance marketing efforts, the analysis of consumer trends and customer-targeting.

EDI continues to have a significant impact on supply-chain management, and encourages the break-up of vertically integrated operations (known as disintermediation ).

The new technology of computer-telephony integration will see the development of sophisticated telemarketing operations within three to four years; CTI will support the growth of home-shopping and will hasten the decline of store-based retailing.

As many as 50% of all retail outlets in developed countries may close within the next five to ten years, as a result of competitive pressure from globally expanding multiples, on the one hand, and the growth of technologically-driven teleshopping, on the other.

-end chapter three-

Monday, September 10 2001