in the industry over the intervening ten years. By contrast, private sources offered extensive analysis but were generally flawed in several ways. First, there were definitional variations from one source to another.<br><br> Second, sampling and grossing methodologies were often weak. A principal reason for this was the very low concentration of the UK software services industry in which the top 40 companies accounted for only 20% of aggregate turnover, and the total number of companies had expanded rapidly from 4,000 in 1980 to 85,000 in 1998 and perhaps more than 120,000 today. There was much less scope for error in estimating the size of the hardware industry, in which the top 20 companies accounted for 75% of aggregate turnover, and private estimates rarely differed by more than 30% from the official ones.<br><br> Third, forecasts of the growth of the IT industry and e- commerce frequently ignored inflation, the business cycle and one-off events. Indeed, when IBM approached a number of market researchers to assess the scale of opportunities afforded by Y2K, several were unaware of the latter. Yet it had been identified in the USA some 18 months earlier and been the subject of rolling public education programmes there, and some of these researchers had published forecasts beyond the end of 1999.<br><br> These methodological flaws led to wide variations in private statistics, of which Keith cited two examples. First, the official estimate of the turnover of the UK software services industry in 1998 was £30bn, an increase of 31% on the previous year, but two principal non-official sources estimated it to be around half this and annual growth as low as 13%. Second, a US company which produces estimates by synthesising other companies 9 research suggested there were 130mn internet users worldwide in 1999, whilst another company claimed 313mn.<br><br> In projecting numbers for 2005, a third company thought there would then be fewer users than this (301mn), whilst a fourth suggested that one-sixth of the world 9s population (1bn) would be on- line. In conclusion, Keith suggested ways in which the best of both official and non-official statistics could be combined. First, he commended work recently undertaken by Goldman Sachs and (independently) Cambridge Econometrics to forecast growth in e- commerce using dynamic input-output models, thereby more closely reflecting the methodology used in compiling official statistics.<br><br> Second, he urged market researchers to be more forthcoming about their methodologies. At present a typical 2-300 page report may include as little as half a page on this, and a recent example described its central assumption as c&big companies have big budgets and therefore spend more on IT than small companies& d Third, forecasts of the supplier industry in the UK, which is largely self- contained with modest net exports, should recognise that the principal limitation to growth is the skills base. Measurement of e-commerce in official statistics Robert Hay of the ONS, whose responsibilities within the Business and Prices Group include developing a strategy to measure e-commerce in business statistics, then spoke about both business and social official data.<br><br> He began by explaining that in September 1999 the Performance and Innovation Unit of the Cabinet Office had published a report cE-commerce @its.best.uk d as a follow-up to the Competitiveness White Paper. This had included a chapter on monitoring and evaluation, with recommendations for a number of investigations into the provision of statistics which were being taken forward under the auspices of a Government Statistical Service (GSS) inter-departmental group chaired by the Department of Trade and Industry (DTI). Robert 9s talk drew extensively on the cE-Commerce @the.ONS.UK d article published in the April 2000 edition of Economic Trends.<br><br> ONS 9s role was essentially to develop business and social surveys to ensure that e-commerce was adequately covered in macro-economic indicators, and the national and balance of payments accounts, and that the needs of customers such as the Treasury and Bank of England for information on the impact of the internet were met. In order to fulfil this role it was first necessary to agree internationally how to define e-commerce. Discussions Monetary & Financial Statistics: June 2000 5 initiated by the Organisation for Economic Co-operation and Development (OECD) had produced a framework sufficient to allow data collection to go ahead.<br><br> In the Business and Prices area a number of assessments of the impact of e-commerce were currently underway. First, of how the Business Register, from which populations and samples were drawn, reflected both e- commerce divisions of existing companies and new standalone dot-com businesses. Second, of how price indices were affected by differentials between the internet and other channels 3 the prices of books and toys purchased on-line were now being systematically surveyed to determine the data collection issues, though sales were currently too small to warrant inclusion in the Retail Prices Index (RPI).<br><br> Third, of the scale of activity, such as buying direct from overseas-based internet businesses, for which no data collection mechanism exists at present. And fourth of the effect of changes in business practices on stockbuilding, capital expenditure, etc. ONS have also taken an interest in a study commissioned by the Office of the e-Envoy from De Montfort University of the availability and quality of market research data on e-commerce, and the extent to which this meets business needs.<br><br> The study listed no fewer than 115 indicators of activity that might be measured. (It was suggested during the panel session towards the close of the seminar that ONS should also comment on the value of market research data, but a DTI member of the GSS inter-departmental group present suggested that the researchers might be unwilling for critical observations to be published.) In the immediate future, ONS propose to obtain inexpensive and comprehensive but highly aggregated information on internet usage from the service providers, and to identify the largest business users of e-commerce through publicly available information. In the next 12 months or so they intend to undertake detailed case studies of those industries which use the internet heavily.<br><br> Longer-term aims include the integration of e-commerce into existing business inquiries with separate questions on for example turnover and capital expenditure sourced from the internet, and a pilot survey of the enablers of e- commerce, such as software providers, in conjunction with the Computer and Software Services Association. In contrast to business surveys, social ones are already gathering information on the impact of the internet. Since 1998 the Family Expenditure Survey has been collecting data on households 9 access to the internet - figures for 1999/2000 will be available later this summer.<br><br> In April this survey was extended to cover information on goods and services purchased on-line, which in time will influence the basket used to measure the RPI, and questions on internet usage, distinguishing browsing for goods and services from purchases, were added to the Time Use Survey 3 the first data will be available in the second half of 2001. In addition, a fundamental revision of the Standard Occupational Classification to provide more detail on employment linked to e-commerce has been implemented 3 the first results from the Labour Force and New Earnings surveys will be available next year and from the Census in 2002. Internationally, few data are currently available from official sources.<br><br> The DTI has for several years contracted a consultant to undertake a benchmarking study in the G7 countries, comprising a telephone survey of inter alia the level of ownership of personal computers and internet access, and the understanding and use of the internet for business activity. And whilst the Australians have surveyed internet service providers (ISPs), and the Dutch carried out some pilot studies, only the USA has begun to publish macro-economic aggregates, commencing with an estimate of retail sales via the internet in Q4 1999. Robert completed his presentation by flagging the Office of the e-Envoy 9s next report in July 2000 which will include a progress report on developments in official statistics.<br><br> E-commerce shaping an industry John Kemble of the ABI echoed the previous speaker 9s remarks about the importance of establishing a common definition of e-commerce for measurement purposes, and quoted that submitted by the DTI to OECD - cThe exchange of information across electronic networks, at any stage in the supply chain, whether within an organisation, between businesses, between businesses and consumers, or between the public and private sectors, whether paid or unpaid. d Thus e-commerce embraced such technologies as Electronic Data Interchange (which has been used in the insurance market for some 20 years), e-mail, the World Wide Web, Computer Telephony Integration (which will develop further with the advent of Wireless Application Protocol mobile telephone technology), public terminals in banks, shopping malls, etc, and smart cards. To demonstrate that e-commerce would continue to expand, John quoted figures compiled by a market research company about a year ago. These showed that whilst only 6% of UK internet users had bought financial services on-line, 20% had browsed for information on them, and 37% of those who had made any purchases electronically expected to buy financial products in the same way in the future.<br><br> It appeared that the experience of ordering computer software, books and music/videos via the internet (between 27% and 40% of users had done this) generated confidence to purchase other goods and services on-line. (In the panel session both John and the final speaker Mark Pearson asserted that there was a growing conviction that the internet was a relatively secure transactions medium, and that it was safer for credit card details to be transmitted electronically with the benefit of 28 point encryption than for the card itself to be out of sight of the holder for several minutes as often occurred when settling a restaurant bill. There was no evidence of credit card numbers being intercepted in transmission; rather fraudsters obtained them by hacking into retailers 9, etc databases, which contained numbers supplied via all media including over the telephone and by post.) The growth of e-commerce reflects its twin roles as an enabler and a driver.<br><br> For example, over the past 30 years supermarkets have steadily been extending their opening hours and the number of lines they stock in response to consumer demand. E-commerce not only enables even Monetary & Financial Statistics: June 2000 6 wider access and choice, but also drives expectations of the levels of service, such as home delivery, and volume of product information that can be provided. And as the internet has enabled greater access to overseas suppliers of goods and services, regulators have been driven to create increasingly level playing fields for trade, which has not only enabled more international business to be done electronically in turn but also via other channels.<br><br> In the insurance industry, e-commerce is enabling fragmentation of broker roles and disintermediation, and the resulting efficiencies are driving further use of this channel. Traditionally, customers have sought general advice, obtained detailed and customised product information, and then made their purchase from one broker. The internet encourages buyers to shop around for both information and prices, and perhaps ultimately to buy direct from the insurance company itself.<br><br> And new technologies will further aid this process 3 for example XML (Extended Mark-up Language) will permit personal details (and other data) to be labelled such that they can automatically be downloaded to several sites rather than having to be typed into each one. There are also efficiencies to be gained post-sale. For example, customers can change their name and address details on-line rather than posting a form which then has to be input by the insurer.<br><br> Or by completing an electronic claim form, they can receive information on mitigating the claim together with details of the nearest approved vehicle repairer, roofing contractor, etc, and in time perhaps a copy of the form could be sent automatically to the police where this was appropriate. John concluded his talk by outlining two statistical challenges and one opportunity. First, the ABI had yet to adapt its sales channel based data collection mechanisms to measure the contribution of the internet, as the latter might be used for only one or two of the three elements of a sale (general advice, detailed information, the transaction).<br><br> Second, as the pace of change quickened, it would become increasingly difficult to compare present and historical data on a like for like basis. On the other hand, XML raised the possibility of internet transactions being labelled in such a way that they could be identified and aggregated for statistical purposes. Developing an e-commerce product strategy The final speaker was Mark Pearson, head of customer voice at Egg, who had recently assumed responsibility for desk research also.<br><br> He identified five sources of information used in his company 9s product formulation strategies. First, traditional ones such as the ONS, Council of Mortgage Lenders, British Bankers 9 Association and ABI. Second, desk research on for example market size and growth.<br><br> Because none of this was dynamic and its quality varied, it was necessary to pick and mix sources rather than rely on one continuously. Third, desk research on the social implications of new technologies. Fourth, Egg 9s own customer databases.<br><br> These revealed who the nearly 1 million customers were (name, age, gender, etc), where they were (e-mail and postal addresses and telephone numbers), their technology literacy (based on how they used the internet), size of their wallets and Egg 9s share of those wallets, how Egg 9s share could be increased (by offering services geared to stated key intentions such as going on holiday, database triggers such as change of address, etc), and the strength and depth of the customer 9s relationship with Egg (recency, frequency and value of interactions, etc). One particularly useful piece of information was a customer 9s ISP as this was a powerful predictor of behaviour 3 for example one provider 9s customers had an average savings balance of £13,600 whilst another 9s averaged only £8,000. (Though it was suggested in the panel session that free internet access offered by many ISPs had encouraged users to subscribe to several providers, thereby creating increasing overlap between the latter 9s customer bases.) The customer database was a potent tool for segmenting and pricing products.<br><br> The final source of information was continuous feedback from customers. Whilst in the run-up to a product launch desk research was vital in for example sizing the market, shortcomings and inconsistencies in this research led to more reliance being placed post-launch on customer feedback. Indeed, between 1 January and 9 May 2000 (the date of the seminar) no fewer than 32 customer research projects had been carried out in addition to the brief satisfaction survey that can be completed each time a customer goes on-line.<br><br> A significant trend in this research was the fall in expectations of what the internet can deliver 3 in particular the speed of access and response to service requests. The challenge was to bring all these sources of sometimes conflicting information together and to extract the main stories quickly. The key difference in developing a traditional and on-line product respectively was the need continuously to obtain and analyse information on the latter and to make improvements in the light of this.<br><br> This was vital as markets changed more rapidly and brand loyalties diminished. For example, the latest customer research survey had been drafted on a Wednesday, distributed the next day by e-mail, and achieved a 40% response rate by the following Monday. The results would be considered and necessary actions agreed at the next weekly development meeting.<br><br> Mark Pearson from E . Monetary & Financial Statistics: June 2000 7 Turning specifically to the launch of Egg Card, Mark explained that his company had begun to size the potential market in April 1999 using market research data showing that 22% of adults in Great Britain had internet access and 5% (of which half were on-line) were likely to purchase their first or an additional credit card during the year 3 from these the current market for e-credit cards was estimated as 1.1 million. To this was added an allowance for predicted annual growth in internet access of 35%, 1 million adults already buying on-line, and 2.5 million considering doing so, to produce a top-end estimate for e-credit card demand in mid-2000 of almost 5 million.<br><br> Many of the principles underlying the launch of an e- product such as Egg Card are similar to those for a traditional product. Opportunities and threats must be assessed 3 in this Egg were comforted by the continued growth of their savings account after access was switched from the telephone to the internet. So too must cost dynamics 3 an example from Mark 9s own area of responsibility of how different these are for on-line products is that at the beginning of 1999 postal customer satisfaction surveys for the then telephone-based savings account were costing £20,000 per month, whilst on-line surveys now cost just £2,500 per month representing mainly the development of the website to accommodate these as data are returned at the customer 9s expense.<br><br> And marketing must be undertaken 3 this was partly via click- throughs from banners on other sites to help distinguish Egg Card from traditional credit cards. However, in the same way that the rapidly changing internet market-place requires continuous data collection and product improvement in the light of this, so it also drives a shortening of the time in which e-products are brought to market initially. Indeed, Egg became involved in a race to launch the first e-credit card, and winning this race has undoubtedly helped it to remain ahead of the competition.<br><br>

less