Digital Intelligence 1997

back to e-Publications

INTRODUCTION

The conference was organised by Infonortics Ltd. of Tetbury, England. It took place on November 17-19, in The Hague, The Netherlands. There were 62 participants. Slides supplied by some of the speakers are on a page on the Infonortics Web site. Many of the presentations have appeared, or will appear, in the Journal of AGSI.

Pre-Conference Workshops

The first workshop entitled "Integrating traditional and software techniques for competitive intelligence" was led by Stephen E. Arnold of AIT, USA, and Marydee Ojala, editor of DATABASE magazine. The second workshop, which ran concurrently, "The people side of technology adoption" was led by Neil Simon of Business Development Group, USA.

Integrating traditional and software techniques for competitive intelligence

Stephen E. Arnold, AIT, Kentucky, USA and Marydee Ojala, Editor, DATABASE magazine, USA

We were given reprints of two recent articles:
Arnold, S. E.; Arnold, E. S. Vectors of Change. ONLINE July/August 1997, 19-32.
Arnold, S. E.; Arnold, E. S. Push Technology. DATABASE August/September, 1997, 36-46.

Steve Arnold spoke first. He said that push technologies are "footprintless"; they differ from earlier ones that left a paper trail. Only 5% of information is in text form, so Steve welcomes the new class of companies offering spiders for sound, image, and animation. Visualisers are needed to help you cope with all the information at your desktop. DR-LINK, SemioMap, and products from i2 were presented later in the conference.

Marydee Ojala said that Northern Light allows you to search mixed sources, for example, news databases and Web sites. Marydee's talk concerned putting the content into context. She discussed three examples, an Australian view of the automobile industry, the competitive advantage gained by a company that recognised well in advance that it would be affected by a gelatine issue arising from the BSE crisis and a world-wide competitive intelligence effort in a chemical company. In that last company, the marketing staff and the Library and Information services staff operated in different cultures. They had to get their heads together to decide which information was important.

The workshop ended with open discussion on sourcing and covering up your footprints. Marydee stressed the importance of filtering what you got from the filters, and following up leads. Steve summed up. If you take six months out from your job, when you return it will be as if you are in a foreign country, because things are changing so quickly. We have moved from pull, to push, to punch technologies. "Punch" has agent-monitoring. Multiprocessing PCs are coming with 6 Gigabytes of RAM. You will be watching TV on your PC. Infonortics can use Framemaker to publish on multiple platforms but there will be live, smart documents with Active Desktop etc. in future. He mentioned DynoBook.

He recommended that we visit Tibco. Tibco is a tool underneath the spider. It is a "killer" tool. Steve thinks that Microsoft's DCOM for objects will dominate rather than CORBA. Newer tools are embedded in the network not in the computer. If you have content, this is a great time to be alive.

Steve was also the first speaker in the conference proper, to which we now turn.

Challenges and opportunities in the world of data

Stephen E. Arnold, AIT, Kentucky, USA

Steve talked about enabling technologies that are seed technologies and are network-based. Network worlds require common standards. The first technology he covered was "spiders". Spiders can now update you hourly and they can query not just text but also spreadsheets etc. The network is the home of the spider. A second technology is visualisation. There are huge volumes of data now available but a picture can transmit hundreds of megabytes of information in seconds. A third technology is "punch delivery", exemplified by Grand Central Station (GCS) from IBM, an intelligent Java-based search engine. Tibco (The Information Bus Company) is now owned by Reuters. Microsoft has future plans based on this (Yukon plus Active Desktop). Steve suggests visiting datachannel.

Applications have been described as digital "feet on the street" or "ears listening in on a conversation". This technique can be used for monitoring in real time. Machine translation has also been combined with punch delivery and you can "read it all" by looking at pictures. Steve gave examples such as looking at the location of cell phones on a map (your boss can actually keep track of where you are) and the automated formatting of reports.

The hard reality of soft information: new techniques for developing business foresight.

Seena Sharp, Sharp Information Research, California, USA

Seena talked about foresight. She took Ford and Chrysler as an example. Chrysler developed the mini-van and is still ahead after 14 years. Although someone at Ford saw the concept first, Ford failed to develop it. Before there are facts, there are trends, and before there are trends, there are changes. Hard information involves facts. Hard information is important but you cannot measure early changes. This is where soft information comes in. Hard information is retrospective; soft information is indicative. Peter Drukker said "While the future can't be planned, events can be foreseen".

Soft information includes opinions, rumours, commentaries, conjecture and musings, complaints, anecdotes and observations. Soft information is not the same as a "gut feeling". There are techniques to identify changes. One is to value non-traditional sources. Read cartoon strips and listen to talk shows. Look at what other fringe industries are saying about your own industry. This is competitive not competitor intelligence. When you buy a house you don't just look at its square footage and its number of bedrooms, you look also at the environment e.g., local schools, transport, and politics in the district.

Another technique is to be alert to your disturbed gut. If the information challenges your assumptions, pay attention and ask what it means. Capitalise on alternative uses, misuses and abuses. Share, share and share (this is knowledge management). Beg borrow or steal: look at the number of companies which have borrowed American Airlines' frequent flyer programme. Soft information is your early warning system.

The importance of human intelligence sources

Douglas Bernhardt, Business Research Group, Geneva, Switzerland

Douglas introduced the concept of actionable intelligence which supports future-oriented decision making. It is important to check on alliances (who is getting into bed with whom) and also to check that your own partner is staying in bed with you. The intelligence cycle consists of processing, analysis and production, dissemination, planning and direction, and collection which links to processing.

Human intelligence (HUMINT) is that which is not published. OSINT is open source intelligence. SIGINT concerns signals, IMINT, imagery, and MASINT messages and signatures. Douglas' talk concerned HUMINT: gathering information from human sources. The KGB's approach to using humans used to take advantage of MICE: money, ideology, compromise and ego. The first three have no place in a CI environment but ego does.

Collection resources include budget, internal analysis and collectors, time, external consulting firms and sources (open, internal, or external HUMINT). Collection management involves understanding the requirements. HUMINT is used for verification of facts. It provides evidence of a rival's motives and intentions. It reveals information hidden from other forms of surveillance and environmental scanning.

Elucidation is rarely adversarial. Its authority derives from the interviewer's flexibility, imagination and people skills. Proven elucidation techniques are bracketing (e.g., age 25-34, 35-44, 45-54), challenging, criticism, neutral phrases (non-threatening statements such as "that sounds fascinating"), restatement and suggestions ("you must have several interesting R&D projects underway in multimedia technologies").

Automatic analysis and presentation of soft information from customer databases

Charles Huot, IBM, Paris, France

Customer calls analysis allows you to increase your knowledge about customers, better understand customer complaints and comments, and learn about customers' perceptions of your competitors. Voice messages can be converted to text using speech recognition. The first step is compression and abstraction. The call transcriptions are processed by IBM text analysis, which gives text indexes, which are submitted to IBM visualisation and IBM clustering. The 80:20 rule can be applied to information overload of spoken and written information: unstructured textual information (80%) and structured numeric data (20%).

Text mining is a new component of the Business Intelligence Solution Family. Pfizer will use GBIS text mining to analyse biomedical documents extracted from various sources such as MEDLINE. Another example is the product FOX. Charles showed a pie chart of customer comments separated into compliments, compliments plus, and compliments plus plus, and complaints in three similar categories. Evolution of the classes by date can also be plotted. Charles gave a demonstration of customer relationship intelligence from customer emails and phone calls and from representatives' comments. He also demonstrated an example of IBM and Hitachi patent portfolios. He clustered patents, coloured the map and displayed actual patent data.

SemioMap, and the concept of semiotics

Claude Vogel, SemioMap Corporation, California, USA

The SemioMap software can be downloaded from the Internet. It requires a Solaris or Windows NT operating system on the server but there are no restrictions on the client since Java is used. SEMIotic analysis of LEXical networks has facilities for text, lexicon, co-occurrencies, and clusters. At the end of 1998, semiotic patterns and narrative building will be available. A SemioMap is a map of inter-related concepts with proximity links. Claude gave an example of data mining 2 million pages on the Web. He showed an island called "data mining" with satellites of related text around it, and other ellipses named "business problems", "publicly available software" and so on. Thus a conceptual map of documents is viewed. Claude demonstrated SemioMap Server 1.3 on 2053 kbytes of data. The Collector program took 14 seconds, Extractor, 9 seconds, Compressor 8 seconds, and Index Builder 12 seconds. He demonstrated a Java applet for the recent explosion on a TWA aeroplane. He clicked on it to explode (no pun intended) to the clusters "mechanical failure", "criminal act" etc. Clicking on criminal act exploded to "national transportation safety board" and so on.

The promise of real intelligence in the network

Michael Weiner, Manning & Napier Information Services (MNIS), New York, USA

Michael was not able to attend and his presentation was given by Mary McKenna. The paper has recently appeared in the Journal of AGSI. Manning and Napier is currently showing two products, MAPIT (Management and Analysis of Patent Information Text) and the Natural Language Processing system DR-LINK (Document Retrieval using Linguistic Knowledge). DR-LINK is linguistically based not just word-based.

Mary talked more about new systems under development at MNIS. CHESS (Chronological Information Extraction Software) uses several different information retrieval engines and decides which does the best job. This project started in October and goes on for 18 months. Another project is summarisation. The software generates summaries based on subject field codes, complex nominals, proper noun categories, most relevant section, and text structure.

Conceptual Interlingua DOcument Retrieval (CINDOR) concerns cross language versus multi-language. The query is translated to concepts. Documents are indexed by concepts. Matching occurs at the conceptual level. This gets around the machine translation problem.

KNOW-IT (KNOWledge-based Information Tools) is a small part of an integrated group effort to construct a high performance knowledge base for the US government. It is an extension of CHESS. MNIS' newest US government grant is for EVA - geospatial information storage and retrieval using learning algorithms to support retrieval.

One picture is worth a thousand words: software for data visualisation

Mike Hunter, i2, Cambridge, England

The Interpol Crime Analysis Working Group in Lyon uses techniques used by the military. The process involves task definition, followed by collection, followed by evaluation, then visualisation/analysis and finally presentation. After visualisation, re-examination may be carried out, leading to another cycle starting with collection. Spiders are completely redefining collection. Visualisation is not just about pictures: it needs an interactive interface and only the right picture gets the information across.

Edward Tufte has written about the necessary features of visual excellence. The company i2 is the market leader in a law enforcement and defence niche and is in business intelligence.

The information decays so it must be gathered quickly for the investigative data sets. These data sets are large, complex and detailed and may be in any source or format. Some of the information is computerised, some not. You have to beware of counter business intelligence. The problem is to decide what all this information is telling you. Techniques are link analysis and case analysis along timelines to identify the sequence of events. Various styles of chart are used. This is an understood and documented discipline. Mike gave examples of manual and automatic link and case analysis and of temporal charts and time lines. Related applications are studying complex money laundering and build-up of shares.

IConnect software connects to databases and does visualisation and analysis. IBase is an analytical database generator. IGlass is a charting tool for finding patterns in your data. Mike gave a long list of i2 customers including Interpol and police forces, and Customs and Excise in many countries.

Mike demonstrated some software in alpha-test. He showed a manual chart of the sequence of events in the assassination of President Kennedy as an example, together with photographs and a video clip. The Visual Basic algorithm is extensible because of OCX implementation. It can be dropped into your own application. Mike demonstrated mining from an Access database of the Barlow Clowes case in the UK. Peter Clowes and his links were taken out and links were extracted from another database before chart generation. Mike also demonstrated an FBI intranet application using DCOM and Internet Explorer. Understanding through visualisation and analysis is what it is all about.

Connecting intelligence to strategy: software-enabled business analysis

Larry Moffett, Management Software Association, Brussels, Belgium

Larry works for Decision Resources but was representing MBAWare from the Management Software Association. MBAWare has built-in management expertise and is designed for use by managers and decision makers. It applies complex planning and analytical techniques to specific business tasks. It offers advice and guidance to non-expert users and runs on the desktop. Management applications are analysis, modelling, planning, problem solving, reporting and strategy. Larry gave four examples: risk analysis, statistical forecasting, SWOT analysis and financial analysis.

In the program @RISK you can use a Monte Carlo simulation and show the probability distribution of all outcomes by testing multiple combinations of input variables in a model. Larry showed a product launch in Microsoft Excel. Forecast Pro was used to analyse a time series of historical data, select a statistical model, fit the parameters of the model to the historical data and generate a forecast by extrapolation. The Five Forces model considers customers, competitors, substitutes, suppliers and another factor. OT is used for industry structure analysis. SW capabilities analysis uses the Michael Porter value chain model. Larry gave a demonstration of Quick Insight. Finally, he showed la Solution Douce (available in various languages) to analyse balance sheets and profit and loss accounts of competitors, customers, suppliers and partners.

Threat assessment databases

Michael Brenton, Memex Technology, East Kilbride, Scotland

Michael was with Memex when he submitted the paper but he is now back with Sterling Software, Information Technology Division. His mission was to provide an automated environment for threat warning and analysis, studying people, groups and countries who pose a threat to US interests. He also had to cope with cultural change.

The database engine is the Memex information engine. The information is tokenised and a vocabulary is created. Zoning concepts lead to field definitions. The user can navigate, optimise the search, and exploit the vocabulary. Statistical profiles and feedback mechanisms are offered. The latter give statistically meaningful terms on which you can rank. The analyst needs such tools and the information needs to be tailored to the analyst.

The user interface uses Netscape and Java applets, which can be used all round the world. APIs allow linking into other Department of Defense systems. The system will go live in April 1998. The forces in Bosnia have used the fuzzy search options and found them useful, for example, for Bosnian names. In the near future, link analysis, temporal analysis and spatial analysis tools will be supplied. In the medium term, natural language understanding and multi-tiered information mining servers will come. In the long term, datascape visualisation and VRML will be applied.

Obstacles to threat analysis include rdbms thinking ("normalise or die"), soft data, changing information requirements, and lack of information repositories (which are needed because you don't know who will be the enemy tomorrow). "Best of breed" is being integrated for use today but they are on the lookout for more ideas and tools.

Lessons in human information processing: how to win clients and motivate decision makers

Neil Simon, Business Development Group, Michigan, USA

This was an extended presentation in the form of an interactive workshop. Its aim was to show us how to understand a customer's unique style of perceiving and expressing information, to be more effective in presenting information to that customer and to build a persuasive business case tailored to him or her. Neil starting by bringing Harry Collier to the front, seating him on a throne and dressing him as a king, labelled "customer". (Click here to view the photograph).. Neil then discussed the basic communication cycle and broke down the tasks involved. The receiver receives a message, then grasps it, then responds. Receiving involves perceiving and processing. Grasping means interpreting and reacting. The receiver has to structure a response then send it.

Neil described the common genetic factors that affect perception. This caused some complaints from one person in the audience who believed in "nurture" rather than "nature" but I myself found Neil's arguments persuasive and his techniques certainly worked in practice later in the workshop. Another attendee wondered whether tailored persuasion is "ethical" but I think he was playing devil's advocate and did not have a real objection. The lively discussion showed that Neil knows how to keep an audience's attention even at 6 p.m. after a long day of presentations.

Neil explained the common factors that affect human information processing, interpreting, reacting, structuring and sending. He harnessed the techniques of neurolinguistic programming. Then he showed us how to do our homework, analyse our goals and study our target customer. Having done this, we could develop a personalised approach. Now we had a chance to practise implementing an approach.

We were divided into four teams, each of which had to design a product that would appeal to Harry Collier. Each team had to find a spokesman to sell the product to Harry for two minutes. We were armed with the knowledge that Harry is a francophile publisher who organises conferences, plays the viola and likes listening to violin music. Some of the attendees who know Harry well may have been at an advantage but this seemed not to bias the competition since one committee member failed to sell his product effectively and another person unwisely gave Harry the hard sell.

The winning team persuaded Harry to spend a free weekend at a château in France to sample the food and to study the château's advantages as a conference location. The pièce de résistance, however, was the "selling" technique of Charles Huot of IBM in Paris. Charles' quiet, conversational approach and slightly French accent clenched the decision.

How to prepare for access to the 57 varieties of corporate data

Jay Ven Eman, Access Innovations, New Mexico, USA

Are you sitting on acres of diamonds? You need to assess the value of your collection and consider the costs of converting it. Jay's presentation covered legacy data that needs converting, why it might need converting, the conversion process, the target environment, and budget and cost parameters. Legacy data possesses latent value and it can be converted into strategic intelligence.

Jay discussed subjective and objective measures for determining the value of your collection. The nature of it has to be studied. Does it include slides, human resource records, archival matter, numeric data. geographic data, or bibliographic records? Is it smelly? Will it scan? Are there paper clips or staples in it? How big is the collection, who will use it, and how will the information be retrieved? There may be added value in having the information available or there may be no cost benefit (for example if the collection is too small, if there is only one user, or if an electronic product will cannibalise print revenues).

The second half of the problem concerns legacy systems. Jay discussed various possible technologies including document management systems, groupware, HTML editors, and Internet browsers. He explained how to manage a conversion project and justify it as an investment, not an expense. He went into detail on costs of rekeying. scanning and so on, and emphasised the factors that add to the cost of conversion. The greater the accuracy needed, the greater will be the cost of the project.

Aqua Browsing: visualisation of large information spaces in context

Anne Veling, Medialab, Schellinkhout, The Netherlands

The world is changing very fast. There is much more information available, more interconnection and more live information. The life span of information is decreasing. Anne discussed new information needs, concept spaces, trees and hierarchies, human and automatic structuring, dynamic data and dynamic structures. It is hard to handle multiple views using trees. Semantic networks allow us to do everything we could do before as well as allowing us to do more.

The Aqua browser is a dynamic solution: a Java applet can be embedded anywhere which handles fuzzy information and visualises dynamic concept space. Anne gave a demonstration and explained the interactions between the Aqua client and the Aqua server. In order to make this report readable by all browsers, I am avoiding reproducing the mathematics involved in the relevance state and semantic networks. The client application needs these applications, it performs visualisation and it handles user actions. The server application performs look-up in a semantic network, a relational database and a text database. It handles expanding and user profiling.

Anne gave a demonstration with Europe coloured blue and countries around Europe in other colours. If France is chosen, it moves to the centre and it becomes the blue zone. The computer has recomputed the concept space. The Aqua Browser is a visualisation tool that can be integrated in search engines such as AltaVista. Users like animation and in-context browsing and this is an attractive way of dealing with dynamic concept spaces. There are technology drawbacks: the system uses Java, it is slow, it is big, and there may be concept overlap and absent relations. A Delphi version (Windows 95) is planned.

Anne closed with some thought-provoking theses. New users will demand far better answers from computers than we can give them with today's techniques. Trees are out; nets are in. Relational databases are archaic: they are good for aeroplane reservations but not much else.

Communication of strategic intelligence in an R&D environment

Olga Svoboda, MINTEK, South Africa

Olga described her presentation as a simple paper covering practical experience. R&D companies are competing in a global market and have a need for global strategic knowledge. Systems cannot be closed: they have to be integrated with others. Companies have to learn. Olga described a learning model. She discussed the principles of the successful knowledge base. The material must be focused towards the organisation's objectives and regularly updated. Accessed must be personalised. Olga talked about networking and complained that IT departments think that connecting to the Internet is marvellous.

Her company does use the Mintek Web site for business information but not for R&D information. Lines from South Africa are costly and GUI's don't work. Hard copy is obtained in traditional ways (e.g., from the British Library). Olga described a minerals processing knowledge base (the corporate memory) from open and confidential sources. A combination of technical information with business information is needed in order to plan the research of the future.

Competitive intelligence using electronic data sources

Gail Hart, Swiss Re, Melbourne, Australia

Gail's company is in re-insurance. She talked about building the Competitive Advantage (CA) system, presenting a diagram of a pyramid with the IT infrastructure at the base and layers of decreasing size built on it, namely, data sourcing and extraction, information processing, knowledge management, learning organisation and CA. Five IT staff manned the project. Lotus Notes was used in the interface.

CI data is primary and secondary, from people and from reference sources. Sources have to be questioned and examined for errors (especially in the case of the Internet). The Internet is becoming increasingly important to the company. Both hard and soft information are stored. Reference tools include directories, databases and virtual libraries. Primary references are annual reports, government reports, speeches, and financial reports. Secondary references include newspapers, industry reports, newsletters, journals and trade associations. Soft information comes from surveys, brokers, experts, observations and presentations.

The end-users wanted to capture the information themselves. Gail listed various news and Web sources and statistics, tables and charts. Favourite Web tools included LookSmart, Web Wombat, Aus Index, Dogpile, Yahoo in Asia, InfoSeekUltra, and HotBot. There has been a surprising increase in the number of Web pages over just one year. Gail listed CI profiles in knowledge bases under the headings competitors, industries and clients.

Challenges have been the rapid pace of business, information overload, competition, economic changes, breaking preconceived notions, improving management decision making, improving benchmarking success and refreshing CI.

Answering questions of identity and activity on the Internet

Philip Wherry, Concept Five Technologies, Virginia, USA

Philip's company specialises in computer security. He talked about masking your identity and understanding and managing the footprints that you leave behind when you use the Internet. He gave reasons why you might want to look at the traces left by people who visit your own site. The easiest way of checking on a visitor is to give him a registration form, generate a password and send it to the registered email address.

Sites have a Domain Name and an IP address. DNS converts a typical URL to an IP address number. You can do this in reverse: a process known as reverse DNS. Falsification of reverse DNS is possible. Double reverse DNS can be too slow to do in real time. You can mask your identity using DNS. You make up a dummy name and register this with InterNIC, then set up both forward and reverse DNS. A low cost alternative is to obtain a normal user account from an ISP but bandwidth is usually limited.

A proxy server can be used to mediate access to the Internet, for security purposes. You can access InterNIC to determine the identity of someone if DNS information is missing. You can also do network route tracing. Identity knowledge can be used in real time for selective presentation of information to people visiting your site.

Because of the state management problem you cannot tell if someone places an order after doing a certain type of search on your site. A cookie can be used for co-operative and semi-cooperative tracking. It can be used to find individual users behind a proxy firewall and can be used to see if the person who logged in today is the same person who carried out a certain action last Friday.

Philip talked about the referring URL and making use of available coverage mapping tools. "Honeypot" sites attract the attention of search engines. Domain hijacking controls are primitive; secure sockets layer (SSL, https) is better. SSL can also be used to identify users.

Using war games to leverage your data

Mark Chussil, Advanced Competitive Strategies, Oregon, USA

Why do some strategies fail? Why is Lotus no longer the leader in spreadsheets? Forecasting works for some problems (e.g., the growth in electronic spreadsheets) but it does not work so well at the product level. More precise data does not necessarily mean a better decision and it can cost a lot of money to get the data correct to that last decimal point. Using a "tried-and-true strategy" does not always work. What if the future is not like the past or what if you want to mould the future? Measuring results and holding managers accountable can help the competition to win on innovation.

Bobby Fischer won at chess because he could see 11 moves ahead. Deep Blue can see many more moves ahead. Kasparov lost to Deep Blue for this reason and because Deep Blue frightened him by making a move that no human would have made. What if Kasparov could team up with Deep Blue? War games simulate your future without risk. They "leverage" [not my word] intelligence. People create strategies and scenarios; computers handle complex calculations. War games thus improve strategic decisions.

Determined competitors add realism and punish complacence. Rolling back the clock boosts learning and aids contingency planning. Speed and the risk-free environment free managers to experiment. The shared experience builds consensus. There are various different types of war game available. Mark listed some styles and their applicability to various problems. The point is not to predict the future but to make better decisions about the future. Mark recommended "The Art of the Long View" by Peter Schwartz.

Marketing information and strategy formulation: a case study of retail banks in the United Kingdom

Judith Broady and Timothy Hayward, University of Wales, Aberystwyth, Wales

The speakers presented interim results of an 18-month research project that started in March. Judith spoke first. They chose banks as a subject because of recent deregulation of the banking sector, new forms of competition and the significant changes in the services offered by banks in the 1990s. Forces contributing to rapid change in the retail banking sector were consolidation, convergence and electronic delivery. In a mini-survey of 13 banks done in 1996, the speakers received an 85% response to their questionnaire!

In the current study many more banks are involved. A 70% response has been obtained in a questionnaire looking at involvement with group or business strategy, use of market planning, role of the information department, types of analyses used, importance of information, bottom-up input to strategy formulation, and models. Follow-up interviews are now being carried out.

An appropriate perspective is marketing. Marketing staff need to stay close to the customer base. Customer satisfaction is studied. Differentiation is one of three strategies. Most people consider that marketing is tactical rather than strategic but the speakers disagree. Drukker says that to develop strategy we need organised information.

It is not necessarily better to have even more data. Models which aid strategic decision making include Kaplan and Norton's balanced scorecard, the business excellence model and value-based management. Marketing involves long term relationships so the speakers want to study its link to strategic planning. Timothy presented some interim results including a histogram of the models used in the strategic process.

Global Gold: planning for profits in export markets

Ruth Stanat, SIS International Research, Indiana, USA

Ruth supplied a large document headed "China Sources" listing Chinese addresses. Reduction in tariff barriers and other factors have increased trade across the oceans. You need to do your homework to reduce risk in trading in foreign countries. Ruth gave an example of a case study in China and Latin America and the production of a strategic opportunity grid.

Asia will be the "engine" of the next millennium. There are strong demands for all sorts of products, and low labour costs in many parts of the Asia Pacific region, but there is no pan-Asian culture. It may be very important to have connections to government. Ruth is editing a book, which has 30 contributors and she gave some examples from it. One concerned a car manufacturer which donated 300 motors to a large iron and steel group, got a local agent and became quite successful. A company with digital compact cassettes was less successful. They did not do their research well enough: the Chinese were happy with the cassettes they already had. The company unwisely sacked a Chinese employee who had government contacts. Another company lost customers because it was accused of spying on the government.

The legal system is not as developed in China as in the West. There are patent and trademark issues. A Spanish company lost a case concerning a plagiarised drink in a Chinese court. A Swiss watch company failed to realise that under Chinese law you must get a Chinese patent before entering the Chinese market. In the meantime a Chinese company came up with a competitive watch.

If you enter the Latin American market you must stay a long time. Re-entry is next to impossible. There are opportunities in telecommunications, energy and utilities. Wal-Mart was unsuccessful in Brazil because it had a problem with suppliers. Ruth described a study on the potential demand for a very sophisticated sort of hospital bed. The Indiana company making these beds did primary and secondary research in 11 countries and Ruth prioritised the markets for them. She did primary and secondary segmentation of the market and did an analysis of high end products based on various factors. Countries were benchmarked and scored against the best country. A weighting scheme was devised. Brazil came out as high priority, followed by Mexico. China was number six on current rating and number four in the long term. Private hospitals are a better market than public ones and urban areas are a better prospect than rural ones.