How to organize a business intelligence service.

How to organize a business intelligence service.

Boris Liflyandchik, Nikolay Likhodeyev
candidates of physical and mathematical sciences

We appeal first of all to those entrepreneurs and managers who, having realized the importance of business intelligence (see our previous article in «MB» N 6, 98), want to create a corresponding service in their company. We hope that employees of companies where the BR service has already been created, after reading this article, may also have some new ideas.

To avoid subsequent disappointment (which most managers of those companies where the BR service already exists experience), before making a final decision, once again make sure that all the conditions listed below are met:

Condition 1. You know what information you need to make specific decisions that are important for the company.

Condition 2. You believe that the information you need can be obtained, and you have a general idea of ​​how.

Condition 3.You estimate that the costs of collecting this information will be many times less than the benefit you expect to receive from using it.

Condition 4. You can find resources to cover the necessary costs.

Condition 5. You do not see other ways to use these resources that could bring greater benefit to your company.

Condition 6. You are ready to receive information that destroys your previous ideas and expectations and/or proves the erroneousness of decisions you made earlier.

As practice shows, it is most difficult to establish the fulfillment of the last two conditions. Moreover, if the problems associated with condition 6 are predominantly subjective, the difficulty of checking condition 5 is that the effectiveness of almost any analytical activity can be quantitatively assessed only in individual cases. For example, it is possible to estimate the effect of information about the non-competitiveness of a new project, allowing to avoid large losses at the stages of experimental testing, production of a pilot batch. But how to calculate the economic effect of information confirming the correctness of a decision that the manager already considered preferable?

Stage 2. Selection of a BR service manager

This stage is the most important. The difficulty is that the requirements for the head of the BR service are somewhat different from the standard requirements for middle managers. In addition to the obvious suitability for analytical work, he must have increased resistance to criticism and other forms of pressure and not be a careerist, even in the best sense of the word. Failure to meet the last two requirements can lead (and often leads) to a conscious or subconscious distortion of the information provided in favor of the opinion of superiors or other employees of the company who are capable of exerting pressure or, conversely, facilitating career advancement.

Very often this post is given to a manager who has worked at the company for a long time and has become unpromising, but who has a good relationship with his superiors, but who has not been very successful. However, as a rule, it is precisely for such a manager that friendship with the management will be more valuable than truth. The authors believe that, after professional intelligence officers or specialists with similar work experience, former scientists who have gone into business should be considered as candidates for the position of head of the BR service. Such people are more inclined to analytical work, more receptive to modeling situations and are often more resistant to failure.

As was already shown in the first article, the most correct option is — subordinate the BR service manager directly to the first or, in extreme cases, second person in the company.

All the listed parameters of the BR service should be mutually linked, so that the process of their development should be prompt. The natural desire for the service to quickly reach the project structure, i.e. to solve a wide range of problems, seems erroneous. The range of BR tasks should be expanded gradually, as evidence of its necessity accumulates. However, the initial budget should be obviously sufficient for solving several of the most urgent problems, which will bring practical evidence of the real usefulness of the service. It is the specific contribution of the BR to the adoption of specific tasks, decisions important for the company — the best way to win the trust of the MWU in the BR, especially when its recommendations diverge from the opinion of managers of other departments.

The optimal structure of the service is naturally determined based on the operating conditions of the company, its strategy, size, structure, etc. Figure 1 shows a typical diagram of this structure for a medium or not very large company.

The company has only one specialized analytical unit — the BR service, directly subordinate to the head of the company, and various functional units, two of which are shown in Fig. 1. The BR service performs its main functions of collecting initial data, systematizing them, storing and analytically processing them to fulfill its main task — analytical support for decisions made by top-level managers (TLM). Some of the initial data comes to the BR service from other units. The BR service not only uses this data in its work, but also supplies them to those units for which they can be useful. Since these data are also entered into the database formed by the BR service, they can be sent to the units exactly when they need them. The units can use the information received from the BR service for their own analytical work (for example, for market analysis in the marketing unit or for evaluating various design options for a new product in the R & D service). In addition, some of the analytical information synthesized by the BR service is also sent to the relevant departments (for example, a review of competitors' strategy to the marketing department).

Fig. 1. Option for the structure of the BR service.

Because the BR service is the central link in the information structure of the company, it has at its disposal information that is so useful to other services that they are prepared to make certain sacrifices to obtain it. It is this interest in cooperation, and not any formal directives, that should be used to solve the problem of extracting large volumes of data noted in the first article. This same interest helps to achieve coordination with departments in carrying out a number of data collection and analysis operations (for example, BR can cooperate with the R & D department in preparing a forecast of the technical policy of competitors).

Based on the specific features of the company's activities and information resources, some of the functions of data collection (e.g. via the Internet), as well as the functions of maintaining the database and redistributing it between the company's divisions, can be assigned to other services, such as the information technology service. Of course, this requires very good coordination of the work of these two services.

Fig. 2. The structure of the BR services in a very large company.

In very large firms and especially in firms with a high degree of diversification and/or territorial disunity, it is advisable to use a structure such as that shown in Fig. 2, which is built by analogy with military intelligence. Each large structural unit of the firm has a BR service that performs the same functions as the central BR service, but in the interests of the MWU of this structural unit. At the same time, these services exchange analytical information with the BR services of higher and/or lower structural units and coordinate their actions with them. Similar to the previous case, for effective cooperation to be realized, the information exchange between services of different levels must be mutually beneficial. Although Fig. 2 shows a two-level BR structure for simplicity, a greater number of levels is also possible, for example, the central service — the branch service — the business unit service. According to ASBIA, a similar principle of organizing the BR has been implemented by the well-known corporation «Motorola».

Also not shown in Fig. 1 and 2 are the «external agents» that each BR service may have. For example, a company that produces complex analytical equipment may attract workers who service the equipment of a competitor company at the enterprise that purchased this product.

External agents can also work as consultants, independently deciding which of the information that has become available to them may be of interest to the company that has engaged them for cooperation. The use of external agents can be especially useful at the stage of formation of the BR service. Obtaining valuable information with the help of such agents is another effective way to gain the trust of the IWU.

An alternative option for quickly gaining authority is to use the services of a company specializing, for example, in information search, assessment of the competitive situation in the relevant industry or investment conditions in a particular region, etc.

Stage 5. Development of the BR Service Work Program

Due to the limited scope of the article, it will consider only one of the most important aspects of the program — the selection of open information sources. Specific tasks and various forms of processing data obtained from these sources will be briefly analyzed in the next, third article.

Due to the variety of analytical tasks solved by BR, documents of almost any type can be of informational value to it — from books to labels, inscriptions and pictures on packaging. In relation to typical situations, the most useful can be corporate reports and reference books, court cases, business and industry press, advertising brochures, patent documentation, scientific periodicals and conference materials. When solving a complex analytical problem, for example, in comprehensive forecasting of the development of a dynamic high-tech industry, the number of documents that can contain useful information can be hundreds, thousands and even millions of units, which makes the task of collecting all relevant (i.e. related to the topic under consideration) documents not only hopeless, but also meaningless. This is precisely why the problem of selecting sources is important.

The criteria for such selection are the information value of documents and their accessibility. The information value of a document, from the point of view of the BR, is naturally determined by the influence that the information contained in it can have on the adoption of a particular decision at the level of the MWUU.

The second aspect of information value is the novelty of the content. The introduction of this subcriterion is due to the fact that the same information can be contained in dozens and even hundreds of different documents (B. I. Liflyandchik. «Scientific and Technical Information», ser. 1, 1982, no. 6, pp. 26 — 30), and the fact that all of them will be found will increase the amount of useful information only slightly. The availability of a document is measured by the total costs of extracting useful information from the document (i.e., searching for, receiving, reading, understanding, and converting it to its final form).

Although the availability of traditional paper documents fluctuates widely, it is obvious that today they are generally inferior in this criterion to documents in electronic form, the search, and to a large extent the analysis and transformation of which can be carried out with the help of computers, i.e. much faster and more accurately. It is also obvious that the trend towards outstripping growth and increasing gains in the availability of electronic documents in the coming years will become overwhelming. Therefore, a modern BR service should be oriented towards the priority use of documents of this type.

Among all possible electronic information sources, electronic databases (DB) are undoubtedly of the greatest value for BR, as they provide the ability to quickly access large arrays of documents from various sources. Although a comparative analysis of various databases deserves consideration in a separate article, it is advisable to give some assessments in the context of considering the organizational aspects of BR.

Many useful data for DB purposes can be supplied on CD-ROM, for example, the well-known CAS, PASCAL databases, databases oriented specifically to the needs of information technology and communications departments, the RDSL database issued by MineSoft (Great Britain), on diskettes and magnetic tapes. However, such DBs are difficult to replenish, and therefore they quickly become outdated. Online databases are much more convenient, search and reading of data from which can be done in real time. The total number of such databases has already exceeded 10 thousand. The largest of them, LexisT-NexisT, contains more than one and a half billion documents (N. P. Likhodedov and L. E. Tolstykh. World information resources for businessmen and specialists. St. Petersburg, 1997). Information systems based on online databases can be roughly divided into three large groups.

Group 1. Information systems implemented as an Internet server. The development of such systems is due to the technology of creating WWW sites, their number is rapidly increasing: every week another 2000 commercial sites are added to the existing 400,000. The popularity of the Internet forces us to enter into it the most diverse sources, for example, patent document databases. Software for searching in such systems is also rapidly developing.

When the site profile exactly matches the subject of the information task, systems of this type can be very useful, especially in case of low registration cost (usually several tens of dollars). However, there are also very expensive systems. Thus, the price of access to data of the information service Business Intelligence and Data Warehouse http://data.psgroup.com, specially oriented to the needs of BR, is $ 1795.

The breadth of the subject and the number of documents in the systems of the first group varies within very wide limits, and in most cases, to solve BR problems, it is necessary to access several of them. This increases the cost of searching for the necessary sites and extends the overall search time.

Group 2. Online systems for the mass consumer (consumer online market).

The systems of this group contain a large volume of materials on completely different subjects: from weather forecasts in different states and price lists for certain services to full texts of some newspapers. The most significant representative of this group is the company America Online, which recently acquired the well-known information system CompuServe in Russia. America Online provides access to more than 50 newspapers and magazines, its subject content: computers and software, reference literature, news and finance, entertainment industry, tourism, scientific literature, sports, etc. Systems of this type are on average the cheapest (access fee for America Online is from $19.95 per month in the USA), but due to their non-professional orientation they do not contain many of the most valuable sources for BR and do not have developed search tools. This significantly limits their usefulness for BR purposes.

Group 3. Professional databases. These are the most powerful of modern information systems. Leading systems contain valuable information on virtually all subject areas. Their general characteristics are presented in Table 1.

Table 1.

Of particular note are data on companies and markets. LEXIS-NEXIS alone contains over 300 company databases. The largest database of this type, Dun & Bradstreet, contains information on 46 million companies from 135 countries.

The Investext database contains information on company activities, economic life reviews for various regions and 53 industries. Information is supplied by over 270 investment banks, major consulting firms and research companies. Over 500 reports are added daily.

The Predicasts Overview of Market and Technology database contains full text or abstracts of articles from over 1,500 publications, professional reviews, economic journals and newspapers, newsletters, audit reports, market and industry reviews, and press releases. There is information on all industries and services.

Market Research Reports DB — these are reports from the largest research firms, such as Business Communication, Euromonitor, Datamonitor, Freedonia, Frost & Sullivan, etc. on various areas of industry, companies, individual products. The value of such reports is that the research firms that create them use unique analysis methods, unique sources and non-traditional forms of collecting information.

High completeness and subject breadth allow solving a very wide range of information tasks by accessing only one or two systems, i.e. eliminating the need to master various search languages ​​and improving the quality of search. The LexisT-NexisT system offers several information products, such as Tracker, InfoTailorct, ReQuester, specially oriented to the needs of BR (http://lexis-nexis. com/esp).

Professional systems have a much more powerful and diverse search engine than systems of other groups (for example, the well-known Internet users systems Alta Vista, Yahoo, Web Crawler, etc.). At the same time, the search engine installed in LexisT-NexisT, when preparing a response to a standard request, looks through over 10 million pages of text in less than 10 seconds. As a result, it is possible to identify all documents on the issue of interest in a very short time, available in the system. Information in the systems of this group is updated daily or even more often, and access to them is possible 24 hours a day.

The completeness of the selection of source documents from various sources for the systems of the third group is also significantly higher than for the other systems. For example, when searching for documents published in the period from November 22 to December 22, 1997, in the system of the first group (on the Financial Times server http://ft. com) and in the electronic version of the same newspaper Financial Times in LexisT-NexisT, only 2256 and 7016 documents were found, respectively, with the keyword Russia contained, respectively, 100 and 247 documents. Such incompleteness of sources accessible through the corresponding server, which is not always known to the user, can lead to serious errors in analysis and thus to making an erroneous decision based on unreliable data.

The services of such systems are paid, and the terms of payment are flexible and varied. Thus, the hourly rate for working with professional databases can be from $75 to $260 for most files (depending on the database used), and the time for solving simple information problems is within 5 — 20 minutes. Since the range of search strategies is unusually wide, the total cost of any search depends significantly on the experience of the researcher and his information culture. When solving an important information problem by an experienced researcher, the costs of obtaining a unit of useful information from professional online databases are on average no higher, but lower than when using databases of the first or second groups.

Although the largest online systems use a very limited number of Russian sources of information, they receive data from world news agencies and information services operating in Russia. For example, in the large abstract database PASCAL, which accumulates data in the field of science, technology and medicine, abstracts of works written in Russian occupy third place (after English-language and French-language ones). Therefore, in general, in terms of the completeness of the reflection of Russian reality, the largest on

place (after English-speaking and French-speaking). Therefore, in general, in terms of the completeness of the reflection of Russian reality, the largest online systems are practically not inferior to Russian online databases, which are still far behind the world level in their development, and are also of significant interest to BR services. Brief information about some of these systems (obtained from the Internet home page) is given in Table 2.

From a very brief comparative analysis of various types of online databases, it can be concluded that each of them may be optimal for a specific database service at a specific stage of its activity. Thus, when solving relatively simple analytical problems associated exclusively with the Russian market, in conditions where the time factor is not very critical, but there are financial or language restrictions, it is advisable to conduct a search on the Internet and/or in Russian online databases.

However, the more complex and broader the analytical task, the more responsible the solution it provides, i.e. the higher the price of erroneous or incomplete knowledge, the more justified is the appeal to large professional online systems. It should also be noted that these systems have a deep archive of legal information up to the 18th century (LEXIS-NEXIS), and patent and scientific and technical information up to the beginning of the century (QUESTEL-ORBIT). Their value becomes maximum when it comes to contacts with foreign companies (remember the story of Golden Ada, mentioned in the first article), and especially when it comes to entering foreign markets (a detailed example will be given in the fourth article). Incidentally, firms like Golden Ada are by no means an exceptional phenomenon. According to Eurogate estimates made in the early 90s, among Western firms operating in the Russian market, only 15% can be considered reliable, 22% are in a state of bankruptcy and 19% are criminal. (Business Petersburg, 18.02.1997)

The enormous information wealth of professional databases allows one to obtain from them a very large number of documents that respond to a specific information request. The task of extracting all the useful information from them can create great difficulties for a not very powerful or not very experienced BR service. Much greater efficiency can be achieved if, during preliminary analytical processing of search results, one does not limit oneself to issuing the found documents, but provides specific, well-founded answers to the questions for which the search was conducted. Such a service is offered, in particular, by the St. Petersburg company ONLINE, which, acting as the official Russian representative of QUESTEL-ORBIT and LEXIST-NEXIST, has accumulated extensive experience in the field of effective work with professional databases.

Table 2.

World Security. July 1998