 |
Internet-EDI?
ICEshop! (Dutch) Knowledge4Free: gratis kennis! |
|
|
Ph.D thesis "The Viability of Multimedia Retrieval Systems for Marketing and Sales" by prof dr M. J. Hoogeveen Chapter 3 Multimedia Retrieval Systems for Marketing & Sales This chapter addresses the next subquestion of this research: what potentially viable MRSs for M&S can we distinguish on the basis of practical examples or as extrapolations of developments in M&S, multimedia systems and retrieval engines? MRS for M&S is used in the sense of an MM system or MM service with a clear retrieval component, which supports one or more M&S processes. The framework for MRSs for M&S, outlined in the previous chapter, is used to answer the subquestion and characterise MRSs for M&S, which are based on case research and state-of-the-art technological developments to be sure that these systems are potentially viable. Most of the business cases used are situated within PTT Telecom BV. The first five systems presented are based on current cases: a Marketing Communications Archive, a Tele Sales Assistant system, an MM Business Catalogue for business sales assistance, an MM Promotion System at a trade fair, and MM Assisted Instruction for self-help training on the job. The next two systems are projections based on current technological developments: a Marketing Documentation Archive, and the Virtual Market. A last system is added to put all the parts together, and present a 'master plan' for PTT Telecom in which all the systems mentioned above work together in one imaginary M&S organisation. Finally, some overall conclusions with regard to the viability of the MRSs, in so far possible, are discussed and the business objectives and perceived success/risk factors for the MRSs are summarised. Diagrams are used to illustrate the characterised MRSs for M&S; before elaborating the MRSs a short explanation of the symbols used in the diagrams is necessary. These diagrams are presented to show the M&S tasks and processes, performed by M&S personnel and M&S units, and which are supported by the MRSs. Further, the diagrams are used to show the system functions of an MRSs, the information flows processed, and the databases included. The basic symbols used are depicted in figure 11. These symbols are: • A task or process performed by an actor: in the present case this refers to M&S tasks and processes. The symbol used is identified by an actor name and a description of the task or process. • A system (sub)function performed by a (sub)system: in the present case this refers mostly to MRSs and MRS functions. A system function is identified by a (sub)system name and/or a (sub)function name. • An information flow: in the present case this refers to the flow of M&S information which often has an MM nature. Actors and system functions have in common that they process and exchange information. An information flow can occur between tasks/processes, between system functions, or between a task/process and a system function. • A database: in the present case this refers to databases containing M&S related information. A database can be considered to be a system or subsystem with one specific function: to contain information. A database is identified by a content specific name. • An external: in the present case an actor or system outside the defined M&S system that exchanges information with an actor performing a task/process or an IS within a defined M&S system.
A simplified diagram, with symbol descriptions in bold type, is given in figure 12.
3.3. The Marketing Communications Archive (MCA) Large organisations use huge amounts of reusable marketing communications material: photographs, videos, graphics, animations, slide shows, etc. This information is often very expensive to produce. The Marketing Communications Archive (MCA) is a centralised on-line archive for all marketing communications that are regularly reused by marketing organisations for advertisement, promotion and PR. A first step towards such an MCA was made by IECT (Internal and External Communication Telecom), the PR office of PTT Telecom BV, within the context of TQM (Total Quality Management). The first ideas about the archiving system were launched in June 1991, and the first version of the archiving system was realised between August and December of that same year. No centralised archive for photographic marketing material existed within PTT Telecom BV before 1992, when the IECT photo archive became fully operational. Preparations were started to make the archive on-line accessible in the middle of 1993. The following descriptions of the M&S business objectives and information processing tasks, MRS functions and MRS implementation aspects are based on interviews with the responsible manager within IECT, the IECT PR official, the archivist, the hired project manager, and on system documentation (IECT, 1993). The main business objectives (O) for the IECT photo archive are related to saving costs, improving access to corporate photographs, implementing quality control, and increasing management control. (A more detailed Cost Benefit Analysis is given in chapter 5). A decisive argument for setting up a centralised photo archive within PTT Telecom was the saving of costs by more efficient reuse of the 10,000-15,000 existing marketing photographs (O1). In the old situation it was very difficult to trace old photographic material. Photographs and slides were believed to be somewhere in a desk drawer. It was highly uncertain whether the photo material could be found or not. Working under pressure of time, it was more easy to order material for a catalogue or for an advertisement than to search for relevant old material. Another reason for not being able to reuse original material was that sometimes the old material was damaged. The production of new photographs is very expensive, as it often means calling in an advertising agency, and hiring photographer's models. Saving costs by avoiding unnecessary production of expensive new material is expected to justify investment in the photo archive (see chapter 5). Other cost savings are expected from the more efficient management and retrieval of the photographs (O2). In the old situation a lot of time was spent by the marketers searching for photographs. A third cost saving point is the prevention of copyright claims after violation of copyright (O3). In the past violation of copyright or usage right sometimes occurred as a result of ignorance due to reprinting a photograph in another type of medium than agreed upon with the photographer or photographic model. Copyright claims after a violation can cost tens of thousands of guilders. Proper administration and settlement of copyrights and usage rights by a central archive administration can prevent such costly mistakes. Another business objective of the IECT photo archive is to improve access to corporate photographic material (O4). Accessibility is important for the marketing units of PTT Telecom that prepare marketing promotions, the advertising agencies called in by the business units, public relations officials, and external contacts like educational publishers. First, good accessibility is necessary to realise the cost savings discussed above. For example, if accessibility to the photographs does not improve in comparison to the old situation then it is highly likely that better reuse of photographs will not be within reach, and publicity personnel will probably continue ordering expensive new photographic productions. Besides, IECT also expects that increased accessibility to the photographic material will result in an increase of the total number of requests and deliveries (4500 in 1993) met by the archive facilities. This will lead to an increase in total archive and delivery costs, but will lead to a decrease in costs per delivery or stored photograph. Second, it is assumed that better access to the corporate photographs will lead to a better and more balanced choice of advertisement photographs, and to the stimulation of advertisement by the marketing units, and free publicity in general. (With regard to the last point, it is reported by the IECT management that external contacts such as educational publishers are increasingly finding their way to the archive for materials). IECT management aims to improve accessibility to corporate photographic material by: • offering access to a complete and up-to-date photo database; • guaranteeing a delivery time of maximal 4 days; • a short selection time for relevant photographs (minutes rather than hours or days). It is felt that accessibility is important for the acceptance of the IECT photo archive by its intended customers. A further business objective of the IECT photo archive is to implement quality control (O5), this was not present in the old situation. Control of the quality of publicity photographs is seen as highly important because it is believed that the quality of the photographs influences the quality of marketing actions and thus indirectly, the response rate to a marketing action. In the IECT case, quality control is seen as removing low quality photographic material, and avoiding too much marketing exposure of the same photographs in the same time period. Another aspect is that it may prove useful for advertising agencies to examine and be inspired by frequently reused, popular photographs. The reuse frequency for certain photographs can be seen as a quality measure, namely a measure for the appreciation of a photograph. The reuse frequency can be monitored by noting the number of requests for a photograph. The final business objective of the IECT photo archive is related to improving management control of costs related to the usage of photographs (O6). In the old situation little management control existed with regard to the management, and production costs of photographic material. In the new situation this has changed. It is anticipated that, in the near future, retrieval and delivery costs will be charged directly to the customers of the IECT photo archive. The charging of costs is important to make internal archiving costs visible to PTT Telecom and its organisational units, and to provide incentives for efficiency. The implementation of cost charging also fits well with the corporate policy to strive for financially responsible, self-supporting units. Future extensions of the business objectives are being considered by IECT management, it is intended that the above-mentioned business objectives will also be applied to other types of communications. In particular, the saving of costs by better reuse (O1) and improving accessibility (O4) may also be applied to the archiving of company videos, animations, and selected PR documentation, like business presentations, leaflets, etc. This will eventually lead to a more complete MCA. PTT Telecom and IECT management decided to set up a central archive assuming it is a viable way to meet the discussed business objectives: to save costs, to improve the access to corporate photographic material (a centralised photo archive solves the problem of where to look for the material), to implement quality control and to improve management control. The choice of an electronic archiving system was made because it was assumed that only by using such a system could the short search times necessary to meet the accessibility objective (O4), and cost savings due to more efficient management and retrieval (O2) be realised. The choice of a multimedia archive (including photographs) was made because it was assumed that this would reduce search times (O4) and thus leads to more efficient retrieval (O2) if photographs are displayed and can be seen by searchers. Electronification of the display of photographs has another advantage: the originals are not damaged by browsing and can be kept safely in the physical archive. 3.3.3. Information processing tasks A number of information processing tasks are implemented during the setting up of the IECT photo archive. Figure 13 shows which tasks are performed by whom on what information within the IECT photo archive. A description of these tasks is given below. Requests are made by customers of the IECT photo archive for the retrieval of photographic materials. Requests are made known orally, often over the phone, to the IECT archivist. During a discussion with the customer the archivist tries to interpret the request. Then, the archivist searches for matching material in the electronic archive, and produces a list of descriptions of photographs together with a reduced print of retrieved photographs. The archivist administers the search result and sends the list and print to the customer. Next, if the customer confirms the order for the photographic material, duplicates of this material are retrieved from the duplicates archive, and if no duplicates are available a reproduction order is send to the Capi Lux photographic service. Capi Lux delivers the ordered duplicates within, at most, three days to the IECT archive. Finally, the complete order of duplicates is sent to the customer. A further task of the archivist is to archive incoming photographs and slides, and extra duplicates. Archiving is a rather complex task. It involves: • scanning, indexing, and storing of digital images of the material; • storage of originals and duplicates; • thesaurus management. Thesaurus management is actually performed by a specialised archivist, a thesaurus manager.
A number of extensions or changes to the information processing tasks are being considered by IECT management on the basis of the business objectives discussed in the previous section. The charging of costs to customers (O6) has not yet been implemented. As yet, it has not been decided, by the IECT management, how to calculate the costs to be charged and when to set up such a charging task. To charge the customer, it is necessary to administer customer orders carefully, and to keep a financial order administration. IECT management is considering giving regular, frequent customers the possibility to search directly in the electronic archive from their own offices, and because of this about 20 external access points are anticipated. Trained personnel from the business units may search the archive on-line, and the ordering process may be also on-line. The main argument for this is that it lowers the threshold for reuse of photographs, which is in accordance with the objectives (O4, O1) discussed in the previous section. Further it will probably reduce the need to consult the archivists of the IECT archive, and accelerates the search and selection processes (O4). A possible new 'product' of the IECT photo archive and another way to reuse photographic material (O1) is to deliver images of photographic material. These images are used by the marketing units producing electronic product information. Delivery of electronic images reduces the need to digitise and code information separately for every MM sales support system. A main issue to be resolved is the standardisation of video and still exchange formats. To extrapolate from these development it is possible to envision an archive in which customers also retrieve marketing communications electronically, and print locally, if necessary. Structured tasks like charging and billing, and delivery, can also be automated. The archivist function can then be reduced to ill-structured tasks like indexing of pictures, and system maintenance tasks. A description of the functions of the electronic archive is given in this section, based on the current situation. An overview is given in figure 14. The main functionality of the electronic archive is related to database system functionality, and because of this the MM system and database system aspects are discussed in combination.
The main system functions are (see figure 14): • Photo capturing: photograph capturing is performed by a photograph capturing system. Photo capturing includes input and digitisation of photographs and slides. • Archive update: after photographs and slides are captured they are archived using the update facility of the electronic archive. The update facility archives images in the photo database, and stores related photo index records in a photo index database. The index terms are selected by the archivist from the thesaurus database (controlled index list) or are entered in free format. • Thesaurus management: the thesaurus manager uses a thesaurus interface to manage the thesaurus database, i.e., to update the index terms in the thesaurus database. (Other database management activities are not directly supported by the electronic archive. They include the back-up of database files). • Search facility: the search facility supports the archivist in searching photographs. Search terms are entered by the archivist or selected by the archivist from the thesaurus database. Searches are performed on the photo index database. Searches can be stored in the search history and can be re-executed. The result set of a search with references to photographs in the photo database is presented to the archivist. • Presentation facility: the presentation facility retrieves images from the photo database; the result set with references to the images, produced by the search facility, is used for this purpose. The presentation facility delivers prints of images (prints to screen or prints to paper) to the archivist.
On the basis of the possible extensions of the information processing tasks, discussed in the previous section, we can anticipate several extensions of system functionality: • direct access to photo searching and photo presentation functionality by authorised customers of the IECT archive; • export functionality for the electronic delivery of images to customers; • charging functionality for charging and billing of delivered photographs and slides; • extended input functionality, update functionality, search functionality and presentation functionality for digital video, audio and animation; Extended search functionality like full text search mechanisms may prove useful for retrieving PR documentation and leaflets. Hyperlinks from the photograph in the digital photo archive to the leaflet where it is used may also be very useful. The MRS implementation aspects for the IECT photo archive are discussed in this section. First, general MRS implementation aspects, then MDBMS implementation aspects are discussed. The electronic archive is a simple MRS. A caption camera is used to capture photographic material. The images are stored in an analogue format on duplicated video disks (one as a backup). A video disk may contain up to 60,000 of such images. Currently, a PC controls the video disk, and is used for data-entry of descriptions and searching through these descriptions which are linked to the stored, analogue images. A TV monitor is used as the output medium to display the analogue images. A thermal transfer printer is used to print the photographs (9 photographs to a page) for selection purposes, and a normal printer is used to print alpha-numeric descriptions. The system software, alpha-numeric descriptions and search requests are stored on hard disk, and a back up is made via an Ethernet-LAN. No really integrated database management facility (see chapter 2) is used for the management of the photo database and the photo index database. The photo database contains now about 6000 images of photographs and slides. (Only those photographs are stored which are likely to be reused). The photo index database contains the same number of records with alpha-numeric descriptions of these images. Only the storage and presentation control are integrated in the PC software. The retrieval engine used is also fairly simple. The DBMS Picbase offers the update facility, search facility, presentation facility and the thesaurus interface used to rubricate the stills. The version of Picbase used only allows for a two levelled rubrication: a broader term and a narrower term. Examples of rubrication terms for the IECT photo archive are: activity - sport objects - building objects - telecommunication infrastructure Other essential attributes of the entity type photograph are a unique reference to the physical archive, and a unique reference to the laser disk, photo production data, background, direction of view, colour, type of object, activity, atmospheric description. An example of a unique reference to the physical archive is: P-130-01254 (P = slide, 130 = size 100x130 mm, 01254 = serial number 1254) New versions of Picbase will support more thesaurus functions. The Picbase thesaurus does not support any international thesaurus standards at the moment. No motivation is given for the choice of Picbase and its thesaurus facility. Other search facilities and other retrieval software are simply not considered. The intentions of IECT to improve the accessibility of the photo archive by offering marketers direct access to the photo archive system brought to light limitations of the current implementation: • Analogue images need to be converted into a digital format to communicate them using a digital system and digital network. To eliminate this bottleneck, a conversion effort has been started to store the stills of photographs and slides in a compressed digital format, JPEG or FIF (a fractal interchange format), on a digital medium. The advantage of FIF over JPEG is that FIF results in higher compression ratios (and thus more efficient transmission and storage). ISDN is the preferred transmission medium for remote access. • The current implementation on a DOS-PC is single-user, and needs to be prepared for multi-user access. To avoid the limitations of DOS, one can imagine a migration towards Windows NT or a UNIX environment. If this happens, it will be necessary to choose a new retrieval engine instead of Picbase. • The implementation of the thesaurus search facility is considered by IECT to be too complex for the archive customers. Adequate use of this search facility requires archivists knowledge of the way the stills are rubricated, and this is a bottleneck for untrained customers. Straight forward QBE (Query By Example) screens could be offered to overcome this problem, requiring the presence of a query language, and less complex (= confusing) screen lay-outs. The next version of Picbase will offer a possibility for searching on attributes of database objects, this will probably suffice to support QBE. • The response times for searching a selection of objects in the database are too
long (many minutes), which is not seen as acceptable for customers. The response time must
be shortened to seconds. This may be done by indexation of the rubrications and database
fields. If the IECT photo archive is extended towards a more complete MCA than an MDBMS should be considered, e.g., the ORACLE DBMS, this is a company standard within PTT Telecom. An MCA like the IECT photo archive is clearly technically feasible, and seems viable if the acceptance of the system by the IECT management and users, and the growing use of the IECT archive by marketing departments and other customers is taken into account. A number of strengths and weaknesses have been reported by the IECT respondents or can be observed. First of all, only a very rough cost benefit analysis was performed beforehand. The business value added of the IECT photo archive was analysed predominantly in qualitative terms and very rough quantitative terms. Nevertheless, this analysis was convincing enough for the PTT Telecom and IECT management. Second, MM and retrieval are reported to have a clear value added. The observed value added of MM is mainly that photographs are better and faster selected when they are presented together with their retrieved descriptions. The observed value added of retrieval is related to reducing search times. Third, the photo archive is seen as a first step towards a full-grown MCA. This growth path was only defined by the IECT management after the first version of the electronic archive was delivered. This makes conversions necessary that could otherwise have been avoided. Other weaknesses reported by the IECT respondents are that there was little IT and automation knowledge available within IECT when the IECT photo project started, too limited requirement analysis and system specification was carried out with an insufficient review of implementation alternatives, for example, with regard to the choice of the retrieval engine.
3.4. The Tele Sales Assistant (TSA) system A telephone sales assistant is someone who can be called for information or help about products and services; in their turn assistants need assistance with the marketing and sales activities, from a Sales Support System, called Tele Sales Assistant (TSA) system by the business unit Consumer Market of PTT Telecom. In general, a TSA system can be described as follows: a TSA system is a front-office sales assistance system which offers support for order intake and complaint handling, that is integrated seemlessly with back office order handling and product management processes. The TSA project started in September 1991 with a preparatory study, from January 1992 to the Summer of 1993 several prototypes were developed to demonstrate TSA system functionality to the Consumer Market management and to further elucidate specifications. A requirement and specification study was conducted during the same period. In November 1993 the TSA system was implemented at two test sites. March 1994 saw the start of a pilot phase in which a more stable TSA computing environment was set up, one that was over the teething troubles. The TSA system was introduced nation wide in the Autumn of 1994. Total development and implementation costs, including training, were about 1 million guilders (HFL). Operational costs are estimated at about HFL. 250,000 a year. Further, it has been estimated by the TSA project team that the implementation costs will be recovered in approximately 1 year from nation wide implementation. The following descriptions of the M&S business objectives and information processing tasks, MRS functions and MRS implementation aspects are based on several interviews with the manager responsible for the TSA project, several hired in developers, and TSA documentation (Peeters, 1993), plus incidental participation. The business objectives (O) of the TSA project are to improve the quality of service to consumers, to increase the productivity of tele(phone) sales employees, to reduce training costs, and to improve internal information processing. Consumer Market management believes that improving the quality of service to consumers (O1) will lead to better customer relations and (indirectly) to more turnover. Improving the quality of service is translated into: • improving the telephone contacts by better informing customers; • improving the accuracy with which questions are answered; • improved dealing with consumer complaints; • shortening waiting times for consumers. The objective of the TSA project is to increase the productivity of telesales employees by an improved return on telephone contacts with consumers (O2). It is believed that an increase in productivity is possible by: • dealing with more callers during a time period; • unhampered order intake; • improvements in informativeness and ability to give convincing selling arguments. An increase in productivity is influenced largely by factors other than the introduction of a TSA system, in particular by marketing actions to stimulate customers to contact telesales centres. Another objective is to improve the knowledge level of telesales employees (O3). It is assumed that telesales employees with a sound knowledge of products and services are more productive. In the old situation, the training of new telesales employees formed a serious bottleneck, in part due to a high turn over of telesales employees and a limited training capacity. Discussions with TSA project team members made clear that some extensions to the business objectives can be expected if the TSA system proves successful. The most likely extensions are applying the business objectives to other types of sales personnel, namely shop employees in the consumer shops, and personal sales personnel, business shop employees and telesales personnel for the business market. It seems very likely that the business function of the SSS (Sales Support System) will be extended, sooner or later, by including support of shop sales. The business units Consumer Market and Business Market both run telecommunication shops, called respectively, Primafoons and Business centres. For shops, the business objectives are increasing sales volume, improving the quality of service to customers, more efficient order intake and improved promotion. An interactive training module as part of the TSA system seemed necessary when thinking about a solution for the training of sales employee (O3). A side effect of such a training module is that training on the job becomes possible and that external training costs can be reduced. Nevertheless, the most important expected result is having telesales personnel trained more quickly. With regard to the first two objectives (O1 en O2) it is believed that improved sales support in the form of a TSA system will contribute to meeting these objectives by removing bottlenecks found in the old situation. One of these bottlenecks is formed by the limitations the sales support used currently: a paper telesales catalogue. The paper telesales catalogue is insufficient to support telesales employees effectively for several reasons. First, topicality of the paper catalogue is not as high as needed. Second, the catalogue is not complete (e.g., news about marketing actions is absent). Third, searching through paper manuals is slow. To remove these bottleneck the choice was made for an electronic TSA system, which • is complete and topical; • gives fast results, a few keystrokes calls up the information required. (It is observed at the TSA system test-sites that searching for information in the TSA system is indeed much faster than searching for the same information in the paper documentation). A TSA system must support order entry and reporting complaints, which can be processed by other employees in the Consumer Market unit, to facilitate unhampered order intake and handling of customer complaints. A multimedia system was chosen as it was assumed that showing pictures of products has important advantages. An important (assumed) advantage is that communication about products between customers and telesales employees is eased (O1) when the telesales employees are provided with the same catalogue information as the consumer ("I want the phone with the red buttons, not the one with the curly string"). It is also assumed that when the presentation of product information is improved by showing product pictures, product specifications become more understandable for telesales employees. Next, it is assumed that improved retrieval and presentation of assortment related information to telesales employees results in an improvement in answering customer questions and improvements in informing customers about products and services (O1). 3.4.3. Information processing tasks Based on the business objectives and choices discussed in the previous section a number of information processing tasks were selected to be supported by the TSA system. Figure 15 shows what tasks are performed by the employees involved on what information. A description of the tasks of involved employees is given below. There are about 500 telesales employees spread over 32 regional call centres. They cope with the requests of domestic consumers and increasingly of business people. The types of requests dealt with include request for information about the cost of a service, handling a problem with a telephone connection, or placing an order. The telesales employee takes a note of the customer request and tries to give immediate feedback. If the telesales employee does not know the answer they can use the TSA system. The TSA system gives up to date information about products (characteristics, selling arguments, current prices, special offers, etc.), newsletters, and current marketing actions. Orders are entered in an order entry system and problems are reported. The handling of problems and orders is outside the scope of the TSA system. As can be seen in figure 15, the TSA system is a valuable information source for other employees, like marketers, sales managers, and product managers. An editor updates the contents of the TSA system, and a teacher the contents of the training module when necessary.
If a decision is made to extend the business objectives to other sales outlets it is necessary to consider to what degree other information processing tasks need to be supported by the SSS and to what extent other information items need to be included. (If very different tasks and information items are involved it is more economical to develop completely separate sales support systems for the different outlets). In the case of support for shop employees some differences can be noted. Shop employees are under more variable pressure of time than telesales employees, and they already work with a number of information systems, like an order entry system. An SSS needs to be neatly integrated, as having many different terminals on the counter (for every IS one) is not workable for shop personnel. An important difference from telesales employees is the physical presence of customers, presentation of SSS information to the customer can be an additional task. In case of sales support of personal sales some other important differences can be found. Personal sales differs from other types of sales in that customer relation management is stressed, personal sales staff are more mobile, and personal sales staff often combine customer care with back office activities like developing sales plans. The SSS for personal sales staff therefore needs to be integrated seemlessly with the software applications they use, like a presentation tool, word processor, an order entry system, and a customer IS. In general, (tele)sales personnel can not be seen as high level IT users, who can be directly confronted with the more complex retrieval facilities. This leads us to another basic requirement for any SSS offered: it should be easy to use by relatively untrained IT users. A description of the functions of the TSA system is given in this section. An overview is given in figure 16. General TSA system functionality is discussed first followed by a discussion of TSA database system functionality. The main system functions are: • Order entry: the TSA user interface offers functionality for order entry to telesales employees. Entered orders are sent on to an order system. • Problem entry: the TSA user interface offers functionality for problem entry to telesales employees. Entered problems are sent on to a complaint system, a strategic corporate resource. • Assortment retrieval: assortment retrieval is a main function of the TSA system. Telesales employees, marketers and product managers use this functionality. The TSA user interface sends retrieval request to the TSA database system, which returns requested assortment information for presentation. • Product news retrieval: telesales employees particularly use the product news retrieval functionality. The telesales employee makes a selection, which is translated into a request to the database system. The database system returns one or more product news leaflets, which are presented to the user by the TSA user interface. • Telesales news retrieval: telesales employees particularly use the telesales news retrieval functionality. The telesales employee makes a selection, which is translated into a request to the database system. The database system returns one or more telesales news pages, which are presented to the user by the TSA user interface. • Training: a training function is offered to telesales employees (see section 3.7. for more details). The answers to questions given by telesales employees are registered by the training function in the TSA database system. Training information is retrieved by the training function from the TSA database system. The TSA system offers a production environment for the: • Production of assortment information: the TSA editors prepares information from a corporate assortment IS to be updated to the TSA database system. • Production of news information: the TSA editor is supported by functionalities to digitise news information and to edit information in the product news and telesales news databases. • Training editing: teachers are supported in authoring courseware and updating the training databases in the TSA database system.
The database system functionality is separated into two environments: the production databases environment on the right and the retrieval databases environment on the left. The reasons for this separation are discussed in the next section. The main database system functions are (see figure 17): • Updating: the update facility offers updating functionality to the production environment for the updating of telesales news, product news, assortment and training production databases. • Data exchanging: the exchange facility exchanges database information from the production databases to the retrieval databases. The database administrator initiates data exchange by submitting an exchange request to the exchange facility of the TSA database system. • Searching: the search facility offers the TSA (user) interface search functionality. The TSA interface sends queries. The search facility returns result sets. • Presentation: the presentation facility accepts presentation requests (including references in a result set) from the TSA interface and returns information objects from the telesales news, product news, assortment and training databases.
Taking as a basis the extensions considered for the information processing tasks, discussed in the previous section, several extensions of system functionality can be anticipated: • direct access to the TSA system by customer, limited to the TSA (user) interface; • commercial presentation functionality, to confront customers with attractive promotions. Direct access by customers puts heavy requirements on the ease of use for the user interface of the TSA system. MRS implementation aspects are discussed in this section on the basis of the subsystems shown in figure 16, followed by a discussion of MDBMS implementation aspects on the basis of the database system functionality shown in figure 17. The decision was made to implement the TSA system in an MS-Windows/PC environment, because the MS-Windows/PC environment is a standard part of T-workplace, and T-workplace is the standard MS-Windows/PC based equipment for a PTT Telecom workplace. The TSA interface and production environment were programmed in Visual Basic. Currently, a major technical limitation is related to on-line transmission: offering on-line pictorial retrieval by a nation wide Wide Area Network would lead to high telecommunication costs. Therefore a cyclical downloading procedure was set up to update the local file servers of the 32 telesales centres (one server per centre). The TSA system software and databases are stored on these local file servers, a 486-PC, accessed by a 10 Mbit/s Ethernet. The TSA databases are downloaded by the business network (alternative transmission media for remote update are N-ISDN, PSTN or the business network Themis). A more severe bottleneck at present is that downloading of large volumes of video data via the Wide Area Network is not affordable. It will take some years before the price/performance ratio has improved enough too make downloading video files affordable. It will take even longer before on-line video retrieval by a WAN will be affordable. The production environment consists of a digital camera to photograph products and a scanner for input of pictures in GIF format, a download interface to the PTT Telecom assortment IS, and an authoring interface to edit the content of the TSA databases. The GIF format was chosen because it is a widespread graphics format and offers an acceptable compression ratio. (Better than TIFF for example). The authoring interface was programmed in Visual Basic and offers the functionality defined in the previous section. The TSA interface consists of a PC with VGA monitor placed on the desks of the telesales employees, a keyboard and mouse for controlling the application, and Visual Basic software for accessing the local databases in telephone centres. The information types used are: • graphics for the human interface; • GIF pictures of products; • text to describe products, to give news about marketing actions and products, and to ask training questions; • and instruction videos (AVI) in the training module. TSA system GIF pictures are presented to the user in two sizes: small and full screen. The full screen picture of a product appears if the user clicks on the small picture which is presented together with product information. The full screen picture shows more visual product details. Instruction videos are not yet included in the operational TSA system, only in a prototype, because they require too much hard disk and network capacity. An option is to use optical disks for training material including full motion full screen video. This would relieve the business networks and the local file server, however, the disadvantage is that changes in the TSA system software are not so easily updated to the training system. For every training database update, new circulations of optical disks need to be produced and distributed. The TSA interface offers retrieval by: • a hierarchical menu structure, which can be rather laborious; especially when the database is growing, this limitation of menu structures becomes a time consuming bottleneck; • graphical browsing through a list of product categories and names; • by offering the possibility to search on key words. Key word search functionality is presented in the form of a fill in the blanks screen (Query By Example - QBE), in which a sales employees fills in, for example, a product name field and/or a product colour field. The inclusion of hyperlinking is considered for the TSA interface, to permit direct jumping to related information by clicking on hot words to trace likely associations. If speech recognition software improves it may be possible to include query by speech in the TSA interface: one can imagine the telesales employee repeating the product name the customer is interested in. Speech input is much faster than input by typing but voice input may disturb TSA interface interactions of other employees. Some TSA system extensions or modifications are discussed in the previous sections. Extending the TSA system towards a SSS for shop personnel would not really require different types of media, however, extending towards a SSS for personal sales does lead to transmission media extensions, as personal sales staff is mobile. This would require retrieval possibilities using a notebook with a communication interface (for ISDN) to access topical data. Spreading the SSS data by CD-ROM or CD-i is an alternative, although this still requires an on-line connection possibility to access information that has a limited life span, such as price information, marketing actions and product news. A network connection is also needed for ordering. The high cost of introducing notebooks with CD-ROM and/or ISDN connections for all personal sales staff is a drawback. The TSA database system consists of an Access DBMS from Microsoft as the retrieval engine, Access database files, and GIF files in directories. In total, the size of the TSA databases is approximately 100 MB (excluding video files). The MS-Windows DBMS Access, a conventional DBMS offering all basic DBMS facilities (see chapter 2), is only used for management of the structured data. Ill structured data (i.e., the GIF files) is retrieved by external references to their path and file names on hard disk. Hence, we can conclude that the extended DBMS approach is followed (see chapter 2). Hidden from the user, the search facility uses a standard DBMS search facility (SQL based). The Access presentation facility is used to present retrieved database objects to the TSA interface. The update facility and exchange facility are used to update the TSA databases from the production environment. This includes assortment updates from Oracle databases (the Assortment IS is implemented by the Oracle DBMS). An alternative retrieval engine, which takes into account the possible growth path of the TSA system towards a more inclusive SSS, would be a real MDBMS. The advantage would be that the integrity of references to ill structured data objects could be better controlled. A disadvantage is that the retrieval performance would slow down, but this can be compensated for by faster PC hardware. The TSA system is feasible and seems viable if we consider the enthusiastic responses of telesales employees and telesales managers, and the strong management commitment. We should, however, be cautious about enthusiastic responses: they can disappear as fast as they come. A number of strengths and weaknesses were mentioned in discussions with the TSA project team. Measuring the value added of the TSA system is seen as important, evaluative measurements are scheduled although it is hard to isolate the TSA effects on quality of service to customers and productivity because there are many other influencing factors. The assumed value added of MR of the TSA system is based on a better recognition and understanding of products by seeing pictures, and shorter retrieval times in comparison to retrieval using paper catalogues. The strong management commitment is seen as a very positive aspect, that is extremely
important for initiating changes in such a large organisation as PTT Telecom. High
involvement of the sales people, who are going to use the TSA system, is seen as a success
factor of the TSA project. It is believed that convincing the sales staff of the value
added of the TSA system, in terms of easy access to topical and complete information in
support of their sales conversations, is a crucial success factor for the TSA system.
Changing (re-engineering) the business processes in combination with the introduction of
the TSA system is seen as a very difficult job. Not so much because of the complexity of
the tasks, but because of the social and political hurdles that have to be overcome. Other
difficulties mentioned are lack of familiarity with MT (MM technology) applications within
PTT Telecom, and integration with the corporate IT infrastructures. |
© 1995-2002 Martijn Hoogeveen |
Figure 11. Symbols used in
diagrams. 
Figure 12. Simplified diagram with the
basic symbols used. 
Figure 13. Information processing
tasks within the IECT photo archive. 
Figure 14. Functional elements of
the electronic archive. 
Figure 15. TSA system in relation
to the tasks of a telesales employee. 
Figure 16. Functional elements of the TSA
system. 
Figure 17. Functional elements of the TSA
database system.