THE TELER PROJECT, A SHORT DESCRIPTION (1)
Gerrit de Bolster, Statistics Netherlands
A large number of enterprises, especially small and medium-sized, express growing concern over the heavy burden imposed on them by data collecting bodies, be they National Statistical Institutes (NSI’s) or other public authorities. It is specifically to tackle this problem that the TELER project was launched at the beginning of 1996. The TELER consortium is made up of three key components - NSI’s, National Producers Associations (NPA’s) members of EUROFER (the European Confederation of Iron and Steel Industries) and EDIFICAS (a Europe-wide federation of accountants)
The TELER project aims at lowering enterprises’ administrative burden, especially in the field of statistical business surveys, thereby improving the work of national statistical services and other collectors of statistical data by using EDI-related techniques in combination with the existing European telematics infrastructure.
The objectives are the following:
1. Workload lessening as the right use of new technologies can
*) help automating the preparation of enterprises’ answers to administrative and statistical data requests,
*) provide transmission facilities;
2. Common standards using the same exchange model for every statistical and if possible administrative requests, even if this model can be specifically adapted to country, enterprise size, industry, legal form peculiarities; in other words, organisational standards in the form of a common language rather than technical standards;
3. Simplification of statistical data collection by clear prior notification of the full set of statistical requests through the definition of what some statistical offices call an electronic combi-questionnaire.
TELER is a 3-year project which, in the end, should create the impetus for different providers and collectors of business data across Europe to get together and set up a framework for making this data accessible and comparable.
2. RATIONALE OF THE PROJECT
Nowadays, most data are communicated through paper questionnaires filled out by company clerks (looking for requested data, transcoding data from the in-house classification into the collector’s own classification, aggregating and computing various aggregates (averages, consolidation), typing, checking, etc.). The paper questionnaires are sent by ordinary mail or fax by the company to the data collector. In some rare cases, data is sent by phone (touch-tone system), or floppy disks, or even EDI messages. Whenever hard copies are sent, the data collector, in turn, has to perform various tasks, carried out by specialised personnel such as keying in the data, processing, analysing and reporting it, etc.
The whole system proves very costly, both for enterprises and data collectors. This burden is borne with more and more reluctance by businesses, especially the smallest ones (for which this burden is hardly bearable). All in all, it could result in serious difficulties for public data collectors in the future. What is more, the system is not rational: data moves from electronic format (in the internal enterprise IS) to paper forms and, in the end, back to electronic format at the data collector’s.
Reducing the burden is therefore crucial for business competitiveness and for government efficiency. The solution rests in surveys’ simplification and, above all, in the automated derivation of data from businesses to data collectors’ information systems. Most data should be computed from management information systems, i.e. all data required by data collectors and already existing in computer-accessible form. Yet, this solution raises a series of problems and difficulties which must be analysed and coped with. The TELER project is concerned with this analysis and will try to find innovative and viable solutions. To demonstrate the feasibility of the solutions envisaged, TELER partners chose to focus on a limited number of organisations and data. TELER therefore concentrates on statistical variables collected by statistical institutes and partner associations (information brokers), i.e. on data required from enterprises by NSI’s, with or without the mediation of National Producers Associations or accounting firms.
Data collectors are still asking for data of all sorts on paper questionnaires. Enterprises fill in those questionnaires, often after referring to their automated information systems. Data requested by data collectors can be divided into two categories: (i) data generally available in the automated information systems of the enterprise, (ii) other data. Figure 1 illustrates the current data flows between enterprises and data collectors.
It will be implemented at CPS, in FR, involving 18 steel works. It will include the building of 3 interfaces enterprise/CPS to cover all cases. The query tools BUSINESS OBJECTS will be used.
Since the demonstrator will be implemented in a variety of situations, the technical environment must be based on common platforms available in all situations: PCs, Windows and off-the-shelf packages and on the telecommunication infrastructure that is available. It will thus be possible to propose standards and software prototypes which operators and data collectors will be able to implement and market in Europe.
In the work already done, an inventory has been made of the existing standards and methods. The data items (or statistical variables) used in different countries in order to gather similar data are now listed using a common framework. The elements to be taken into account in order to create or use common models and standards are now present. Thanks to an evaluation of the hardware and software platforms of future users, the main targets for future dissemination now have been identified.
The main functions of the demonstrator are:
a) Building a repository
b) Data entry
c) Data import and control
d) Data and metadata handling
e) Data transmission
f) Data storage
The main test criteria are: usability, security, user friendliness, time to implement the packages, national and cross-national inconsistencies, resulting data quality, cost elements and reliability of data transmitted. Additional subjects for consideration, not included in the demonstrator during the trials, are (g) help desk support functions, and (h) information feedback. These subjects will be covered in the recommendations for further enhancements and exploitation.
4. CURRENT STATUS OF THE WORK-PACKAGES
4.1 Progress in WP01 (co-ordination and management)
Management and co-ordination have been changed according to agreements with DG XIII during and following the Annual Review of the project in October 1996. The main modifications are:
*) setting up a Technical Board, in which technical solutions are discussed with an emphasis on comparability and common models;
*) setting up an Editorial Board, which reviews and controls the contents of all deliverables.
4.2 Progress in WP02 (dissemination and participation in external groups)
This activity included, among other things, the participation of the User Group to discussions on the project. The User Group is allied to the SERT ad hoc group, an initiative of Eurostat on the use of telematic networks for statistical data collection among enterprises. SERT includes all European and EFTA countries' statistical institutes, a few government departments interested in automated data collection and a number of expert companies (consultants and software houses). Two meetings of the User Group have been held in Luxemburg (13/12/96 and 20/6/97). Deliverables No. 1 and 2 have been disseminated to group members on this occasion and are available upon request.
This activity also included the presentation of the TELER project to various international meetings, such as:
*) IDA discussions held on 28/2/97 in Brussels
*) The SNTE working group (3/97 in Luxemburg)
*) The Working party on EDP - UN Conference of European Statisticians (2/97 in Geneva)
*) Plenary meeting of the UN Conference of European Statisticians (6/97 in Geneva)
*) The Expert Group meeting on "Applying new methods and technologies in international statisti-cal databases", convened by the UN (5/97 in New York)
*) "What Financial Controllers and Accountants need to know about IT" (2/97 in Singapore)
*) 20th UN/EDIFACT JRT Meeting - Presentation of INFENT application at JM6 and JTAG - (4/97 in Singapore)
*) "Data collection re-engineering", working party with Dun & Bradstreet USA (6/97 in Brussels)
*) The ISI 1997 Conference (8/97 in Istanbul)
*) The DOSIS meeting, convened by Eurostat (9/97 in Luxemburg)
In addition, the project leader participated in a concertation meeting organised by the EC in Brussels (2/4/97). Articles and papers were published individually by members in specialised magazines.
Specialised meetings with software houses with a view to future exploitation plans were also organised (see WP11).
4.3 Progress in WP03 (user requirements)
The objectives of this work-package are:
*) Specify the users’ requirements (NSIs, other data collectors, enterprises).
*) Specify the enterprises’ environments and expectations regarding business surveys.
This work-package was finished in 1996 (Deliverable No. 1).
4.4 Progress in WP04 (development of variant 1)
This work package is being carried out by the participating NSIs. Its main objectives are:
*) To build the data model for this variant and to define the functions to be supported by the EDISENT software.
*) To develop EDISENT, and in particular to install the combi-questionnaire for use in NL, SE, DE, FI, PT, ES and IT (at enterprises and NSIs), to install the EDIFICAS coding system for the use in NL, DE, SE, IT (at accounting firms and NSIs) and NL again (at accounting office and Tax authorities).
*) To verify the EDISENT software (NL).
In order to meet these objectives the following tasks have been carried out during the reporting period:
*) Further definition of the NSI data model (data types, relations and definitions), resulting in the NSI organisational data model which provides the basis for the electronic combi-questionnaires (completed).
*) Detailed definition of data types to be involved in the trials in NL, SE, DE, FI, PT, ES and IT, resulting in an enhanced version of the NSIs' Access databases (completed).
*) Preparation of the contents of the combi-questionnaire in NL, SE, DE, FI, PT, ES and IT (in progress).
*) Definition of the functions to be included in the EDISENT software (in order to be able to demonstrate the electronic data flows between enterprises and NSIs, as well as between accoun-ting offices and NSIs/tax authorities (global specifications completed, work on detailed specifi-cations in progress).
*) Building and adaptation of the EDISENT software (a first prototype to be used under MS/DOS was shown during a work shop at the NSI meeting in May 1997; building of the second version to be used under MS Windows is in progress).
*) Building and testing of software tools for the translation of EDISENT system and help texts into the various national languages.
*) Building and testing of software tools for the translation and implementation of the combi-questionnaires in the respective national versions.
*) Adaptation of the prototype, including the installation of the combi-questionnaires for the use in NL, SE, SE, FI, PT, and IT (in progress).
*) Verification of the EDISENT software (in progress).
*) Contributions to Datacare's efforts on specification and building of the data communication module to be used (global specifications completed, work on the detailed specifications is in progress).
The results achieved so far have been documented in Deliverables D02 and D03. Deliverable D04 is in preparation. In addition to this, many working documents have been made and the various NSI Access databases have been updated. It should be noted that, in addition to the regular Consortium members, INE Spain participates in the NSI activities on a voluntary basis.
4.5 Progress in WP05 (development of variant 2)
The objectives of this work package led by EDIFICAS are: definition of the operational data model for accounting firms, definition of the global specifications, development of a prototype and verification. This variant is more focused on meeting the needs of SMEs. One of the main issues outlined in WP03 is that accounting systems do not provide the expected facilitation for statistical data collection, even in countries where standard chart of accounts exist by law. To bridge this gap, a bottom-up approach was used, finally fully complementary to the top-down combi-questionnaire approach of the NSIs. The coding system proposed by EDIFICAS being the missing link between the two approaches. An organisation data model, then an operational data model were thus built. The required functions were also determined with a view to embedding them into an already existing software package, published by CCMx (who claims an 80% share of the market of packages for accounting firms in France).
These models and prototype will be tried in demonstration trials, among which one will include data submission to two different public administrations (statistical institute and tax authorities), with data covering the balance sheet and profit & loss statements plus data specific to the French enterprise structural survey.
During this work package, an agreement was reached with the French statistical institute who will accept to collect elementary data instead of aggregated data and check sums (the corresponding functions of aggregation and checking will therefore have to be implemented in the software at the data collector). This agreement with INSEE (who is not part of the TELER consortium) is a very encouraging sign for future exploitation of TELER and the confirmation that public data collectors are able to change their own paradigms when usual paper form checks are no longer necessary. It shows that facilitation can cleverly be shared or split, increasing the win-win objective at responding parties and turning the administrative data provision obligation for enterprises into a partnership with collecting administrations.
4.6 Progress in WP06 (development of variant 3)
The objective of this work package led by Eurofer is to develop a basic data and process model, only run with off-the-shelf software, applicable, after possible adaptations, by all industry sectors in all countries, for collecting data electronically from enterprise information systems, before or after trans-codification, and storing them into a national professional association’s relational database capable of providing the enterprises themselves and public administrations with the required information on electronic medium carrier, with the help of an end-user query tool.
Only for trial operations, has the model been limited to delivery statistics and ECSC questionnaires submitted to Eurostat. Even though the model has been conceived in common by three national professional associations (F, GB, GE) it will be demonstrated only in France. Under this work package the following tasks were achieved:
*) participation in the common data model
*) building the Eurofer conceptual and organisational data models, on the basis of the common model, but adapted to deliveries, including triangular operations
*) building the three possible applications of this model in the steel industry in each of the participating countries (Access databases)
*) building the operational data model to be demonstrated in France
*) building the list of required global functions of the software on the basis of the data models
*) preliminary work at CPS (the French member of Eurofer) for trials (revising, clearing, changing, mapping nomenclature tables)
*) search for standards
*) implementation of the tables in the French Access database
*) selection and installation on a PC of the off-the-shelf software to be used in the demonstration (Sybase, AMC Designor, AION-DS, Business Objects).
The main results in terms of data models and description of process are shown in deliverables D02 and D03.
4.7 Progress in WP10 (technical co-ordination and common models)
The bulk of the work in this package was devoted to the building of a common model, which means the sharing of common concepts by all Consortium members, whatever their origin (statistical institutes, accountant and professional associations). Its difficulty to achieve technical co-ordination stems from the discrepancies in Europe in the situations of data exchanges between businesses and government in terms of national practices, of organisation and of technologies used. The main technical co-ordination work was done by the Technical Board under the management of Statistics Netherlands for the NSI part and CESIA for the Consortium as a whole, with all partners being active to enrich the common model.
The collaborative work carried out in this work package had two results:
*) a common conceptual data model and its relationship with variant-specific organisational and operational data models, as they are elicited in Deliverable 02 (first 2 chapters)
*) in addition to the overall coherence, a significant enhancement and enrichment of each variant.
4.8 Progress in WP11 (exploitation plan)
This work package started earlier than foreseen in the initial Work Programme. At the present stage, the objective is only to sketch out the main orientations of future exploitation. Therefore a document was issued showing the 3 main beneficiaries of TELER (each organisation participating in TELER, "sister organisations" and Business and Government circles) and the achievements so far, such as :
*) non-participating NSIs are kept informed about TELER progress;
*) participating NSIs started promotion of the TELER project in general and EDISENT software in particular within their own countries, by contacting enterprises and their professional organisa-tions;
*) awareness activities take place , e.g. among the French Accounting Boards, with press releases and local and national conferences;
*) presentations are given at international conferences (see above)
*) contacts have been made by Eurofer with MECASTAT, a professional association representing 6000 enterprises of the French mechanical engineering industry
*) contacts have been made with software houses by CESIA (Axime-Sligos, GSI-ADP).
This document will be updated on a regular basis. This work package is only in a starting phase; its full development is expected for 1998.
(1) This description is based on several reports from the TELER project
(2) This is only possible for data that are available in the automated accounts; for other data the same tool can be used (adding the missing data by manual data entry) or other like touch tone.
(3) BISE and SISE are acronyms for Base d'Information Statistiques d'Entreprise (Statistical Enterprise Information Basis) and Système d'Information Statistique d'Entreprises (Statistical Enterprise Information System). The concepts of BISE and SISE have been introduced in the SERT study 01 that CESIA conducted on behalf of Eurostat.
(4) EDISENT is an acronym for EDI between Statistics and ENTerprises. The name is used for both the concept and the software module that will be used to transform the data in the enterprises' automated accounts into data that are collected by the NSIs and others.