Data Warehouse for PPI The implementation of a data warehouse for PTT Post International - The Netherlands Tadek R. Miller PTT Post International The Netherlands Date: April 1998 Version: Final Author: Tadek Miller PTT Post International BV
Contents 1. Abstract...Fehler! Textmarke nicht definiert. 2. Organisation... 2 2.1. The TNT Post Group... 3 2.2. PTT Post International... 3 3. The business processes... 4 3.1. Primary processes...4 3.1.1. Intake... 4 3.1.2. Sorting (handling)... 4 3.1.3. Transportation... 4 3.1.4. Distribution... 4 3.2. Secondary processes... 5 3.2.1. Accruing... 5 3.2.2. Invoicing... 5 3.2.3. Accounting... 5 4. PPI business control mechanism... 5 5. PPI systems architecture... 5 6. Systems development and data management... 6 7. Data Warehousing at PPI... 7 8. Conclusions... 8 9. References... 9 1. 10. How to contact the author... 9 1
Abstract PTT Post International (PPI) is part of the Dutch PTT Post organisation and deals with the collection, sorting, and distribution of international mail. These activities are done in cooperation with other postal operators in the world. PTT Post recently acquired the Australian TNT company that specialises in time sensitive distribution and logistics of corporate mail. International mail is handled and remunerated in accordance with international regulations and multilateral or bilateral agreements. In order to determine our turnover and to remunerate, all incoming, outgoing, and transiting mail is registered. To accomplish this a number of information systems were developed. Within systems development reporting is a special issue. Wherever possible SAS software is used to generate management information. PPI is in the middle of setting up a data warehouse based on a mixture of legacy systems and newly developed systems. Currently, we use the data warehouse especially for analysis purposes like performance and quality control. In the near future we will also use the data warehouse to do analysis on product-market combinations. In this paper the position of the data warehouse within PPI's systems architecture and data management, and the relevant information systems in use with PPI will be discussed. The focus will be on how to organise, use, and maintain the data warehouse within our organisation. 2. Organisation Until recently, PTT Post was part of KPN (Royal PTT Nederland NV); the holding company which is listed on the stock exchanges of Amsterdam, New York, London, and Frankfurt. KPN consists of two divisions: PTT Post and Telecom. Last year PTT Post acquired TNT. TNT is an Australian company specialised in just-in-time delivery, logistics, and handling and distribution of corporate mail worldwide. The merger of PTT Post and TNT resulted in a new company: TNT Post Group (TPG). In August 1998, KPN will be split in two independent companies: TPG and KPN Telecom. Both companies will be listed on various stock exchanges. 2
2.1. The TNT Post Group TNT Post Group (TPG) Post TNT International Mail TNT International Express TNT North America etc. Letters International etc. Figure A - The organisation chart of TPG In Figure 1 the TPG organisation chart is shown. TPG is the name of the new holding company of PTT Post and TNT. 2.2. PTT Post International In Figure 2 the organisation chart of PPI is drawn. PTT Post International Information Technology International Relations Finance Operational Support Marketing/Sales Operational Units Figure B - The organisation chart of PPI 3
3. The business processes In general two types of processes can be distinguished within PTT Post International: The primary and secondary processes. The primary processes are dealing with the intake, sorting, transportation, and distribution of incoming and outgoing international mail from and to all countries in the world. The secondary processes are the administrative processes such as accruing, transportation billing, invoicing, and accounting. 3.1. Primary processes 3.1.1. Intake International mail is handled in two large handling units, one in Amsterdam (mostly letters) and one in Rotterdam (parcels). The collection of international mail within the Netherlands (mail from mail boxes, post offices etc.) is done by the business unit letters. They deliver the international mail to the handling units in Amsterdam or Rotterdam. In the handling units the mail is registered and sorted. PPI has separate contracts with organisations like publishers that ship regularly large amounts of mail and/or parcels. These customers often deliver their international mail directly to the handling units. Inbound mail from other countries arriving by plain, boat, or truck is registered at the front entrance of the handling unit. 3.1.2. Sorting (handling) 3.1.3. Transportation Depending on the type of outgoing mail, it will be sorted according to international regulations (e.g. UPU - Union Postale Universelle) and according to bilateral or multilateral agreements (e.g. the REIMS agreement). Incoming mail from other countries is delivered to the business unit letters for further distribution within The Netherlands. Mail from outside the European Community has to pass the customs first. Except for large parcels, outgoing mail is bagged and tagged with labels containing information on the contents, service level, type of mail, weight, etc. 3.1.4. Distribution Depending on the transit quality required by the customer, international mail is shipped to other countries either by air, truck, or boat. Transportation is not carried out by PPI itself but is contracted out to transportation companies like KLM Royal Dutch Airlines. Therefore, PPI contracts volume space with various transportation companies for longer periods of time. Within other countries, the distribution of international mail coming from The Netherlands is usually done by local postal services. However, more and more private distribution companies are contracted by PPI to distribute the mail in other countries. 4
3.2. Secondary processes 3.2.1. Accruing All receiving countries register their incoming international mail. The receiving country bills the sending country for all distribution costs, according to international regulations, governed by the UPU or other agreements, and based on weight, quantity, mail type, and/or service level. 3.2.2. Invoicing Based on the registration of mail during the primary processes, PPI either makes invoices for other countries or checks the invoices received from other countries. Based on the same information, invoices from transportation companies for distribution services rendered are also checked. 3.2.3. Accounting Using the invoicing system, customers are billed according to their contracts with PPI and the amounts of international mail shipped. Accounting is done based on accruing and other cash flow, just like in most other companies. 4. PPI business control mechanism For budgetary control, operational units such as the handling units in Amsterdam and Rotterdam, are regarded as cost centres. This means that they are responsible for working against low cost prices. The sales units within PPI are responsible for profit making. All costs made by the operational unit are charged to the sales unit. Therefore, the operational units need to know which sales unit is responsible for each customer or product. On the other hand the sales units need to know the cost prices of handling, transportation, and distribution. Products and costprices are determined and maintained in the production portfolio per handling unit. 5. PPI systems architecture The architecture of the information systems at PPI is drawn in Figure 3. 5
'$7$:$5(+286( $FFRXQWLQJ */ 2SHUDWLRQDO $JUHHPHQWV $FFUXLQJ 2$6 5287,1* WUDQVSRUW FRVWV GLVWULEXWLRQ FRVWV 2(6 206 LQWDNH KDQGOLQJ WUDQVSRUW 'LVWULEXWLRQ 2SHUDWLRQV Figure C - Overview of information system at PPI These are not the only systems within PPI; there are other information systems like the sales supporting system and other financial systems. However, the systems in Figure 3 can be regarded as the back bone for PPI. The operational and financial systems in Figure 3 are currently being re-designed. This large project is carried out under the responsibility of the Information Technology department of PPI. 6. Systems development and data management Within PPI, information systems are developed using Uniface as development tool, ORACLE as RDBMS, and the SAS system as reporting tool. The Data Management group within PPI's Information Technology department, maintains all data models and assures the interface between newly developed systems as well as with legacy systems. Within PPI conceptual data models are not maintained. The combination of Uniface, ORACLE, and the data management tool ORACLE Designer, makes it possible to maintain the physical databases at low cost. 6
7. Data Warehousing at PPI The data warehouse at PPI is loaded with information from the operational databases of the different information systems in use by PPI. In addition we include vital information from outside the company in the data warehouse. For example: Price Waterhouse collects data on worldwide mail transit times. This data is used by many postal organisations in the world to gain insight in the quality of their postal services. In the near future the data will actually be the basis for remuneration between countries, resulting in financial advantages for postal organisations that perform well in terms of short transit times. The data gathered by Price Waterhouse is presented to the International Postal Council (IPC) in Brussels. PPI receives the data in a summary of this information. In addition to the measurements done by Price Waterhouse, the Dutch marketing company Intomart measures for PPI the main mail flows. Intomart uses advanced devices like RTF (Radio Frequency Tagging) and quiver chips to collect the data. The data from Price Waterhouse and Intomart is imported into the data warehouse. Based on the data in the data warehouse, managers produce reports on the quality of the services delivered by PPI. In addition the results are compared with those of the competitors. The data in the data warehouse is further analysed in order to get a better view on the business processes and to determine bottlenecks in the operation. In other words the data warehouse is used as a management tool to optimise the production processes. As the example above clearly demonstrates, information on business processes is critical for PPI. However, users of information have one thing in common: today they do not know what information they need tomorrow. At PPI we make distinction between information represented by reports and information gathered by data analysis. Information on reports have the tendency to be steady over a longer period of time while the information as the result of data analysis often varies from day to day and from question to question. Therefore, within PPI we were looking for a tool that was able to deal with both types of information. At PPI the data warehouse is considered as an operational system. In other words the data warehouse is treated like any other system, including functional control. We have appointed a special data warehouse administrator, who s task it is to keep the data warehouse up to date, to help users with questions, to adapt reports where necessary, to expand the data warehouse, and to keep users and management informed about new developments in the data warehouse. The corporate Intranet at PPI is used to inform users about the the contents of the data warehouse. Potential users find information on data warehousing in general and how they can become a user. The interaction between users and the data warehouse administrator is important and of special concern. Once users discover the possibilities of a data warehouse, they become hungry for more and more information. As the data warehouse is constructed very flexible, it is easy to add, for example, additional variables to the subjects in the data warehouse. However, there are limits to the costs involved in the maintenaince of the data warehouse. The question arises how the cost of maintenaince is controlled. As incidated before, the 7
data warehouse is treated like any other operational system. This means that requests of users to change report layouts, adding additional variables, etc., are handled as change requests. As change requests put pressure on the managers budget they will only request changes or additions when there is a clear return on investment. This procedure ensures a high quality data warehouse with useful data and a small budget for maintenance. The users, together with the data warehouse developers and the responsible managers, decide on the construction and contents of the subjects in the data warehouse. We attempt to design the data warehouse according to the needs of the users and not according to what technically is possible. The design processes are done iteratively. At PPI, the data warehouse developers are involved in the development of new information systems at an early stage. Having the developers of the data warehouse and the information systems work closely together, improves the quality of the information household within PPI as a whole. The tool that managers use to access the wealth of information in the data warehouse is of utmost importance. Several vendors claim they have the ultimate end user tool in order to make queries and reports. They all have certain benefits but also disadvantages. PPI chose to work with CIDS (Company Information Delivery System), an application developed in SAS/AF by PW Consulting. CIDS consists of two modules: CIDS Data Warehouse Manager and CIDS Reporting & Analysis. The CIDS Data Warehouse Manager is a tool to construct and maintain an enterprise data warehouse and CIDS Reporting & Analysis is used by managers to analyse data and to generate reports. In the paper of Schipper (1998) both CIDS modules are described in more detail. Presently, within PPI several data warehouse user groups can be distinguished. In the marketing and sales department CIDS Reporting & Analysis is used to analyse customer data. Managers of the operations department use it for quality control whereas the controllers of the financial department use it for extracting financial information and internal billing. Using CIDS Reporting & Analysis enables us to standardise report layouts. To get users acquited with CIDS, new users participate in a two-day hands-on workshop where they learn how to substract and analyse the data in a data warehouse. In addition, the data warehouse administrator is always available to assist managers with question that might arise. This paper dealt with the view of the business on the data warehouse. In Schipper's (1998) paper, the construction of the data warehouse is described in more detail. 8. Conclusions A data warehouse is a critical system for a business. Important decisions are made based on the information gathered in the data warehouse. Though the contents of a data warehouse is static, the construction is dynamic and done in co-operation with users and management. Growth is needed, but never without a good architecture. Controlled growth using change requests is recommended. Also, continuous interaction between systems developers, data warehouse developers, and users ensure a tuned operational system and satisfied managers. 8
9. References Schipper, H., 1998, Using CIDS, a SAS/AF application, to build a data warehouse and to provide OLAP tools for managers at PTT Post International. Proceedings of SEUGI # 16 at Praag. 10. How to contact the author Tadek R. Miller, Manager Data Management PTT Post International P.O. box 1996 2121 XX Hoofddorp The Netherlands Tel.: +31 23 567 51 38 Fax.: +31 23 567 51 50 E-mail: t.r.miller@ptt-post.nl Web: http://www.tntpost-groep.nl/ SAS and SAS/AF are registered trademarks of SAS Institute Inc., Cary, NC, USA. Uniface is a registered trademark of Compuware Corporation. Oracle and Oracle Designer are trademarks of ORACLE Corporation. Other brand and product names are registered trademarks or trademarks of their respective companies. 9