GISdevelopment.net ---> GITA 1997 ---> Project planning, implementation and management

Managing conversion control centers

Martin Ridley
British Telecommunications Plc
The Angel Centre, 403 St. Johns Street
London ECIV 4PL, UK

Simon Mapes, David Wren
Brown & Root EPR Conversion Control Centre
54-60 South Street, Dorking
Surrey RH42HQ, UK


Introduction
British Telecommunications plc (BT) provides communication services in the U.K. (England, Wales, Scotland and Northern Ireland). Privatised in 1984, BT has embarked on many efficiency initiatives. The company has significantly reduced staff over the past ten years and is continuing to implement new systems and procedures aimed at maintaining a leading position in a highly competitive marketplace.

The External Planning and Recording project (EPR) is aimed at automating the design and recording of external network records. BT has a total of 35 million lines in approximately 5,600 exchanges (wire centers). Manual records are maintained in 55 regional drawing offices, and consist of approximately 200,000 geographic maps, 400,000 duct and cable diagrams, 2.5 million duct records, and numerous other source documents. Drafting standards, while company-wide, have been adapted to suit regional variations, especially in dense city centre areas such as London, Manchester, Birmingham and Glasgow.

The EPR project & conversion iSSUES
BT investigated AM/FM systems during the 1980’s, and identified data conversion as a critical component of the EPR project. Extensive discussions with other companies in various stages of system implementation, and with conversion vendors showed that conversion could negatively impact the project in two ways - the cost of conversion and the risk of failure.
  1. Conversion Cost
    Forming a major part of the total project budget and schedule, conversion cost has a significant impact on the cosffbenefit analysis, to the point that it may prevent justification of an AM/FM project. An earlier BT project based heavily on a graphics-oriented AM/FM solution was analysed but the total conversion costs (estimated to be ‘well over El 00 million’) rendered the project unfeasible.


  2. Conversion Cost Reduction
    The design of the EPR system was influenced by the conversion cost issue, with a database-driven, minimal graphics, orientation. BT’s national and local requirements in the U.K. service area were addressed with a multi-tiered application. The dense and more active areas required full functionality with full data conversion, while the rural low-activity areas required less functionality and a correspondingly lower implementation cost. BT also carefully considered and evaluated the use of off-shore conversion facilities for their potential cost savings. BT were convinced that the use of such facilities would be both socially and politically acceptable and were therefore included in the project execution strategy.


  3. Conversion Risk of Failure
    Learning from the experiences of others, it was apparent that data conversion was often treated as a project add-on that frequently failed primarily due to lack of adequate planning and project management experience. Conversion specifications needed to be detailed and accurate, and skilled resources needed to be in place to cope with the demands of the invariably tight conversion schedule. Project management was identified as a critical requirement for the EPR project.


  4. Conversion Risk Reduction
    Extensive qualification of suitable data conversion vendors was carried out. The capability and experience of each vendor was assessed thoroughly before tenders were issued and contracts placed. To establish consistency in the data conversion, conversion specifications were developed that included a detailed source record analysis and correlation to a new data model. The data conversion support activities were also identified and quantified as follows:

    • collection of source documents from the 55 drawing offices around the UK;
    • identification of local recording practices for each drawing office, addressed as specific addenda to the core conversion specification;
    • analysis and resolution of problems arising during the conversion process;
    • checking of converted digital data to an acceptable quality level.

    While it seemed to be universally accepted that these activities should be carried out internally, there were several drivers that led BT to contract the EPR Conversion Control Centre (CCC), the conversion support activities, and the management of the seven data conversion vendors to an external organisation. The factors that led to this decision were as follows:

    • corporate downsizing made it important to retain internal expertise, gained during the implementation, throughout the continued operation of the EPR system;
    • BT EPR project staff would potentially lose their current career-path positions, and would have been unhappy at being tasked with shorter-term roles in data conversion;
    • BT staff were not experienced in data conversion, therefore project management and production management skills were not readily available;
    • BT team members were to focus on the EPR application software and specifications, while the shorter-term conversion tasks were to be contracted externally.

    The EPR project also has major features which necessitated a uniquely tailored solution for BT:

    • complex conversion process with variation in pradlces between 55 drawing offices;
    • need for a high level of accuracy and completeness of work in data conversion;
    • wide variety of interested parties both from within BT and from external contractors, all of whom were to be involved in key aspects of the conversion process;
    • world-wide geographical spread of those involved in related aspects of the conversion work;
    • large volume of original source records and data to be processed;
    • demanding project constraints (project finance, schedule and quality).

    The decision to outsource the conversion management and conversion support activities has proven to be a significant factor in the successful progress made on the EPR project to date.
Management and technical challenges at the control center
Taking into account the project objectives, the necessaty technical processes and component activities of the conversion control centre were defined:
  • document management mocesses associated with collection, control and dissemination of paper based records in electronic format.
  • conversion suDDort and helDdesk activities required to support data conversion vendors, specifically in conversion specification management and query response.
  • data acceptance Processes required to ensure converted data meets BT’s requirements for accuracy and completeness, including sample data testing and acceptance statistics.
  • proiect control txocesses implemented for monitoring progress and planning ahead to mitigate the impact of change on project costs and schedule.
  • qauality management processes required to ensure a pragmatic approach, including quality assurance, quality control and a philosophy of continuous improvement.
Each of these activities are fundamentally linked in the conversion process. In order to manage the complexity and interfaces between the above processes, a management philosophy has been adopted at the control centre to treat the overall conversion process as a production line, with converted data being the finished ‘product’.

The key objectives for the overall conversion process or ‘production line’ are a controlled level of data throughput, acceptable quality levels of converted data and the control of project costs to BT. It was recognised that project priorities could change over time, and that technical innovation would be required to manage this change throughout the life of the data conversion. This has been achieved throuah a hiah level of {ethnical and management skill, allied to an appropriate use of technology and automation.



Scanned images or paper copies?
Commercial sensitivity and ongoing access to original records has determined that only copies can be removed from BT drawing office locations. The requirement for minimum disruption to the continued network planning task highlighted the need for an effective document management solution.

A major element in the technical approach of the EPR Conversion Control Centre lies in the use of scanned electronic images of original source records, a technology known as Document Image Processing or DIP, in preference to more traditional paper copy or microfilm based solutions.

A typical paper based solution for data conversion would have required multiple copies of source records to be taken and stored prior to shipping to a selected data conversion vendor. Source records or copies would also need to be retrieved manually from the master library for use in the conversion support and data acceptance processes. The costs involved in a paper based approach are centred largely on:
  • extra staff required to handle and control the additional document movement throughout the conversion project in relation to standard drawing office practices;
  • office space required for storage of and access to the source records and copies;
  • additional costs in creating paper copies at the drawing offices and transporting them to the chosen data conversion vendors.
A DIP system was proposed as a key technology for the EPR conversion to reduce the life costs of document management and to maintain accuracy of document identification. The DIP system option involves the use of portable equipment to capture source records electronically at each of the 55 regional drawing offices (equivalent to a total storage of over 500 drawing cabinets). The optical image based capability in the conversion control centre involves electronic retrieval and plotting of scanned documents through a workstation network in suppoil of the conversion support and data acceptance processes. The DIP system option offers the following significant advantages over the use of simple paper based photocopies:
  • document management is simplified considerably, reducing effort by avoiding the need for manual extraction and re-filing of paper records;
  • the movement of copied documents is simplified, thereby reducing transportation costs;
  • conversion support and data acceptance processes are streamlined by simultaneous read only access to all scanned images;
  • all scanned images are first generation copies, thereby avoiding quality degradation;
  • network security is provided against loss of individual documents, with a readily established audit trail of access to documents;
  • electronic storage of images reduces space requirements significantly, with lower long term costs of archiving.
Mainstay processes of conversion control
  1. Record Retrieval
    A site survey at each record drawing office is undertaken to assess any physical constraints plus an assessment the location and condition of the source documents. This information is used to plan an accurate scanning schedule in order to minimise disruption to BT drawing offices, as well as providing a required scope of work to each scanning team. Source records are collected by four well-drilled mobile scanning teams, comprising four team members (with knowledge of BT drawing office practices), each team capturing more than 5000 documents per week.

    Quality control of the captured images and electronic cross references are also performed in the field. The images for complete telephone exchanges are transferred to data tapes and are dispatched to the conversion control centre. Upon arrival, data tapes containing the scanned images are loaded onto the DIP system optical disk jukebox. The DIP system can access and display any one of over half a million documents held on-line (via standard networked PCs) within an average of 15 seconds.

    This capability provides significant benefits to all subsequent processes in support of the data conversion. Prior to dispatching exchange data to the conversion vendors, a second image quality check is undertaken to ensure the best possible data is provided to the vendor for conversion.


  2. Conversion Support
    The aim of conversion suppofl is to provide the data conversion vendors with a help desk facility, manned by a team with detailed knowledge of BT drawings and drawing office practices. The team answer conversion related questions from the conversion vendors, thus relieving this burden from BT drawing office staff. This also compliments the role of the conversion control centre as a project communication hub.

    The main stay process of conversion support is to deal with questions, known as Conversion Exception Reports (CERS). To ensure a fast turn around of results, each data conversion vendor is allocated a dedicated fax line within the control centre. A CER tracking system together with the DIP system, ensures that all CERS are recorded and have a response within a 48 hour service level agreement. Statistics from the CER tracking system help to identify common questions that may be best addressed by an addendum to the conversion specifications.

    Conversion specifications are the corner stone of any data conversion project. To ensure that every interested party focuses on what is best for the project, process change requests (PCRS) are raised by all parties involved in the project. In the early stages of the project the majority of PCRS resulted from the conversion process. Conversion control centre staff also produce PCRS as a result of findings made during data checking operations. The conversion support team co-ordinate the collection of PCRS which are then passed onto BT technical staff to create specification addenda.


  3. Data Delivery
    Data delivery processes are focused on post conversion quality control checks and the upgrading of data to the latest available version of the EPR system. In an ideal world, checking of convetied data would require as much effort as the conversion work itself. Available schedule, budget and resources necessitate an alternative approach based on a sampling methodology (I SO 2589-1 / BS 6001 ).

    The checking process was designed to be consistent, fair and auditable, and to meet the statutory requirements of the New Roads & Street Works legislation (UK 1991). It was recognised that a perfect conversion was not realistic, therefore a series of Acceptable Quality Levels (AQLs) were established. The checking team all have an intimate knowledge of BT drawings and drawing office practices. Each team member draws upon this knowledge to advise data conversion vendors on discrepancies that have been detected in conversion. Detailed reports and statistical results are generated for every submission showing vendor performance in relation to the required acceptable quality levels. From date of receipt, a service level agreement ensures that exchange data is checked and a result returned within 45 days.

    The economics of standardisation and change control of data sources dictate that data conversion and system development activities are run in parallel. Vendors are therefore required to convetl data to be compatible with software that is several versions behind the current system. Converted data that passes the acceptable quality level is subjected to a data upgrade process that is linked to ongoing system development. The data upgrade scripts define the changes between software versions thereby enabling checked data to be compatible with the latest database schema. In addition to data upgrades, a series of technical end user reports are generated describing the checks undertaken. This doument also notifies the end users of residual errors that require post-conversion rectification.
Support processes
A number of supporting processes play a key role in meeting the technical challenges presented in the conversion management task.

An open style of project management empowers the process owners within the control centre with a high level of decision making ability. This has been a major success in all areas at the centre and enables the project manager to act as both leader and mentor, assisting project members as well as providing ‘thinking space’ for activities such as fomard planning and development.

From the perspective of document management and provision of an audit trail, both paper-based and electronic documentation arriving and leaving the conversion control centre is recorded by document controllers. Despite the fact that one of the benefits of EPR is to reduce paper, the conversion process itself is heavily reliant upon paper primarily due to the nature of the source documents.

One aim of EPR is to standardise record keeping. Source record analysis of local custom and practice eliminated many clarification issues that could have been raised during data conversion. It was recognised that regional variations exist throughout the source documents held in each of the 55 regional drawing offices. These variations have been fed back into the conversion process via addenda to the conversion specifications. A by product of this analysis is the creation of a document that addresses typical end user questions such as “What will our data look like in the EPR system?”

The cost of a poor conversion specification is, at minimum, the cost of data conversion plus the cost of rectification. This high value warranted a dedicated team to administer and control management of technical specifications. This is no trivial task. There have been in excess of 500 specification addenda plus two major base line upgrades consolidating these addenda into new versions of the master specification documents. Specification management is a continuous task as the conversion process runs in parallel with system development. This poses significant change management issues which can only be controlled successfully if strong project management practices and sound administrative processes are in place.

To be able to run a project as a production line process, it is imperative that processes and work instructions are developed in a quality assured and contolled manner and are subjected to regular audit. The conversion control centre has produced a Project Procedures Manual currently containing over 2500 separate auditable tasks developed to ISO 9001 standards. A continuous improvement programme empowers and encourages all project staff to challenge processes with the aim of refining control centre operations.

Scheduling and planning of EPR data conversion are both major tasks. In terms of a production line, the whole conversion process can only move as fast as the slowest process. The currency of data in the production line is time critical as BT documents are in are in a state of continual use and amendment. The timing of events to avoid delays is therefore critical in keeping control of costs, schedule and the posting of backlog.

Project structure
Flexible staffing is necessary in the conversion control centre to manage available resources to tackle fluctuating workloads in each process. Mobile scanning teams maintain the throughput of document collection required to meet the planned conversion schedule. The makeup and structure of the control centre team demonstrate this flexibility.

Team-based work allows activities to be carried out in parallel with minimum requirement for extra skilled staff. Peaks of activity in the key processes are resourced by cross training of individuals involved in conversion support and data acceptance. All staff are actively encouraged to discuss potential challenges in order to provide the best overall solution for the project. A single location for the conversion control centre allows movement of staff with common skills between tasks.



The management approach is based primarily on the need to contain project costs, however a number of further opportunities for containment of costs and reduction of risks have been identified :
  • optimisation of the sequence in which exchange data is to be processed;
  • progressive reduction in the effort of checking of converted data, by close monitoring of detected error rates and quality assurance of the checking process;
  • redeployment of skilled staff as resource requirements for activities change through the conversion project life cycle;
  • programme of continuous improvement to streamline processes and procedures;
  • ongoing refinement of automated systems to improve control and work efficiency.
Critical success factors for a conversion control center
Each data conversion project is unique and each provides a fresh challenge to both the utility company and the conversion specialist. As most information held by a utility is unique to that organisation, it is unlikely that a previously successful conversion solution to a different project, even from the same industry, will provide exactly the correct solution for you. The lessons learned on conversion projects, including the successful and the not so successful, have been equally valuable in determining the following critical success factors for establishing a conversion control centre.
  • Data conversion work should be treated as a separately resourced project and not an extension to current working practices. Do not assume that the skills necessary for data conversion exist in your current organisation. Learn what you can from the experiences of others and utilise their skills where possible.
  • Risks that are not managed or understood can prove costly. Increase your probability of success by transferring evaluated project risk to those who can best manage the specific risk.
  • Do not underestimate the total cost of converting your data. Experience has shown that 60-707. of ~ project costs can be expended solely on the conversion of data. Conversion costs can be reduced through simplifying your data model in less critical areas of your network.
  • Plan to run a pilot project that includes the complete life cycle of conversion from retrieval of records through to implementation of converted data onto the target system. Evaluate the reengineered process and be prepared to review any lessons learned to refine the process.
  • Establish a dedicated management team with appropriate skills in project management, systems development and operational experience. Do not treat conversion management as a side line task of an existing manager.
  • Manage data conversion activities, from record retrieval to data delivery, as separate processes focused on delivery of converted data to the target system. Understand the interfaces between each process (similar to a manufacturing production line) and identify the bottlenecks to optimise the total conversion process. Delays in the schedule can kill a conversion programme.
  • Maintain strict control of source records throughout the entire conversion activity. Commercial sensitivity and the need for continual access to documents may preclude the use of original records outside the drawing office. If you intend to use conversion vendors, assess the cost benefits of using original documents versus scanned images versus paper copies as the reference document for the vendor.
  • Where possible, achieve a stable and functionally complete target system. Otherwise, ensure that your system development programme and objectives are well understood by all involved. Your ability to manage ‘scope creep’ will be the difference between success and failure.
  • Establish a set of baseline conversion specifications. Maintain strict control of the specifications and any subsequent addenda. Failure to do so can result in an exponential increase in costs of both data conversion and system development.
The way ahead for EPR
The future is very encouraging for BT and for the EPR project. Following the successful completion of an operational pilot study, the next project phase involves the conversion of 850 telephone exchanges representing 70?40of BT’snationalcustomerbase. Technologywillplaya majorrolein meeting these challenges. One area of potential gain is an extension to electronic exchange of information and data, with the benefits of reduced costs and faster turnaround times. The current DIP system (optical disc technology) has an on-line storage capacity of 40 gigabytes. Current projections indicate a mass storage capacity requirement in excess of 120 gigabytes. Networked RAID server technology (Reduced Array of Inexpensive Discs) is one potential solution providing larger on-line capacity and faster access times.

These advances will be complemented by a continuous improvement philosophy that encourages individuals to challenge and question current working practices, and continues to devolve decision making down through the project organisation.
© GISdevelopment.net. All rights reserved.