This post was extensively updated and incorporated into my new book, The Data Conversion Cycle: A guide to migrating transactions and other records for system implementation teams. Now available on Amazon.com in both Kindle format for $4.49 and paperback for $6.99.
Continuing the series on the data conversion cycle. In order to conduct a meaningful risk assessment of a planned data conversion effort, we need to consider the not just the legacy source system and the target, but the remainder of the project. A risk taxonomy identifies and defines relationships between activities and resources as classes, activities, and attributes. The team conducting the risk assessment can then use the taxonomy as a source of questions, to determine what is unknown. In this installment, I’ll outline a risk taxonomy for data conversion that includes sources of risk from the larger project and operating environment, as well as related activities. This holistic view of risk will allow the project team to better identify the unknowns that matter.
For example, the team members charged with defining the requirements for the conversion can use the taxonomy to identify and anticipate unknowns in each of the activities in the Requirements Class. A leading practice is to phrase a question, using the attributes of each activity. “Is the project scope stable enough to let us accurately define what records should be converted?” “Is the target system configuration complete and stable enough for us to map existing records to it?”
The Requirements Class
This class includes the activities that define the requirements for the data to be converted and loaded to the target system.
- Define scope. This is the analysis done to determine what records will be migrated. Attributes include stability of the project scope, completeness of the analysis, and accuracy of the detailed scope description.
- Map data elements. This is the analysis done to determine where each data element in scope will be placed in the target system, and what changes or transformations must be performed on the source data in order for it to be meaningful. Attributes include stability of the target system configuration, correctness of the mappings, and completeness of the requirements.
- Correct invalid mappings. This is the corrective action taken, based on the validation of a load, to update the mappings. Attributes include version control, quality control, and updates to the validation plan for the next load.
- Staff the activities. This is the people management process that assigns team members to perform the tasks. Attributes include availability of functional subject matter experts, source system experts, and target system experts.
The Source Extraction Class
This class includes the activities that produce data records, formatted in a way that will allow them to be loaded to the target system.
- Develop legacy system extraction processes. This is the development of queries, reports, and related artifacts that will select specific transactions, format them appropriately, and make them available for loading to the target system. Attributes include software quality, availability of hardware and network resources, and performance.
- Extract records from the legacy system. This activity consists of executing the extraction processes from a snapshot of production and making the resulting data available to the team members who will load the target system. Attributes include access to the system containing the snapshot, sequencing, performance, defect handling, and secure file transfer.
- Correct invalid processes. This is corrective action taken, based on the validation of the load, to update the extraction, re-formatting, and file transfer processes that produce the data and make it available to the Load team. Attributes include version control, quality control, and updates to the validation plan for the next load.
- Provide resources. This is the management activity that assigns team members and provides them with access to server resources required to perform their tasks. Attributes include availability of operations team members and developers, availability of disk space, program access to the source system, and availability of secure file storage.
The Load and Validate Class
This class includes the activities required to load the records extracted from the source to the target system, validate that they were loaded correctly, and initiate corrective actions.
- Load data records to the target system. This activity consists of executing the target system load processes on the data records provided by the Source extraction team. Attributes include software quality, extracted data quality, availability of hardware and network resources, and performance.
- Validate loaded records. This is an inspection activity, designed to ensure that the data stored in the source system has been properly moved to the target system, and is ready to support planned operations in production. Attributes include coverage, accuracy, tools, and access to the target system.
- Correct invalid records in production. This is a records management activity. Attributes include accuracy of historical records, quality of the validation process, and access to the production system.
- Provide resources. This is the management activity that assigns team members cleared to access the data, and provides them with access to the loaded system required to perform their tasks. Attributes include availability of qualified team members, security controls, and access to the loaded target system.
The Project and Environmental Factors Class
This class includes those project activities outside of the data conversion cycle processes that influence or drive the activities listed above.
- Manage project scope. This is a project management activity. Attributes include change control, configuration management, and communication.
- Manage system design. This is a target system development activity. The design of the system is a primary input to the conversion scope and mapping activities. Attributes include completeness, validity, and maintainability.
- Manage implementation. This is a target system development activity. The configuration of the system greatly influences the mapping and transformation of the source data to make it usable by the target system. Attributes include completeness, stability, modularity, and traceability of changes.
- Manage Operations. This is a data center activity. Attributes include production system management, operating schedule, and network access.
Plainly, this list is just a starting point. There are a number of potential risks in these classes, but based on the environment, the scope of the project and the state of development, there may be many more. The project team should review the list, asking questions as they go, and identifying the unknowns as issues or potential risks requiring qualitative analysis. Next week, I’ll conclude this series with a look at how we can measure progress in the iterative data conversion cycle.