Repeatability, Predictability, Timeliness
Many factors go into successful data migration, the main one being good communication with the client, this is encapsulated in our BIDM methodology. However, from a more technical point of view, we find three things that stand out in making the project a success.
Repeatability
Utilising proprietary data manipulation tools, automation makes the whole process of data migration repeatable and predictable. With the strict use of
- Automated source and target system refreshes
- Data extracts
- Script execution
- Data quality checking, tagging and loading
- Auditing and reporting
The data migration process can be run end-to-end as often as is practical. In a typical data migration, data objects will be repeatedly run through this process until the results are correct, and predictable every time. When working on critical systems with large data, repeatable automated processes are the only way to go.
Predictability
To ‘guarantee’ the result of the final data migration, the main cut-over and system go-live deployment, the expected result should be known beforehand. This concept is extremely difficult to obtain, given
- The ever-changing nature of the data
- Unknown data quality errors introduced just prior to the final migration
- Target system changes
- Migration requirement changes
- Changes in project resources
At DataMC our whole methodology is geared towards building a predictable outcome for the business during the final data migration, the systems go-live event. The ‘repeatability’ of the built data migration solution is there to support multiple end-to-end migration runs, to enable all likely scenarios to be encountered before the go-live migration.
Timeliness – the Runsheet
As much as having a ‘repeatable’ and ‘predictable’ data migration is critical to achieving a successful data migration, so too is the timeframe in which the go-live migration will take. This is only achievable with a thorough and accurate runsheet listing:
- The order of all tasks to be completed
- The times taken for each task
All aspects of a data migration should be known before a go-live event is initiated. Under no circumstance should this be attempted without an accurate and comprehensive fully scheduled runsheet being developed. With a runsheet, all timeframes should be known and achievable, without it, the outcome and success of a project go-live can not be predicted in advance.
Explore the DM REVOLVE user interface
