As our clients know, we love nasty data challenges - otherwise you’d have done it yourselves already !
We’ve got a 4 minute video all about migrations that you will enjoy here…
So why are System implementations/upgrades/migrations one of the most difficult projects to deliver ?
Let us count the ways !
Legacy systems are not well understood and often have difficult data access.
System vendor or decommissioned system has no incentive to help.
Data is very patchy, generally only maintained to support the minimum operations.
A common example is addresses being all on one line to work well in window envelopes, believe it or not !The new platform needs a totally different data layout and lookup values requiring significant re-shaping - as our video says, you need to get a square peg into a round hole.
The migration needs to happen multiple times for different testing cycles.
Data problems are only found right at the end when data doesn’t upload, requiring the whole process to be started again.
Because of this, most migrations quickly compromise by limiting which historical data is migrated, only loading test data or cutting the number of test runs.
This project was typical ;
Legacy platform with limited data connectivity and little technical support.
Large amounts of customer data held in documents in shared drives.
A new SaaS platform being implemented with frequent system changes.
The repeatable migration path was as follows ;
The repeatable testing path was as below;
Due to unavailability of Quantios Core API’s, as shown we had to populate upload spreadsheets which raised its own issues ;
Limits on record counts uploaded meant that we had to create client/account/transactions by batch
Changing requirements meant that only partial data was taken across, with other datasets being stored
Difficult to get access to business experts to clarify and resolve data discrepancies.
Anyway, despite all the challenges the project was successful !