Data Migration or Data Conversion are two terms commonly referred to the task of moving data from the legacy system(s) to the new information system.

Data Conversion is every Project Manager’s nightmare. Technical folks, on the other hand, love it; nothing beats a good challenge.

Data Conversion effort/time/cost is notoriously difficult to estimate without ‘getting into it’. If you are faced with a data conversion project involving older legacy systems, manual data sources, or from different technical formats, running a pilot may be a good option to reduce your risk and help you plan the overall effort.

A pilot conversion has these objectives:

1) to prove the data conversion methodology and assumptions
2) to better understand the current state of the legacy data with respect to it’s ‘cleanliness’ and suitability for migration/conversion
3) to determine which data will be cleaned up prior to conversion and which data will be cleaned up after
4) to better estimate the overall conversion effort
5) to understand the ‘time’ required to carry out the conversion. Some conversions may take several days of computer processing time, espeically spatial data conversions.

A good pilot conversion should contain:

1) a small subset of the overall data source(s)
2) the most comprehensive in terms of covering all legacy data sources and types
3) the dirtiest of the data if possible
4) a complete conversion methodology including a final load into the new system
5) contain enough data to facilitate the testers who are testing the new system

Did I miss something? Have you done something different?

Send feedback to robert@ajae.ca, I’ll print it in the next newsletter!