Implementing a new retail system is difficult. A new system can result in many great benefits; improved business processes, access to more and better data, improved analysis, more intuitive results, etc. A new system brings hope and excitement for the future. Regardless of how great the new system is (will be), it cannot be successfully implemented without proper and fully converted data. Retail systems have a lot of complex data and converting it is no simple task. By following these 8 simple steps, you will significantly improve your chance of a successful conversion, with little or no disruption to the current process.

8 steps to a successful data conversion:

  1. Define Scope - Scope must be clearly defined up front. What is being converted and what is not? How much history is being converted? What can be done manually rather than automated? It may not be possible to convert ‘everything’. Scope is critical and the project’s success will be dependent on having this clearly decided upon.
  2. Identify data sources Where is the data coming from? If you have multiple systems or data sources, each data source must be clearly identified.
  3. Identify data destinations (APIs) Where is the data going? You must identify the source systems or databases and the appropriate APIs to get the data in there.
  4. Define requirements/mappings - This is what I believe to be the most critical step. Detailed requirements must be defined for each element within each conversion; what data mapping is required? What business rules must be applied? What data transformations need to occur? Each and every data element must be scrutinized and documented.
  5. Define Estimates – Assign appropriate estimate of effort to the development of each conversion
  6. Define Scope of Test – Scope of the testing must be defined. It may not be viable to convert ALL data as a test or it might be done in a number of steps – sample, small dataset, and then larger or full dataset. All of this must be defined and included in the overall plan.
  7. Who and how to validate – How is this all going to be validated? What results are expected and who in the organization is going to validate the results? Key users/sponsors must validate and feel confident with results.
  8. Develop Project Plan/Assign Resources – And lastly, build the project plan for all development, testing and validating tasks; identify the sequence and dependencies of all tasks; identify and assign the appropriate resources to each task on the plan.

Implementing any new system relies on a successful data conversion. Following these 8 steps will lead to a complete successful data conversion. What do you think? Is there anything else you would caution about or add to the list?