The riskiest part of a development comes in one of two areas, the incorporation with other systems, (particularly where you have near-real time transactions) and most gravely, the ERP in sequence relocation. You can build out your processes in the system correctly, you can ensure that all the reports are spot on and you can appropriately train your users. However, if your ERP Data migration does not implement perfectly, then you are going to have unhappy users and a failed implementation. Since this is a critical part of any implementation, let’s look at some key data concepts and areas to watch for to ensure that you have a good understanding of the migration process.
The key concepts you first need to be familiar with are the two types of data that you and your team will be commerce with. First, there are the Master Data records. This is the data that unusually changes and is the core of the system records. Records such as Companies (Vendors and Customers), Contacts, Inventory Items, Bills of Materials, and GL Accounts are all Master Data. All other Transactional Data relies on the Master Data.
The second type of records is the Transactional Data. Orders, Purchases, Work Orders and the like are your Transactional Data. These are your day-to-day transactions. As mentioned, these utilize (or link to) the Master Data to create a whole record. It should be understandable that you want to ensure that the Master Data is clean and weighed down first before you load the Transactional Data.
The key to success is proper planning. Before you go on board on data immigration it is important that you catalog all of your data sources that you will wander. These should be part of a complete data plan document. Key things that should be included in your Data Plan are:
1. Mapping validation
2. Match reports and record count validation
3. Correct any discrepancies identified in the validation process
When planning your ERP Data migration, some things you should consider:
How much data do you need to bring over? Characteristically, you will bring all of your master data, but possibly only two years of transactional data. This could vary on a table-by-table basis. For example, you may want to bring over five years of customer orders, but only two years of purchase order data.
In summary, it is critical that you have a really well-thought-out plan that addresses not only the migration but also built-in contingency. Is there a rollback plan? Then, once you are in the migration, be sure to run a significant amount of tests on the data to ensure it is valid data, it is properly mapped to the correct target fields, and that the sequencing of the data is carefully considered. Once tested, make sure you have a go-live validation to ensure that even the well-tested data came into your construction environment correctly.