By Neil McGovern

Very little about implementing enterprise resource planning is easy. But there are at least a few ways to make the most challenging tasks a bit more manageable. One of the many demanding requirements involved in ERP implementation is converting data from legacy applications to the new client/server application. When you define the scope of your data conversion, you help improve your approach.

Define the scope of your data conversion
Data conversion is target-centric; in other words, only the data required to make the new application operational is converted. This is not always the original intention, but as time pressure builds on a conversion, the optional parts tend to be stripped out.

To decide what data needs to be brought over to the system being implemented, you need to understand your business processes, both old and new. A comparison between the two will help you determine what data the new system needs to import from the old one. Typically, data will be brought into the new application from more than one legacy system. You need to identify all relevant data sources, document the systems on which they run, and make sure you know who is responsible for maintaining and updating the data.

Questions to ask
You can now start to scope your data conversion in more detail, paying attention to questions such as:

  • How much data will you need to convert?
  • How complex is the process likely to be?
  • How much work will be involved in cleaning the data? (Typically, much of the data in an older legacy system will be redundant, inaccurate, or corrupt.)
  • How should the data be extracted?
  • How should the data be converted, transformed, or formatted?
  • How should the data be loaded into the new system?

The tricky part
It is important to note that data conversions are notoriously difficult to scope. This is because legacy data is often poorly documented and poorly understood, and it is usually more complex than anticipated. Also, it is rarely clean. Changes to applications over time, together with a lack of control over data input, mean that your legacy systems may well contain a substantial amount of poor-quality, inconsistent data. It may not be possible, therefore, to plan everything out in detail before the conversion starts. However, it is far better to obtain partial answers than to proceed blindly in the hope that everything will turn out for the best.

You should also bear in mind the possibility that decisions about data conversion may produce insights that cause the new business model to be revisited and changes to be made. This provides another reason for taking data conversion into account at an early stage in the project.

Neil McGovern is the chief technology officer at Convoy Corporation. Convoy provides a suite of software tools to address the challenges of migrating legacy data to enterprise applications.

Share your experience

If you have already implemented ERP, what would you recommend to make data conversion go more smoothly? Post a comment below or send us a note.