Configuration and Customisation

• Technical Evaluation
• Vendor supplied information

1. Review data quality on existing system.
2. Identify new system data requirements.
3. Identify data to be modified or created.
4. Produce a plan for data preparation.
5. Produce a plan for data conversion.

• Data Conversion Preparation Plan

This is likely to be a lengthy exercise in data investigation. You need to understand:

• What data needs to be corrected before conversion
• What missing data needs to be added before conversion
• What new data needs to be created before conversion
• What data needs to be modified before conversion

In the Preliminary Requirements Phase work should have commenced on correcting data and doing some preparation. This was managed as a separate project and by this point, much of the work should be complete. At this stage it may be appropriate to roll the remaining work, and resources into this project.

Typically you can run queries to address the quality of the existing data. This may take some days or weeks to identify problems. There will be some decisions required as to how much data cleansing is to be undertaken prior to conversion. Users typically have an expectation that the new system is going to have pristine data and in many cases, they will end up with the same problem data as they had in the old system. If this is the case, expectations need to be managed. The amount of effort required to cleanse the data should not be underestimated. With key users involved in the setting up and testing of the new system, they may not have time to clean up existing errors without impacting other parts of the organisation. If at all possible it is better to use temporary staff to clean up data.

The second part of the exercise is to identify what new data needs to be created, and what existing data needs to be modified for the conversion. Some of these activities will require decisions on how the business is to operate. For example, something as simple as deciding whether priorities for transactions should be "1, 2 and 3" or "High, Medium and Low" can generate considerable discussion. Once again, the effort should not be underestimated.

Aside from new data, existing data may need to be modified. A typical example is where telephone numbers that were a single field, need to be split into area code and phone number. This can prove very complex if there has been no standard process for recording data. Even if there was a standard process, it was probably never validated so the could be all sorts of combinations e.g. (02) 1234 567, 02-1234567, 021234567 plus errors such as 21234567. Some can be fixed by running a script but others may require manual intervention. Depending on the importance of the data, you may need to run considerable checking after a system change.