Don’t Neglect Proper Planning and Testing for Data
Let’s face it. The primary reason for the software you’re implementing is so that you can get data into and out of the system reliably. Think of your software as the plumbing in your house and your data as the water that will flow through that plumbing system. Despite the obvious risks of plumbing a whole enterprise and then testing with fake or insufficient data, I am continually amazed at how little effort most companies expend on managing the Data Readiness and Data Validation aspects of their project. The consequences of that lack of preparation can be disastrous!
Data is fluid. Since many implementations (such as for a new ERP system) take place over an extended period of time while the business is in full operation, it is not unusual for the data to change significantly over the life of the project. Customers, Vendors, Employees and Materials are being created and updated. Customer Pricing is changing and Territories are re-aligned. Therefore, you must plan for sufficient and proper “Mock” Data Conversion Load Cycles to ensure you have handled and validated the latest data available and have addressed all possible variations. Too often, the project team plans only one or two partial data conversions mocks, but this approach is NOT a promising plan for a dry basement after Go Live!
Data is easily contaminated. Without the proper management of activities such as Data Governance, Data Profiling, Data Cleansing, Data Conversion and Data Validation, even small amounts of rotten data present in the legacy system can create a mucky mess in your brand new system. Careful attention must be paid to planning, communication, execution and monitoring of these data related activities to adequately de-risk your ERP Implementation.
You-know-what flows downstream. Small data problems at the beginning can become really big problems as the data moves through your systems, being processed and combined with other data. For example, a missing Sales Territory on a handful of migrated high volume customer records may mean that your combined Sales by Territory reports may be inaccurate. Make sure you plan sufficient time to thoroughly and formally test ALL of your downstream applications, databases, analytics and reporting with migrated real data from your ERP Test instance before you go live, leaving enough time in the project schedule to do upstream corrections and re-testing for any defects found.
To boil it down — Following this advice will make your Data Readiness Strategy a well-constructed pillar of the project from the very start and will help build business confidence in their data. You will be better prepared to turn on that data faucet at Go Live!