Project Planning: Four Considerations for Data Readiness

Project managers (PMs) are often asked to pull together project plans and timelines before knowing much of the project’s detail. To do this, PMs routinely rely upon past experiences with comparable projects of similar scope. Knowledge of the project’s requirements, resources and competing priorities help build a reasonable plan from the bottom up. However, in most cases, the timeline is already predefined, and PMs are asked to make the project fit into that timeline with a given budget.

This is when the planning and estimation process can get tricky.

In most retail software implementations, several data elements are required from multiple legacy systems to support the new software. This includes foundational data, such as hierarchies, as well as transactional data, like sales.

Fully understanding the legacy data requirements to support the new system and knowing where to source that data is critical to building a sound timeline.

Because most systems vary in methods for storing and exporting data, some sort of transformation is often required to make the data in System A compatible with System B. The level of data manipulation required is not known until detailed system data mapping sessions are complete. Time should be set aside to conduct a proper assessment to account for data transformation efforts, especially if a compatible middleware solution such as Microsoft Azure, Mulesoft, or TIBCO is not already in place.

When it comes time to move from the legacy system, transactions that are already in process or “in-flight” must be converted to align with the new system. The manipulation of existing transaction data through conversion can eat up a significant amount of time in development, testing, and validation activities.

In some cases, it’s possible to cut over to a new system and leave the old system in place, running in parallel through a series of phased rollouts. This approach is often the preferred method of deploying a new solution as it limits risk and exposure. In this approach, the business is running two production systems simultaneously and will need a way to assess the outputs of both systems to ensure accurate reporting.

All data is susceptible to errors, anomalies, or incompleteness for one reason or another. The process of data cleansing works to improve the quality and accuracy of the data to ensure optimal results in future analysis. Before bringing legacy data into a new system, the data should be cleansed to ensure expected results from the new system.

An early assessment of overall data cleanliness must be weighed carefully to ensure time is allocated for any cleansing activities that may be required.