In my previous post, I suggested that the justification for data governance must be cast in the context of key corporate value drivers, and noted five high-level dimensions of value. However, the value proposition for any data management activity must both drill down into the specific ways that value is “materialized” within standard operational activities or through better decisions based on analytical application results. Conceptually this sounds great, but to make it actionable you must characterize corporate value in a discrete way. This can be done by organizing a hierarchical description of value creation, and we often break it down into these five categories, as shown in the following table:
Category |
Description |
Example Activities |
Revenue generation | Activities that improve profitability through increasing income |
|
Cost management | Activities that improve profitability by reducing expenses |
|
Risk management | Activities that reduce operational, financial, compliance, or other types of risk |
|
Customer experience | Activities that enhance the full lifecycle of the customer experience |
|
Productivity improvement | Activities that optimize the use of existing resources |
|
Within each of these categories we can continue to drill through the activities to the point where we can identify the data dependencies. For example, we can increase revenues through increasing same-customer sales volume. That could be broken down into increasing the average number of items sold per customer, or increasing the average sales total per transaction, among others. To drill down further, increasing the average sales volume by transaction might imply increasing the number of items sold per transaction, upselling larger-sized items within each transaction, or cross-selling additional complementary items, etc.
In order to increase cross-sold items, the sales process must be aware of the customer profile, and the inventory of items that customer has already purchased. Other analytical results might help drive that sale process as well. In other words, each value driver is supported by a number of different activities, each with its own set of performance measures as well as dependencies on specific data sets. The objective, then, is to understand the different levels of the value hierarchy and map the specific activities and tasks to their underlying data sets and then evaluate the usability of those data sets for each particular business purpose – the topics of our next upcoming posts.