Software vendors, white papers, and news articles exist in excess, all hailing the benefits of a single, centralized data repository to support business operations. These include advocating the benefits of solutions in the form of master data management (MDM), enterprise resource planning, (ERP), business intelligence (BI), and/or Big Data solutions. Often times, one of the first areas in which to focus are the benefits from establishing and maintaining a centralized customer data repository. To illustrate the point, a recent news article identified “dissolving internal silos to have one view of the customer” as a key big data opportunity of 2015.
It’s easy to understand why the customer domain receives so much focus. By establishing a single view, it allows for holistic understanding of a customer’s activities, and allows for more sophisticated up/cross sell opportunities. In order to achieve this capability, multiple systems are transformed/consolidated into a single repository. Then, data users are educated to utilize this new source of record for all business purposes. Common definitions are applied throughout the organization so that data is shared and used in a consistent manner, as well as the added benefit of increased data quality,
Let’s take a hypothetical look at how a bank’s marketing department realizes these benefits. Having a centralized customer view can be extremely helpful with the upfront data mining and selection process of customers for an upcoming marketing promotion. In this example, the customer repository is used to identify all customers that have both a checking account and home loan but do not currently have a bank-issued credit card. This customer population is then targeted for an upcoming cross-sell credit card promotion. By having all the customer attributes, products, and activities all in one location, the process is streamlined and efficient.
But what about applying these same concepts to internal data needs?
Often times, this is where the business processes, and corresponding data, becomes disjointed. Marketing will take the identified population and create the targeted promotional offer – again, for illustrative purposes, let’s assume it’s a mail promotion. The offer will be reviewed and executed. Then responses will be measured over a period of time based on the number of customers that sign up for a particular offer
This seems simple enough, right?
Actually this is where things can get really messy, as a variety of questions will be asked as to the success of an executed promotion. These questions will evolve over the life of the promotion, and typically become more complex and require the analysis of data from multiple departments and systems the farther out from the promotion’s launch date a question is asked.
To continue on our example, one of the first questions asked is the promotion’s take rate. This measures the number of customer that signed up for the promotion, and looks to simply understand who, in the credit card system, has been flagged for this offer. Next, the question of customer retention will be examined (i.e., How long did a customer remain on this promotion)? This requires an historical look of activity, and will take into account all customer that accepted the offer, those customers that have switched to a different offer, or no longer have the credit card.
Financial questions will also be asked (i.e., “How profitable was this campaign?”). To do this, the credit card revenue and write offs will need to be understood as it relates to the promotion. Additionally, the cost the bank incurred to launch the promotion will need to be known (this may not be stored in a system). Operational impacts will also be examined, such as how much customer service support was required to handle these customers, was it significantly different than that of other customers, and involves examining the call center activity (yet another system) as well as understanding the statistical significance of the support needs.
The end result is often a fragmented internal data infrastructure that must be manually consolidated and reconciled to answer important business questions. Questions that will be accelerated as a significant investment was made up front to create a single customer view to support business activities. Questions must now be asked as to why have the same investments not been made for these critical internal processes that support these business activities, as well as what is the impact to the business by not making these investments.