This article originally appeared on the BeyeNETWORK.
The availability of universal access to a unified view of each customer in an organization is intended to enable the “360-degree view,” believed to be the key to unlocking the potential value stored across the numerous data sets and reports floating around the organization. Although this holy grail of customer relationship management may be ever elusive, the use of a master data management (MDM) program to enable customer data integration and consolidation does move the organization closer to enabling customer analytics that can supplement and enhance an evolving customer knowledge practice.
The processes of customer analytics enable business analysts to gain greater insight into how historical customer activity can inform ongoing customer interactions. Assessing customer profiles can drive decisions based on predictions of customer behavior. Customer segmentation and profiling, in which clustering algorithms are invoked using customer records (possibly enhanced with geo-demographic data) as input, provide a set of customer groupings. These groups are subjected to review by business analysts to determine reasonability of the segmentation and the corresponding characteristics selected as the variables for segmentation. This processing and review is repeated until the analysts are satisfied that the model for classification relies on an appropriate set of dependent variables.
That classification model can then be used for customer classification. New customer records are submitted to the analytic model and are assigned into a segment. That segmentation process can inform a number of operational and strategic business processes, such as:
- Targeted marketing, which includes methods to match customer behaviors and profiles and evaluate the likelihood of specific response to specific types of offers. This is intended to reduce the size of a targeted market segment while focusing the efforts on those with a high affinity to the marketed product, which should thereby increase the response rate.
- Offer analysis, integrated into different customer-facing applications to support a customer service representative’s ability to solve problems or enhance the customer relationship.
- Cross-selling and up-selling, which provide approaches to analyze transaction histories and customer predispositions, evaluating success patterns and looking for common behaviors to find affinities between customer profiles and products as well as opportunities to make additional sales or sales of higher-end products.
- Customer lifetime analysis, which assesses the long-term purchasing patterns associated with certain customer segments, identifies dependent variables and characteristics, and projects the value and the duration of the relationship with any particular customer.
- Churn/attrition reduction, which helps in reducing the loss of customers while providing profiles that can help in avoiding customers with a predisposition to defection.
Clearly, master data management is critical to this process since the algorithms used to build the classification model must have access to all the information about a customer along with that customer’s transaction history. As part of the customer data integration process, variances in the values of identifying attributes over time that allowed duplication of customer data will be resolved, with the resulting master data registry providing transparent access to a unified view of each customer. However, MDM can play a role as part of predictive business analytics in a few interesting ways.
First, the definition of master data objects is not limited to those “things” involved in transactions, but can also include derived concepts that become business critical. Customer classifications and profiles are good examples – many different operational and analytic applications can be improved through the use of customer profiling and analytics; and providing a synchronized view of those profiles provides a level of comfort that the use of the consistent profiles will engender trust in employing predictive analytics. This suggests that analytic models, customer classifications and the set of relevant variables are all candidates for mastering.
Second, new customer records are subjected to classification, and the accuracy of the classification process depends on high quality customer data. MDM service levels are defined for ensuring the quality of customer data so that when the clustering analysis is initially performed, the resulting models can be trusted and the business user can have confidence that the classification will be correct when new records are submitted.
Third, the cross-pollination of using master data for both transactions and analytics relies on more than just the customer data and the profiles – it also depends on the associated use cases, business rules and process workflows triggered within operational activities. It is not sufficient to just classify a customer. In any particular situation, a customer is predisposed to certain actions based on his or her classification, and the business rules are used to guide the customer’s behavior that is beneficial to all involved parties (that is, the customer buys what he/she was looking for and the company makes a profit). In essence, these business rules implement the organization’s business policies.
Data mining and predictive analytics features are rapidly being embedded and integrated within many business productivity applications, and this opens the door for agile organizations to exploit the combination of analytical models in real time. This means that MDM environments relying on predictive analytics must have a service layer that provides data access services as well as policy management and deployment services. Companies implementing MDM will evolve their service layer to deliver consistent processes for policy execution involving customer interaction, even across different business applications or interaction channels.
David is the President of Knowledge Integrity, Inc., a consulting and development company focusing on customized information management solutions including information quality solutions consulting, information quality training and business rules solutions. Loshin is the author of The Practitioner's Guide to Data Quality Improvement, Master Data Management, Enterprise Knowledge Management:The Data Quality Approach and Business Intelligence: The Savvy Manager's Guide. He is a frequent speaker on maximizing the value of information. David can be reached at firstname.lastname@example.org or at (301) 754-6350.