Informatica Perspectives

Did you ever have one of those moments where you didn’t know you knew something until you were asked?  I was asked recently to address a question about Master Data Management (MDM) for a Customer Data Integration (CDI) initiative.  As I reflected on my experiences, it dawned on me that over the past 17 years I have been involved in well over a dozen CDI projects, but in the end they all boiled down to three distinctly different strategies.  Each strategy is distinguished by its technical approach, architectural complexity, and value proposition.

The most common approach is Customer Analytics.  This approach is the simplest of the three and involves a data-mart or data warehouse to establish a consolidated repository of customer data.  Data from operational systems is replicated to the repository on a periodic basis and then used to generate historical reports, perform trend analysis, or more sophisticated regression analysis to uncover hidden relationships between otherwise disconnected information.  The most common value proposition is to use the consolidated information for more effective marketing campaigns but some organizations also use it for operations (customer servicing) or to gain financial insights (customer profitability).

The second approach is Process Integration.  In practice this usually involves integration across more than one channel, so we could also refer to this as Channel Integration. In any event, this approach is technically and architecturally different from the Customer Analytics approach since it relies on an Enterprise Service Bus (BUS) or SOA infrastructure that enables access to data in operational systems at the point of contact with the customer.  In this case data is not maintained in a replicated repository but rather accessed in real-time from the system-of-record.  The most sophisticated implementations of this strategy also involve a process state engine to orchestrate long-running processes across systems and across channels. The value proposition for the approach is either improved customer service which results in greater retention and growth or the ability to perform real-time cross-selling of products to increase customer wallet share.

The third approach is a Mass Customization.  This approach is the most complex to execute since it involves extremely tight coordination between business and IT functions as well as collaboration across multiple business units. It also involves making trade-offs between product lines (i.e. sub-optimizing performance in some areas) in the interests of maximizing revenue and profits from defined customer segments.  Technically this solution is also the most complex since it involves a CDI repository, formal regression analysis and predictive models for customer profitability, retention, risk and propensity to buy, a centralized strategy decision engine, and modifications to operational systems to adjust behavior on a customer-by-customer basis in response to instructions from the decisioning system.  The value proposition for this approach is maximizing corporate profits by treating your best (most profitable) customers the best while minimizing operational costs for the rest.  If you’re interested in this approach, check out the book Angel Customers and Demon Customers by Larry Selden and Geoffrey Colvin.

This topic is too deep for just one blog posting. I’ll have to add this to the list of potential topics for another chapter in the next ICC book.  In the meantime, reader comments and discussion is welcome.