2006 May 01
Putting Customer Data in Context
David M. Raab
DM Review
May, seek 2006
Customer data integration systems are the central corporate mechanism for capturing and distributing customer data. Like other elements in the services layer of a modern information architecture, tadalafil they are independent of any particular application system yet available to all.

So far so good, see but pretty darn vague. What information should the customer data integration actually transport?

Most answers would start with attributes like customer name, address, age, income, and interests. Business customers have different but similar attributes such as industry, number of employees, and ownership.

More ambitious definitions add transactions such as purchase history and communications. These are not inherent attributes of an entity, but behaviors that occur over time. They are distributed through a CDI system because they help other systems to better manage customer interactions. A telephone agent should know about complaints received via email and orders placed over the Web.

But once you define customer data as anything that helps to manage interactions, the scope expands dramatically. Interaction management is based on understanding customer needs, so everything that indicates needs is included.

Needs are sometimes stated explicitly. But they are more often inferred from behavior and attributes. Behavior is generally the more powerful indicator because it’s more specific—the customer purchased a particular product so she must have a need for whatever that product does.

Or maybe she doesn’t. She might have purchased the product for someone else or for something other than its intended use. When someone bought religious candles at the grocery store, was it because she’s religious or she needed candles? Would it help to know that she purchased them the day before a hurricane or a day after the Pope visited her town? Would it help to know whether the store had other types of candles available? What should you offer her next: a religious statue for the garden or a portable power generator?

In other words, data has a context. (I think someone clever once said that data in context is information. If not, I’ll say it here and you can quote me.) Knowing the context is important for customer management because it clarifies which needs a particular piece of behavior truly indicates. If the CDI system is the vehicle for distributing useful customer information, the CDI data model has to include context as well.

Extending the data model is easy. What’s hard is gathering the context information in the first place. The grocery point of sale system doesn’t need to know about the weather or the Pope; it just needs to record the transaction and maybe a customer loyalty card number. You have to get the context from somewhere else and later connect it to the transaction.

Source systems do capture some context data. Consider location. People behave differently in their local neighborhood, near the office and on vacation; when they live near to a store or far from it; when they’re calling from their own house or a friend’s or the highway. Operational systems capture lots of location data. Retail point of sale records include the store. Telephone switches can identify the calling number. Web systems capture the IP address. Cellular systems know your exact latitude and longitude. Order processing systems know the billing and shipping addresses. Assuming privacy needs are respected, this information can be an important part of the customer data context.

Other context information may be captured elsewhere within the company. A demand forecasting system might track weather and holidays. Merchandising might record stock-outs and product sales rates. Marketing often captures competitive behavior and advertising promotions. When internal sources are not available, external feeds may be. You can easily subscribe to feeds for economic indicators, stock market movements, gasoline prices, and major news events.

Some context information can be merged with other customer data directly within the CDI system. Date and location are often simple to match to transactions, although they may need some interpretation to apply meaningful categories such as “at home” or “pre-holiday”. Other context information is more subtle, such as market baskets, purchase sequences or comparisons of promoted vs. purchased products. This will require processing outside of the CDI system itself, but must still be part of the information the CDI system passes on to help manage new customer interactions.

* * *

David M. Raab is a Principal at Raab Associates Inc., a consultancy specializing in marketing technology and analytics. He can be reached at draab@raabassociates.com.

Leave a Reply

You must be logged in to post a comment.