Text, Online Behavior Often Overlooked Sources Of Data

While purchase and demographics data are quantifiable, customer feedback is less readily crunched. But a white paper from IBM suggests analyzing what your customers are saying could open up the benefits of data mining to a wider range of departments.

Most organizations already have rich data sources of customer-generated text. An organization can link email correspondence and interactions with customer service reps with a customer account. It can then use word pattern recognition software to link phrases and sentiments with behavior, helping marketers identify potentially dissatisfied customers, according to IBM.

Doing so allows an organization to put the customer into a retention campaign with either personal outreach or promotional offers tailored to mitigate the customer’s concerns.

Text analysis can also help business-to-business marketers identify knowledgeable individuals (who are hopefully purchase influencers) within a large organization. If an organization hosts an internal bulletin board—or even is able to access external ones—comments laden with jargon or acronyms can indicate someone with hands-on knowledge of how an offering is, or should be, used.

Marketing material can be sent to this person extolling the virtues of a given offering. Collateral can also incorporate suggestions for moving recommendations along the purchase approval process.

A marketing department can further look at text comments and modify the language it uses in its pitches—eliminating unused terms, or building in those with which it was previously unfamiliar. At a minimum, text analysis can help a sales force stay current with industry terminology.

Similarly, customers’ online behavior can give insight beyond whether or not they made a purchase. A high instance of customers leaving a site if they are unable to find what they are looking for might indicate the need for a page redesign. Rather than simply serving up a page which says that a search term isn’t found, a site could provide alternative offerings, or access to a live service rep.

Marketers can also examine the online interplay of a customer with an organization. An organization might recognize a detrimental pattern, such as email sent to the company right before a customer returns a product or closes an account. This could help the company trigger a retention effort before the action is taken.

What these data sources have in common is that they bring information beyond transactional data into a customer’s file. Doing so allows an organization to manage its data mining activities from a single location.

There are. IBM argues, many advantages to this. For instance:

* It allows modelers to better know about previous data queries, and avoid repeating their efforts

* It requires that data be standardized, allowing a richer source of information, as well as a wider range of data mining activity

* It enables data mining best practices to be applied to queries made from all facets of the organization; and

* It provides quality control mechanisms when data is updated.