When you’re filling out the survey that came with the inkjet printer you just bought. When you get to the questions about satisfaction with your “purchase experience,” you’re at a loss. The questions you wish they’d asked aren’t there: Like how long it took you to find a salesperson at the store who knew something about inkjet printers or how you felt when you assembled the printer, but couldn’t find the paper guide.
Businesses are largely stymied when it comes to collecting this kind of customer feedback efficiently in any useful quantity. Current data-processing and data-mining technologies limit the type of input businesses can accept from their customers. That input must be well-defined so that it can be plugged into the narrowly specified fields of the databases that drive today’s customer relationship management systems.
Today’s CRM systems are incapable of accepting large amounts of vital customer information because this information is contained in unstructured text – e-mail, correspondence, comments on survey forms, and telephone call records. Using humans to analyze this text is no solution – it is expensive, time-consuming and yields inconsistent results.
Forms and checkboxes, however, are not the way people communicate best, because they pigeonhole responses into rigid categories that do not adequately reflect what people want to say. People best express their wishes, attitudes and preferences using normal speech.
By limiting themselves to database-driven feedback mechanisms such as checkboxes, businesses miss out on information that is subjective, qualitative and imprecise, but rich in content and therefore essential to understanding their customers. They are paying a high price in the marketplace for this lack of essential information: lost sales opportunities, customer migration and inability to plan intelligently.
Technologies exist that can unlock the information in customers’ outside-the-box communications to help businesses build sales, retain customers and make smart decisions. These technologies, collectively called text mining, can allow businesses to process and act on human forms of communication cost-effectively in ways that were previously not possible on any large scale.
Text mining is the automated analysis of unstructured text to converted it into data that businesses can act on.
Text mining is different from data mining. Through data mining businesses analyze structured information to identify relationships and trends. When confronted with unstructured text, however, data mining techniques are of little value. By contrast, text mining uses highly developed natural-language processing techniques to extract information from text, complemented by specialized statistical methods to organize it effectively.
Text mining technologies are based in part on natural language processing, a branch of artificial intelligence that identifies the content of text through advanced linguistic analysis. It tolerates the ambiguities, inconsistencies, and even grammatical and sentence errors common in language to determine the underlying structure of text with a high degree of accuracy. And text’s structure can reveal aspects of “meaning.” With the latest NLP capabilities, CRM software will:
• Recognize that different words can have the same meaning in certain contexts. For example, it will determine that “go for a car” and “really like a car” express similar attitudes.
• Identify “porblem” as a misspelling of “problem.”
• Discern customer’s feelings, attitudes, and intentions by distinguishing the subtle differences among, for example, “prefer,” “lean toward,” and “like a lot.”
• Differentiate among and quantify expressions of time such as “soon,” “immediately,” and “in a few weeks.”
A CRM system that can analyze unstructured information in this way will be able to categorize texts and organize them to:
• Respond to questions posed in plain English. A call center manager can ask, “How many calls today were customer complaints?” and get a report of complaints organized by subject.
• Automatically deliver information to the appropriate destination according to its content. For example, all e-mail messages that contain complaints about pieces missing from printer boxes can be routed to a specific customer service representative (CSR). Messages about poorly trained dealer personnel can be forwarded to the partner relationship manager.
• Analyze documents and generate concise summaries to enable companies to manage large volumes of information efficiently. A district sales manager can instruct the system, “Give me a five-line summary of each salesman’s monthly report from the mid-Atlantic region for the past three years. Include sales booked and projected” or “Tell me how many people in California said they might be interested in buying a car within about three months.”
• Group large sets of text according to similarity of content – without predefined categories. It can, for example, categorize call records over a certain time, highlighting calls from customers who want to return goods and who want refunds rather than credit or exchanges.
A CRM system enhanced with text mining will be able to accept, analyze and extract data from the e-mail message from the disappointed customer who bought the inkjet printer. It will be able to suggest to the CSR courses of action to placate the customer. It will export the data to the customer’s profile and include it in company reports that management can use to improve quality control and manage its dealers.
With text mining, customer service centers will be able to process more inquiries and generate more sales per CSR because a large portion of the analytical tasks and judgment calls shifts to the CRM system. In addition, inquiries can be directed to the most appropriate CSR, for example, one handling inquiries about a specific printer. Transferring basic analysis and judgment functions to the CRM system ensures that the CSRs uniformly follow standard procedures in dealing with customers.
Today “free text” message response often means sending the customer a response filled with checkboxes. With CRM systems incorporating text mining , businesses will be able to interact with their customers outside the box – in natural, human ways that give full opportunity for personal expression. The consequences will be far-reaching. Businesses will improve the efficiency with which they respond to customers’ needs, building customer loyalty and boosting sales. Customers will finally get the fast, genuinely individualized attention that has long been promised them. Managers will make decisions grounded on a thorough analysis of comprehensive information. Only then will CRM begin to realize its real potential as the driving force in a company.
