September 2000

BY DOUG WARNER, RIGHTNOW TECHNOLOGIES, INC.

Effective customer service is all about delivering the answers customers are seeking. How information is stored, augmented and organized will determine how effective any organization's customer service efforts will be. The best knowledge base technology uses customers as the driving force behind the service organization. Using customers to guide the content and structure of the knowledge base improves the level of service and decreases the customer service representative's workload.

What Is A Knowledge Base?
Generally speaking, a knowledge base is an organized collection of information. A common example of a first-generation knowledge base is the card catalog in a library. The books in the library are listed alphabetically by title, author and subject. Someone searching for the book "Moby Dick" by Herman Melville will find the same entry if he looks under the name "Melville, Herman" or even under "Ahab, Captain" or "whaling." Unfortunately, this approach works only for those few who come to the library with some foreknowledge of the desired book's title, author or contents.

A more up-to-date example of a second-generation knowledge base is the list of frequently asked questions most companies use for their Web-based customer service. In their most basic form, FAQs are questions and answers with no discernible order. This format lacks even the simple cross-referencing function available in the library card catalog. Some of the more advanced FAQ lists are actually knowledge bases with cross-referencing capabilities similar to the library card catalog, but there is a better way.

A third-generation knowledge base places the most useful answers in the most prominent place in the FAQs list. Consider a situation in which someone with a customer service issue, such as shipping delays during the holiday season, accesses a Web site to try to find information within the knowledge base. Since this question is commonly asked during November and December, customers visiting the site during that time should immediately see the answers to their questions before they even get a chance to ask them. Since this question is usually never posed in July, other questions should have a more prominent placement on the list when the site is visited during the summer. However, in both cases, customers should still be able to ask their questions in words familiar to them and receive immediate, accurate answers.

Making Use Of The Customers
One way to reduce the workload on customer service representatives while improving customer service is to implement a high-quality knowledge base for Internet customer service. Following implementation, use your customers to drive the content and organization of the information within the knowledge base. Customers will often provide feedback on how to improve their online experience, provided the process is simple. Assimilating this feedback and acting on it results in your customers helping each other with little effort or intervention by the customer service representative.

Methods exist to prompt for feedback from customers. These methods are easy and work well within the organic knowledge base paradigm. These include explicit customer feedback on an individual FAQ or group of FAQs, customer questions asked directly of the customer service representative or surveys sent shortly after a question is answered.

These methods include "implicit" and "explicit" ranking of FAQs, which refers to the dynamic ordering of question/answer pairs based upon what customers find most helpful at any given time. The "implicit" rankings are collected by analyzing the paths customers take through the knowledge base without actually asking for help from a customer service representative. Since all customers automatically provide this information, this method returns the bulk of the information. Requesting multiple-choice-style feedback on the usefulness of a particular information item collects the "explicit" rankings. A reasonable percentage of customers will provide feedback of this nature. Assuming a customer responds negatively to the multiple-choice-style "explicit" ranking, the customer can be prompted for a short answer about why a particular FAQ was not helpful. Responses to this prompting will arrive in lower volumes because only those customers providing negative explicit feedback are being asked to respond and because only a portion of those customers will provide detailed feedback.

A customer asking a direct question of a customer service representative can also be considered feedback about the knowledge base. A direct question says either the knowledge base contains the question and it is difficult to find or the knowledge base does not contain the question. In the first case, it's possible to use a match between the customer's question and an existing answer in the knowledge base to adjust the ranking in the knowledge base for that answer. Or, the customer service representative can reword the question/answer pair to this particular question, making it easier to find in the future.

However, when the answer is not found in the knowledge base, a customer asking a direct question of a customer service representative is an important piece of information to a company. This is the most important case in which a customer service representative actually influences the knowledge base. Since asking a question requires more effort on the part of the customer, these requests should arrive fairly infrequently. Finally, a questionnaire sent to customers who accessed the knowledge base can be used to adjust the rankings in the knowledge base, but surveys often require a fair amount of customer effort and will likely result in the lowest level of feedback.

Consider all these feedback methods from a larger perspective. With the largest bulk of responses coming into the knowledge base, the amount of information relayed is rather small and easy for the customer to provide. As the responses become more involved for the customer to provide, the number of them available for updating the knowledge base decreases. The weight given to the customer responses of each type should reflect this relationship by putting more emphasis on the responses that require more customer effort and less on those requiring less customer effort. The result is that the more effort customers spend on the knowledge base, the better the knowledge base becomes for future customers and the workload for the customer service representative remains low.

Future Directions
With the advent of the World Wide Web, customers suddenly became able to answer their own questions by reading a list of frequently asked questions. As techniques for improving Web service advanced, the Internet customer service knowledge base appeared. However, early knowledge base implementations did not use the wealth of information available from the customers because the knowledge bases were crafted in a brittle fashion based upon expert systems and similar approaches.

With current technology, knowledge bases contain features that allow customers a larger role in shaping the information available online. In the near future, we should expect an even greater shift toward customers finding their own answers, helping other customers and relying on customer service representatives for only the most complex questions.

Another area ripe for explosive growth is the presentation of related information. In even the most advanced Internet customer service knowledge bases, there exists only a simple list of FAQs. Since this helpful information exists in the knowledge base, it makes sense that a more advanced method for customers to find answers would be to "surf" through this related information with greater ease and comprehension. It might be much easier to see a relationship between two pieces of information given a graphical presentation of the relationship.

Following that example, it should be possible to view the entire knowledge base as one large network of related information. Finding the answer to a question is then as simple as following a path toward questions increasingly related to the information a customer is seeking.

Further improvement to Internet customer service knowledge bases will come from integration with other e-commerce tools such as surveys, marketing and shopping carts, elements generally outside the traditional area of customer service. To the extent that information from customers can be gained and used, customer service knowledge bases can only benefit.

The field of Internet customer service has come a long way since the days of static FAQs. Powerful new systems are available using organic knowledge bases to reduce the customer service representative's workload while increasing service to the customer and the volume of questions answered. The future holds even more advances and opportunities to strengthen customer loyalty for those companies that implement the most intuitive and customer-friendly systems.

Doug Warner is the Knowledge Base Project Lead at RightNow Technologies  in Bozeman, Montana. Founded in 1997, RightNow Technologies automates customer service and technical support operations for Internet-connected organizations.

[ Return To The September 2000 Table Of Contents ]