What Makes Self-Service Really Work?

A recent Forrester study showed that more than a third of companies rate their self-service capabilities as below average to poor. These findings are confirmed by a recent KANA-IBM study of online service capabilities, which showed that 95% of websites couldn’t answer a simple question, and only 6% offered escalation to email.

What makes some companies successful at self-service, while most others fail? Why are some companies able to achieve the benefits of Web self-service—a better experience for customers and reduced costs for the enterprise—while others fall short?

The recipe for a good self-service experience consists of a handful of basic ingredients. At the core is a healthy knowledgebase that is used by customers and agents. Access methods to this knowledge that make sense to your users must be added. And quick, simple escalation choices to agents who have a customer-centric view must be available.

In a perfect world, the optimal self-service experience begins up-front, prior to a user even becoming a customer. In fact, the Patricia Seybold Group states that "users with a positive search experience spend 270% more on e-commerce sites."

The website should be easily navigable, with a consistent user interface propagated throughout the site. Breadcrumbs and recently viewed pages should be displayed to help orient the user within a site.

If the right information cannot be found, the user should be allowed to escalate the question to an agent using the communication channel of choice: voice, email or chat. The user’s session history—all searches and pages viewed—should be captured so the agent can understand the research that the user has performed on his own.

Once the user becomes a customer, any visit to the website should be personalized. The site should have memory of all customers’ transactions—their orders, their past and pending service requests, transcripts of chat and email interactions with agents, as well as history of any self-service interaction.

Make Knowledge Findable
Too often, searching a knowledgebase is presented as the best way to research an answer to a question. Unfortunately, searches often overwhelm the user with too many answers and may not be appropriate for all categories of users.

If search is used, clarifying questions should be available to narrow relevant results and guide the user to the right topic. Spelling suggestions for mistyped words should always be available.

Techniques to make search "smart" should also be incorporated. Automated learning can be built into a knowledgebase to promote documents to the top of a results list based on use or expert weighting. Navigational aids, such as automatically generated document summaries, help users quickly find what they are looking for.

In addition, repeated searching can be simplified with aids such as the ability to save searches, replay past searches, and auto-complete search terms for customers typing queries into search boxes.

Yet, self-service sites should have much more than search to make content optimally findable. They should show a list of the most frequently asked questions, service alerts and new content so that users have access to the most up-to-date information at all times. To make knowledge an integral part of a user’s life, many sites also allow users to subscribe to content of interest.

Ensure Great Service Experiences with Relevant Knowledge
A knowledgebase is the core of a self-service experience. In a traditional implementation, a knowledge author writes a solution which is an answer to a question that he believes a customer will ask about a particular product or service. This solution is routed to a reviewer, who has the authority to approve the content for the knowledgebase or reject the article back to the author for corrections.

The challenge of this offline authoring model is the relevancy of knowledge, since authoring is performed by someone who is not on the front lines constantly fielding customer questions. This linear authoring flow also introduces a delay between authoring new solutions and having them available within the knowledgebase.

Along with the need for redesigned process, there are several best practices that are being adopted in order to increase the relevancy of solutions and to promote a sense of community between customers, agents and knowledge authors:

• Feedback forms can be appended to all solutions. Knowledge solutions can then be reworked to bring them more in-line with user demand.
• If an agent is unable to find the right solution within a knowledgebase, the agent is able to author a new solution on the fly, creating "just-in-time" knowledge. This allows the solution to be captured with the customer’s exact vernacular.

In this model, solutions are reviewed as they are reused by other agents. Agents take collective responsibility for the quality of solutions. Appending the agent’s name that last modified a solution helps recognize agents who contribute to the knowledgebase.

• Expert users can be allowed to post content directly to the knowledgebase, in effect turning each knowledge solution into a wiki. Expert user contributions are identified and can be rated so that poor contributors can be restricted from adding knowledge content, and star contributors can be recognized.
• Forums can be integrated with knowlegebases that allow users to recommend content to be added, instilling in the user community a sense of empowerment, especially if new content is tagged as recommended by a user.

Customer loyalty is shaped by great service experiences. These basic ingredients of usability, findability and a healthy knowledgebase baked together help provide accurate, complete and personalized answers to user questions. These ingredients are a good starting point to make your users passionate supporters of your products and services.