With so much data available about customer behavior today, it has become increasingly difficult to figure out which information customer support teams should pay attention to when looking to improve their documentation or knowledge base. Nevertheless, developing (and continually expanding) a comprehensive knowledge base remains an important step for attracting, acquiring, and retaining new customers—so where should you start?
A growing trend with customer support teams is the interest in analyzing customer intent data. Learning about customer intent, which includes social engagement data, (such as Facebook “Likes” and Twitter “Retweets” for your knowledge base content) and site search analytics, gives your team the opportunity to provide better curated recommendations and improve your company's knowledge base.
Netflix uses Facebook likes, Google+ shares, and Tweets to gauge which articles from their help center are most useful and/or popular.
While social engagement is an important suggestion of intent and satisfaction, your view of that intent is limited. A simple share or a like are not as explicit as what is arguably the most important source for raw data on customer intent—namely user search behavior.
Analyzing search behavior is like directly asking your users, “What are you looking for?” And while this realization is key, the next step is determining exactly which pieces of information from your site search data are most revealing about what your customer support center is or isn’t doing well.
Basic site search data, like query volume, is readily accessible in Google Analytics under the Behavior tab. Additional site search data made available by other platforms include searches that delivered no results, and top performing pages in search results.
Additional site search data made available by other platforms include searches that delivered no results, and top performing pages in search results.
The most common search queries in online knowledge bases are either short phrases or questions. For short questions, take note of the queries that are product focused.
For question related queries, your approach will be a little different. First, look for words such as “how,” “can,” “who,” “why,” “where,” “when”, and “what”. These are questions generate open-ended responses and will potentially require you to create numerous types of content to effectively supply your customer with the right information in the knowledge base. Questions with “can,” “is,” “are,” “will”, and “does” will require a yes or no response and less content types. This is an important distinction so that you can build the proper plan for creating content that answers these questions in your knowledge base.
Once you have your list of meaningful queries, conduct your own search of these questions to see what results your customers are getting. If you’re not satisfied with what is being delivered, then refine your content strategy to also include plans for filling in the gaps.
It is important that you spend enough attention and analysis on queries that do not generate any results. This means that a customer has come to your site, searched for something, and not received any results to click on. This is a poor experience to provide to customers and generally forces them to reach out to a support agent (in the form of a phone call, email or chat). Even worse, this experience can frustrate a user and leave them uninterested in learning more about your product, causing users to churn and consider alternatives. A good way to prioritize content for these queries is to do some analysis to see how much customer outreach, through email, phone calls, or chat, had to do with the same topics and queries from the no results list. If the number is meaningful, then you know that your customers are taking additional action to reach you and there’s opportunity to do something about it.
In the ongoing challenge to support your users and help them get the most value out of your product, it is important to create a wide array of content that users can easily discover and consume as they encounter problems. Given this challenge, pay close attention to information that gives you direct insight about what your users need—such as social behavior and search data, since these stats are generated directly from your users and not inferred or guessed at on your end. With this information in hand, you can take the first steps toward building a knowledge base that answers the questions your customers are searching for.