Text analytics helps find what you're searching for

Commentary--Enterprise search is still growing, evolving and improving. Its main purpose continues to be to help users find the answers to business questions hidden in a complex myriad of sources. Questions and queries, such as ‘How did the analysts react to our Q2 earnings results’ or ‘Tell me about the Blackberry Bold’, are the basis to an enterprise search system returning accurate results. However, as search becomes more and more of a commodity, many people are beginning to ask “what’s next” for the enterprise search industry. How can we make it even more compelling and offer a new level of search experience? Search experts have come up with an answer by combining search technologies with the fast-evolving area of text analytics.

Text analytics compliments enterprise search by helping the user uncover the questions they may not think to ask – the all important ‘what do I need to know’ question. In fact, those familiar with text analytics would argue that enterprise search is more important than ever to maintaining a competitive edge, and that text analytics will play an increasingly large part in that equation.

The merging of the two disciplines helps users solve one of the biggest problems with enterprise search alone: Search requires a user to know exactly what it is they are looking for. Or put another way, search on its own isn’t capable of suggesting questions and lines of inquiry to pursue. Text analytics, however, does not require any questions and in fact works best as a tool to discover new pieces of information that a user didn’t even know existed.

For example, when applied to the PR/marketing intelligence industry, thousands or millions of customers review, blog, and share information online about products and services they both like and dislike, and this information can easily impact a brand’s reputation. We have seen the viral nature of social media influences and quickly learned that text analytics fits perfectly into that arena. Not only because it can grab hold of the people, products or companies within a content source, but it can also provide insight into the positive or negative sentiment of those entities. If you did not know to ask (or search) if your company’s brand was at risk, you would never know the information was being spread around the globe hurting your company’s image.

A classic example of this was Apple and the iPod Nano screens scratching (Sept 2005). If, as a PR and marketing professional, you were monitoring this brand you would have noticed a large peak in conversations online about it, but without manually looking into all the documents on the Internet you would not have known why. Text analytics applied to the same data set would have shown you that terms like ‘scratching’ and ‘screen scratched’ were showing and the sentiment of those documents towards Apple was almost uniformly negative (unusual in the blogosphere). If Apple had know this it would have been able to get out in front of the problem before it crossed over from the blogosphere into the mainstream media, and they would have been better prepared to deal with the issue. In this case, it took some time for Apple to address the issue, but luckily they recovered from the blogosphere’s negative comments and sentiment.

There are several other consumer examples illustrating the need for social media monitoring from airline carriers and frustrated passengers waiting on tarmacs, to Target™ announcing its plans to no longer address bloggers’ comments. In all cases, text analytics and social media monitoring (SMM) would have outlined the issues and enabled corporations to search for and identify potential damage to their reputation in a much timelier manner.

So, how does this all translate into a successful search solution? Imagine that when you do a search instead of a simple result set coming back to you (with a few contextual ads thrown in) you get a list of documents accompanied by some additional navigators to help you really drill into the content. Items, or entities, such as people, places, companies and products would appear that are all tied into the original search queries, but may not have been obvious when the search was originally conducted. In addition, the solution would be able to suggest facts about your area of interest you might not know – and would not have thought to ask. A tag cloud of all the people mentioned with a pie chart of the sentiment of the result set as well as a list of the top concepts mentioned are just a few simple examples of what adding text analytics to enterprise search enables you to do.

We are embarking on a time when getting a “good” result from search will just not cut it in regard to providing contextual information that can help drive business action. Text analytics, combined with enterprise search, though, delivers the results companies expect while maintaining a true competitive edge. This allows for enhanced analysis, understanding, and utilization of the information attained as well as superior reputation management.

biography
Jeff Catlin is the CEO of Lexalytics and has over 20 years of experience in the fields of search, classification and text analytics products and services.