Google has just announced the launch of Knowledge Graph (GKG), a new search functionality intended, in its words “to help you discover new information quickly and easily.” As with any new Google innovation on search, we should take a look at it attentively.
GKG works in two ways: first, it identifies “real-world entities” and, then, it “summarizes relevant content” of any topic around “key facts”.
The identification of “entities” is based on the fact that a set of two or more words in the search box may form a separate and distinct expression than the simple sequence of words that compose it (so “Taj Mahal” is an independent entity, different from the two component words).
In turn, an expression may refer to different ideas or concepts (an Indian monument, a musician or a restaurant, for instance), that should not be confused.
The results list will help the user to disambiguate the possible different meanings of an expression and then will show the most relevant information on each identified concept as “key facts” that, in very little space, contain the basic information about the concept, including images and other related materials, without having to go to any web sites, as it used to be the case.
The content of these “key facts” is based on a massive analysis of user logs, which has identified the more frequently searched issues consulted on each concept, as well as an impressive data processing capacity.
The performance is absolutely amazing, just as fast as you type or click.
How do these developments affect our business?
Google has acquired a unique ability to find and present factual information in the web.
Prima facie this may be a threat to our business, since a significant amount of our content is freely accessible online and therefore can be easily located.
However: (a) a significant part of our content remains outside the tracking capability, as it is not accessible by conventional search engines, and (b) many of our users are looking beyond just the information for commentaries or explanations, which is not an implicit relationship between concepts and, therefore, is beyond the current reach of seekers.
Therefore, our task of improving specific search capabilities for our business remains important.
Let’s share some thoughts about it:
1. The importance of analyzing the logs (Big data analysis)
In order to know how our customers look for information and which information they are looking for, it is critical to study the logs of users, as they are the image of their way to deal with a search engine and to describe their needs into a search box.
Also, as a result of this analysis we can draw conclusions about the content that may be included in our products and the analysis and added value we might add to them, even anticipating trends and user needs.
2. The importance of recognizing concepts in search
This identification is the basis to offer appropriate responses to the queries. We need to understand what users are looking for.
This is a well-known topic by Wolters Kluwer Spain as we have worked in it for several years. Currently we have a semantic search engine, optimized for the legal domain in different languages, that identifies and recognizes different concepts (and their synonyms). Nowadays this technology is successfully being used in Spain, France and Portugal, and could expand in other countries with relative ease.
3. Superseding the classical model of results list
The lists of thousands of results are no longer a model. Instead, advancing the information beside the search box, as quick data or condensed lists, with only the most relevant and accurate information, will be the new paradigms.
Again Wolters Kluwer Spain have anticipated part of that trend with some new features as the documents suggestion (that anticipates the most likely documents for the query) or the iReport (presenting only the most accurate documents in connection with the search ran).
But a long way remains still ahead. Let’s think of new features, specially adapted to mobile devices, such as presenting pre-prepared documents to the user, texts and tools related to their queries, that avoid taking new steps to their search.
4. Anticipating user’s next steps after the search performed
Even if our customers’ searches cannot be answered and solved with a picture, related documents or new queries related to the first one made, should facilitate discovery or serendipity of new approaches to the customers’ problem.
How can we progress further? Helping answer customer’s next question before he just asks) for it, as pointed out by Google? Surely yes, but this imposes a thorough job with the user logs.
Finally, we should not forget the progressively increased integration of our documentary products with new workflow tools (as Kleos or Iter), which will also require some additional work to optimize the functionality.
What do you think about it? How can these new steps taken by Google be leveraged for our business?