Dmitri Soubbotin, CEO of Semantic
Engines, the maker of SenseBot!
Part 1 of 2
Various types of user queries
There are a number of different types of queries that users make when searching on the Web. A common classification distinguishes between transactional, informational, and navigational queries, although the borders between the types are often fuzzy. (A thorough analysis can be found here.) A typical navigational query is made when a user is looking for a particular Web site, e.g. “Newark airport long-term parking”. A transactional query is one in which the user is looking for a site to perform a transaction, e.g. buy a product.
An informational query is the one in which a user’s goal is to understand a concept, or form an opinion on a subject. Examples could be: semantic search engines, social networking, Iraq war.
One can say that in the case of an informational query, the user is searching for an answer, even when the query is formulated as a combination of keywords and not as a human-language question. A query like “When was Galileo Galilei born?” is looking for a direct answer, whereas the query “Galileo Galilei” is probably looking for an answer in the form of an overview, answering a few questions at once: who, when, what is he famous for, etc.
Major search engines have only one type of result – a list of links to matching Web sources. In order to get an answer to the query, the user needs to dig into individual results. Let’s look at an alternative way to serve results that would allow the user to get to the answer faster.
Query example: “Marseilles Hotel Miami”
Imagine you are planning a trip to South Beach, and have a few preferences in mind. You want to be close to the Art Deco district, preferably even stay in an Art Deco hotel; you’d like to be right on the beach; you want to be close to the entertainment district but are picky about the overall ambiance. Most of these parameters are a bit vague, and cannot be entered in a typical travel site search. You can tell that this is an informational query, even though at the end it may result in a transaction (booking).
Let’s say a friend mentioned a hotel he stayed in, so you type “Marseilles Hotel Miami” in Google. As usual, tons of results come back, most of them from hotel booking sites. The only fact you can get from glancing at the first page of results is the location – Art Deco district. This is not much information, so now you have to start clicking on the links. But here is a time-saving alternative – an overall summary of all search results from the first page. (You can view the example here, starting with the Google results and displaying a summary below. The example was generated by our Firefox browser add-on that integrates with the Google results page.) The summary contains a lot more useful facts than the Google page (highlighted in the example for illustrative purposes): you can see that Marseilles is a 3-star hotel, located in the entertainment district, right on the beach and boardwalk; it is an original Art Deco style building; and other detailed information.
Glancing through the summary, you can also see which of the sources consistently give you more useful information, and drill into those through the links within the summary. There is a certain amount of “noise” in the summary (non-informative sentences), but the overall usefulness outweighs the drawbacks.
Of course, not all travel searches need a summary, but if you are researching a destination rather than just checking price and availability, you might benefit from a summary and save time.
Let’s now look at a pure informational query, where an overview of the topic is even more appropriate. We will search for “Ecotourism”, which is a concept that has a number of different definitions. Yahoo dutifully returns a list of links, but you cannot get a good idea on the subject just by glancing at the first page of results. The sources themselves are not bad, but the way they are represented on the search results page is not very helpful. The chunks of text displayed next to the links are not very informative (this situation gets even worse when a query contains more than one term). So it looks like you still have to spend your time drilling into a few sources. Here is an alternative: look at this summary, constructed automatically out of the sources on the first page. Note a “Semantic Cloud” above the summary, which is a tag cloud consisting of key semantic concepts extracted from the sources. It gives you a quick visual overview of the topic; for example, you can spot in an instant Costa Rica as one of the tags (and it is indeed a famous ecotourism destination). Some other interesting concepts include: conservation; water; wildlife; locals; etc. Clicking on a concept in the cloud will force the system to regenerate the summary with a focus on that concept.The summary gives a pretty decent overview of the topic (disregarding some amount of noise). Even a person new to the subject of ecotourism should be able to maintain a conversation about it after reading through the summary. The summary will also show which of the topics are more useful than others – note that the system used only 5 out of 10 sources returned by Yahoo. The user can follow through to any source he likes right from the summary. You can see other examples of summaries here.A new type of result
The summary described in the examples above can be seen as a different type of search result altogether. The essence of the top relevant sources is automatically extracted and presented to the user, in many cases obviating the need to drill down into individual sources. The “lightweight” sources are excluded from the summary, even if they are highly ranked by major search engines.The primary goal of a system producing this type of result is saving the user time by giving him a quick overview of the topic. Even when the overview is imperfect, the benefits are there – you can get perhaps 80% of the answer in a few seconds, and in many cases that 80% will suffice. Typically, queries of the informational type are the ones that benefit the most from a summary.
But why do we need a new type of result? Looking at the search market share, it may seem that most people are content with Google or Yahoo or Microsoft. However, various satisfaction surveys indicate that users spend too much time searching and not always finding. That’s why dozens of “alternative” search engines (comprehensively covered at AltSearchEngines.com) are trying to create a more efficient way of searching. Among these, there are a number of “semantic” search engines.
Part 2 of Alternative Search Results will be posted tomorrow.
Caronte Web
Federated Search
Highlight Health
MSearchGroove
Ramenos Blog
Reel SEO
Software Abstractions
The Genius Files
The Next Web