Sometimes when you perform a search at Google, you may notice that search results shown from some forums might include additional information, such as how many posts are in a thread, how many authors participated, and when the last post was made, like in the following search snippet:
If you’ve seen those types of results and wondered why Google might include that kind of information, you’re not alone. I’ve been wondering as well.
A recent Google patent application provides some details on why Google is showing that kind of information, how they identify discussion threads, and the kinds of information they may potentially show in search results.
Providing Posts to Discussion Threads in Response to a Search Query
Invented by Tomislav Nad and Jonathan Wilson
Assigned to Google
US Patent Application 20100030753
Published February 4, 2010
Filed: July 31, 2009
According to the patent filing, the kinds of information that might be shown from a discussion thread could include:
- Times when most of the posts were made on the discussion forum,
- Authors that posted most of the posts.
- Number of authors that posted,
- When the most recent post was made,
- Number of posts made by the same author, and;
- Number of replies to posts.
The real question though, is why? Why show a different kind of search result when it involves a “discussion thread?”
We’re told that one advantage of showing this kind of result is that it provides a searcher with an “integrated view of a discussion thread that may include many relevant posts, which might otherwise be shown as scattered web pages in the results provided by the search engine.”
So, rather than separate posts from a discussion thread, and treat them as if they were separate pages in search results, Google is highlighting the fact that a search result is from a discussion from a number of authors and may contain more than one relevant result to a query.
It’s also possible in some instances that Google may also show links to additional discussion threads on the same forum under the first snippet that may also be relevant to what a searcher was looking for.
The patent filing also provides us with a fairly wide definition of what a discussion thread might be, including:
- A blog that can receive comments from viewers;
- A threaded discussion in which messages that share a common theme such as a subject,
- A micro-blog in which “users send brief updates that include text, audio, images, and the like, for publishing.”
- and the like.
We’re also given some hints about what Google might look for to determine whether or not a page it has found on the web might be a discussion thread.
One of the criteria the search engine may look for is the appearance of certain keywords that appear upon a page, such as “forum,” “subject,” “thread,” “post,” “posted by,” “reply,” and so on.
Another signal that might be used involves looking at the structure of the web page, to see if it is “characteristic of pages that include posts to discussion threads.” Those could include dates and times when posts were made, user names associated with those posts, links back to a forum home page, and similar indications that a page is hosting a discussion.
It’s not unusual to see search results containing maps, images, news, blog results, and other kinds of information taken from many of Google’s vertical searches such as Google Maps and Google News. If you haven’t looked at Google’s specialized search for Groups, you may not have seen that they not only include “Google Groups” (and old usenet group) threads, but also results from forums across the Web, which use the kinds of search snippets described in this patent filing.
One of the most interesting aspects of this patent filing is the definition of “discussion thread,” which looks like it not only would include forums, but could include blogs with comments and microblogs. That could be something to possibly watch out for in the future.