A blog post at Search Engine Roundtable last week told us that Google’s Ranking Algorithms Dynamically Change Based On Query & Context. That post was reporting about a conversation that took place during a Google Webmaster Central office-hours hangout with John Mueller from Google. John’s answer to search engine ranking signals in queries included the following statement:
That something because we have so many signals we can combine them in different ways and depending on the query and on the context sometimes we can take more like this and sometimes we take more like that. That from my point of view makes a lot sense to focus on more on all things around that query to try to figure out what is relevant for users at this point in time, which might not be the same thing as a later time, which might not be the same thing for other queries, which might not be the same thing for other users.
A patent was recently granted to Google that looks at some different potential search engine ranking signals that are a little different than ones that we often talk about, such as information retrieval scores and links, targeting what appear to be better search user experiences. The description of this patent tells us that it focuses on “providing search results in response to a query.”
Effectiveness as a Ranking Signal
This patent provides search results by determining an effectiveness measure that is a measure of the effectiveness of the resource to present media content related to an entity mentioned in a query. For instance, if you are searching on Google for a song by a particular artist, and you are doing that on a smartphone which has a music app on it which contains songs from that artist, Google may take into account the number of steps involved in using that music app to listen to the song you are searching for when ranking and presenting search results (or movie results or television results, if your search involves those media.)
Social Networking Affinity as Part of Effectiveness
Google may also look to see if you might have more than one music app on your device, and consider if you have an affinity for one of those apps over the other one (tending to use it more frequently.) That affinity may play a role in which app may be selected in search results.
Google may also recognize an affinity by looking to see if you may have a connection to a social network with the publisher on one app over the other app (if you have two apps installed on your smartphone).
This use of a social network to see if an affinity exists reminds me of Authorship on Google+, and how you would see relevant search results when signed into Google from people whom you might be connected to if they were considered authoritative for a topic. We saw that Google stopped the authorship program on Google, but it seemed like a connection at Google+ could still have an impact on search results, as I described in Has Google Decided that you are Authoritative for a Query? It is interesting seeing that a social network connection might influence search results. The patent tells us that this involves a social network connection, but doesn’t point out Google+ specifically over networks such as Twitter or Facebook. However, it is more likely that Google might use a social network that they have control over the data from, such as Google+ over a Facebook that they can’t crawl or a twitter, which may provide a firehose of tweets, but doesn’t necessarily provide Google with other data associated with participants in that social network.
Think about effectiveness as a ranking signal and how it might be used in other types of queries as well. The patent describes a query that includes a request for a restaurant reservation. It might choose a restaurant where it takes fewer steps to make a reservation, and rank that results higher than other restaurants.
Advantages of this Patented Process
The summary of the patent tells us about the following advantages from the patent:
- A top-ranked search result is more likely to satisfy a user’s query.
- A top-ranked search result would require the same or fewer user interactions than other resources for the other search results.
- Search results are for entities that are physically located closer to a user than entities that are physically located farther away from the user.
- Search results that include links for resources specific to a requesting user device and the software installed on the user device, e.g., the user device will be able to display the resources properly and allow a user operating the user device to interact with the resources as a developer intended.
- Search results take into consideration the software that a user operating the user device prefers.
This patent, granted by the USPTO this past Tuesday, is:
Ranking search results
Inventors: Steve Chen, Sara Su, Michael Aaron Safyan, Jason B. Douglas, and Samuel Shoji Fukujima Goto
US Patent: 9,767,159
Granted: September 19, 2017
Filed: June 13, 2014
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for providing ranked search results responsive to a query. In one aspect, a method includes receiving, from a user device, a query specifying an entity, determining two or more resources each of which provide media content related to the entity, for each of the resources, determining an effectiveness measure that is a measure of the effectiveness of the resource to present, to a user of the user device, the media content related to the entity, ranking the resources using, at least in part, the respective effectiveness measure, and providing, to the user device, a presentation of search results for the ranked resources.
Better Search User Experiences
The patent describes effectiveness with a little more detail.
It tells us that when a searcher looks for a content provider, it can be more than a band. It can include things such as “airline reservations, hotel reservations, train reservations, etc.” It tells us about effectiveness as if helping us fulfill our situational needs, like this:
In some examples, the system may determine that when a user visits a web page or uses an application, the user needs to perform six actions to listen to a song on the web page or the application and when a user visits another web page or uses another application, the user needs to perform three actions to listen to the same song on the other web page or the other application. When the system receives a search request for the song, the system may provide a search results page with the other web page, or other application ranked higher than the web page based on the quantity of actions required to listen to the song. We use the term web page, or the term application, to broadly include a web page that provides access to a particular resource, an application that provides access to the particular resource, or any other appropriate type of system that provides access to the particular resource.
The patent goes into more details about social networking affinities, how much information resources are provided if a site has a song on it that people tend to only listen to part of rather than the whole song. Effectiveness can make a difference to a searcher’s satisfaction. It is interesting seeing it considered as a ranking signal alongside things such as relevance, but it does provide a better search experience if it can be identified in a useful way.
I am also reminded of a patent from Google that ranks websites based on how well the databases of those sites respond to searchers queries, which I wrote about in the post: How Google May Rank Websites Based Upon Their Databases Answering Queries. That one tried to make sure that it was referring searchers to sites that could more likely provide them with appropriate answers to their queries. This one aims at providing results that are faster and potentially easier to use, and possibly preferred sites (ones that someone might have an affinity for). Google seems to be targeting good user experiences. And returning to the idea of authorship connections in a positive way.