A newly published patent application from Google describes how a combination of different types of Google-generated profiles associated with a searcher might influence the results that they see. The description in the patent filing is substantially the same as some I’ve written about in the past involving personalization from Google, in my 2006 post Google Personalization Methods.
But I couldn’t help but think of the role that Google Plus might play in personalized search from Google as well while reading through the patent. Is information from my Google Plus profile used in personalization? Is other Google Plus information part of personalized search?
To recap the area that the patent filing covers, it begins by telling us that the purpose behind the personalized search is to help in filtering the very wide range of relevant results we might see for a query. Since most queries are usually fairly short (2-3 words), the results of such a query might return an ever-growing number of results as the Web continues to expand. A searcher might be overwhelmed by the number of results they see if those aren’t organized based upon things like their relevance to a user’s query and their importance under something like PageRank.
And yet even PageRank has limitations. As the patent filing notes:
In reality, a user like the random surfer never exists. Every user has his preferences when he submits a query to a search engine. The quality of the search results returned by the engine has to be evaluated by its users’ satisfaction. When a user’s preferences can be well defined by the query itself, or when the user’s preference is similar to the random surfer’s preference concerning a specific query, the user is more likely to be satisfied with the search results.
However, if the user’s preference is significantly biased by some personal factors that are not reflected in a search query itself, or if the user’s preference is quite different from the random user’s preference, the search results from the same search engine may be less useful to the user, if not useless.
For example, someone who enters the term “blackberry” into a Google search box might see a bunch of results related to the type of phone, without any fruit-related content insight. But someone who has a history of searches for food and cooking might be much more interested in results that have nothing to do with phones. We’re told that a search engine shouldn’t make it hard for searchers to find the types of things they might be looking for by forcing them to drill down deeper into search results or to refine their queries.
A user profile might be attached to a searcher that might come from sources like:
- Previous search queries submitted by the user,
- Links from or to documents identified by the previous queries,
- Sampled content from the identified documents
- Personal information implicitly or explicitly provided by the user.
When I search for “blackberry,” I just see phone-based results. Then again, I’ve been looking at a lot of phone-related pages over the past few months while considering a new phone to purchase, and I’ve even posted a few posts at Google Plus related to mobile phones. I’ve looked at a lot fewer pages involving food and recipes, and I don’t think I’ve posted more than one or two posts about food.
The steps that might be involved in personalization, from a high level, would include the following:
- The search engine receives a query from a searcher
- The search engine identifies a set of documents matching the query
- The initial rankings for pages use PageRank and text associated with both the document and the search query
- The search engine also generates a “personalized” profile (likely before the search in question) for the searcher and correlates the user profile with each of the identified documents
- That correlation between documents and the user profile produces a profile rank for each of the documents, indicating how relevant they might be to the searcher
- The search engine would then combine each document’s generic rank and profile rank into a personalized rank
- The documents would then be reordered based upon their personalized ranks, and displayed
Instead of using a single focus for a generated user profile, those profiles might be made up of many sub-profiles, each of which may characterize a searcher’s interest from different perspectives. These could include:
A term-based profile with a number of terms, each term carrying a weight indicative of its importance relative to other terms.
A category-based profile using multiple categories, possibly organized into a hierarchical map (like the hierarchy you see DMOZ organized into).
A link-based profile with several links that might be directly or indirectly related to pages or documents identified in a user’s search history, with each link having a weight indicating the importance of the link (like PageRank).
The patent filing is:
Personalization of Web Search Results Using Term, Category, and Link-Based User Profiles
Invented by Stephen R. Lawrence
US Patent Application 20120233142
Published September 13, 2012
Filed: November 11, 2011
Abstract
A system and method for creating a user profile and for using the user profile to order search results returned by a search engine. The user profile is based on search queries submitted by a user, the user’s specific interaction with the documents identified by the search engine, and personal information provided by the user. Terms for the user profile may be selected from the documents accessed by the user by performing paragraph sampling or context analysis.
Generic scores associated with the search results are modulated by the user profile to measure their relevance to a user’s preference and interest. The search results are re-ordered accordingly so that the most relevant results appear on the top of the list. User profiles can be created and/or stored on the client-side or server-side of a client-server network environment.
The claims section of this patent filing is different from those in the patent applications I wrote about in 2006. This patent filing is a continuation patent of two others that were abandoned by Google. The first was filed under the name “Personalization of Web Search,” which was filed in September of 2003. The second shares the same name as this present one and was filed on May 12, 2010. Looking quickly at the patent office’s database for the history of those versions of the patent, both had some issues that challenged the language of the claims within them.
The claims in this newer version don’t explicitly point to or describe the use of information from Google Plus, and yet reading through them it seems that Google Plus could be used to add to the information that Google might look at to personalize Web search results.
The patent filing tells us that people don’t usually share or update their profiles on the Web very frequently, and don’t always do a great job of sharing things that they are interested in, such as interests and likes and so on. How accurate and up-to-date are your profiles at places like Facebook?
But it also tells us that the kind of demographic information that might appear in profiles might still be very useful, such as information about education, employment, and location.
There’s a definite preference from Google to look at things like search history, including the things we search for, the things we click upon, what we might bookmark, and more in patents from Google that involve personalization.
Google could use information about the things we write about and the links we share at Google Plus quite as easily as it might use our Web history to learn about our interests and preferences. It might also look at our interactions with others on the social network as well, and the people we interact with.
While our interactions on other social networks outside of Google Plus might also be considered as well, it’s much easier for Google to take information about our social interactions on Google Plus and use it to add to a user profile that might be used to create a personalized rank for what we see.
Might Google be using Google Plus not only to help determine what we see in Search Plus Your World results based upon our social activities but also in the personalization of those results?
If not now, I would think there’s a very real possibility of it happening soon.
It would really be useful for the user as less typing would be required which means less long tail queries would be entered. Person who has a previous history of searching for recipes related to Blackberry might instantly see food and recipes websites when he/she enters the term Blackberry. Now, there is no need to enter “Blackberry recipes” “Blackberry” -phone and other types of searches.
The search results are constantly being made personal, data from gmail as well as Google + would be used present results. This means more qualified traffic would be diverted towards a website which is good from webmaster point of view.
I’ve created a few unique profiles myself on chrome with different purposes. Personalization is happening for sure with just logged query history (the accounts give different results b/c I use them for different types of searches), but since it’s hard for me to maintain 3 G+ accounts of myself at once and see the changes over time (and conclude changes are from G+ and not from my query history), I’d be very interested to see if anyone else would be able to conduct a simple experiment to validate the G+ influencing personalization theory.
Anyone interested in creating brand spanking new G+ profiles that like varying things and then test out the personalized search results of these G+ profiles?
I’ll make it easy for you Google by shutting off personalization. Don’t worry, I’m smart enough to drive relevant results by adapting queries. You just worry about the integrity of the top ranking slots for MY relevant search queries.
Insightful post Bill with the key takeaway for me to pay close attention to Google+ as it will influence future search results well beyond Google Search Plus Your World in a profound way that is hard for most to image now.
It seems to me that Bing is/has been doing this with Facebook and Twitter data for a while, at least at some level. If Facebook gets serious/successful in search, which is a BIG IF, then would Google really be able to defend this patent based on previous use by others like Bing and Facebook prior to the filing date?
The advantage that Google has over Bing and Facebook is the huge advantage in sheer search data which could be richer than the social graph information for segments who use search many times throughout the day on Google but very rarely tweet or like/share something on Facebook. Deep pockets don’t hurt either! Thanks for sharing.
Hi Rick,
Thanks. Yes, a person or business that is active on Google Plus, and builds a strong presence there on topics related to what they offer on the rest of the Web and interacts with others, could definitely benefit. Hard to say how Bing might or might not be incorporating social networking activity into personalized search. They definitely work to include social connections in Bing’s search results when you’re logged into Facebook and searching, but I don’t know if they are using the things you write about on Bing to personalize the results that you see.
On a much more macro level, I feel like Google Search + Your World or whatever they’re calling this whole phenomenon, is going to move us more toward tunnel vision. Rather than exposing people to what (debatably) objectively are the “best” or “most relevant” answers to their query, people would instead be exposed only (or primarily) to things influenced by their social connections. The inherent problem there is that we become less “well-rounded” for lack of a better term. A timely example of where this could go all wrong is an election year and its resulting flood of political content — only exposing people to things their friends and family like/+1/share vs. what everyone else and non-Google users get easy access to. It’s late and I’m not articulating well, but you get the idea.
Excellent read as usual Bill. I have become more focused on trying to use Google + everyday as I know that our profiles are eventually going to become a key component in search.
Hi Bill
I think that’s a very good idea. I’d definitely suggest not doing that at the exclusion of other activities though, such as blogging and adding informational content to your site in addition to the transactional pages that you have, or working with other social networks such as Twitter and Facebook to also build relationships and build a positive reputation for your expertise involving the areas that you provide services in. You want to make sure that you have a good integrated online marketing plan that encompasses multiple channels if possible, with a lot of emphasis on your own web pages, since you have the most control over them.
I have also began to use Google+ on a more daily basis recently. It has been somewhat in the shadows compared to other sites but I believe Google+ will become necessary for companies in the near future. Thanks for the helpful information.
I think personalization of results and realizing that your iterations/searches are being tracked are factors that push an increasing number of people to use private browsing.
I think it’s Vic Gundotra who said (not quoting him but trying to paraphrase what is said) that Google+ is the backbone of Google. Having said that, I’m not sure Google has already started pulling information from G+ as you stated above. Also, I thing in the future Google Now may use G+ results as well
Indeeed you’re right on this, but I can’t help but feel the same as Eliseo VN, it’s like I’m being tracked down on what I usually visit or search for, nevertheless this is the way to increase the relevancy of search engine results, relevent to each user and its personal preferrences so tracking down or making use of the information from Google plus is definately there.