Reviews offering a ranking Boost?
In the Past, Google Has Hinted Reviews Might Be Used to Rank Results.
In 2013, I wrote a post about how Google might consider using requests for driving directions as something that might boost local search results at Google in the post, Driving Directions vs. Reviews as Ranking Signals for Google Maps. That post was based on a white paper describing a strong correlation between requests for driving directions and positive reviews for a business. The paper I had written about was, Hyper-Local, Directions-Based Ranking of Places (pdf). The post also looked at a Google patent named Directions-based ranking of places returned by local search queries.
Related Content:
In that patent, we were told that for business types where there just weren’t many requests for driving directions, Google might use other signs of popularity for those businesses, such as reviews. It was interesting to see that mentioned, however, it didn’t tell us more about how reviews might be used to boost local search results. (Reviews are something I have seen many people engaging in local search suggest as something Google uses in ranking local businesses, but without too much detail.) I was hoping to see something from Google that told us more.
This week, I finally saw a Google patent telling us more. The patent delves into how reviews might be used to boost local search rankings. If those reviews are from people Google recognizes as local experts on a business type in a certain geographic area, they may be used to boost local results.
Why Focus on Local Expert Reviews?
Like most patents, this one identifies a problem, and then a solution they have selected to solve that problem.
Here is the problem that looking at local expert reviews is intended to resolve:
Local search engines are generally used to obtain information about businesses and other points of interest related to some geographic area. For example, a user may submit a query for “Italian restaurants in New York,” and in response, a local search engine typically attempts to identify responsive content, such as websites of Italian restaurants in New York and reviews of such restaurants. Generally, however, local search queries implicate substantially more web resources than a user is interested in viewing (e.g., there may be several hundred websites of, or about, New York Italian restaurants), and identifying the subset of responsive web resources likely to warrant the user’s attention can be difficult.
The solution seems to be identifying local experts on a business type in a locality and the places they recommend:
Often, some users are particularly knowledgeable about certain topics relating to local search queries, for instance, some users are experts in a particular genre of restaurants, a particular city or neighborhood, a type of retailer, or a combination of these categories (e.g., the New York Italian restaurants responsive to the above-mentioned query). Such experts, however, can be difficult to identify, particularly at web-scale, as many users with relatively little expertise will hold forth on various topics; some users will self identify as experts to propagate spam for commercial benefit, and the number of local-search topics on which one may be an expert is very large.
Some of this local experts reviews patent include obtaining reviews of a reviewed category of businesses in a reviewed geographic area by reviewers; determining that reviews from more than a threshold number of reviewers have been obtained and, in response, identifying at least some of the reviewers as being experts in the reviewed category of businesses in the reviewed geographic area; receiving a query from a user; identifying a geographic area and a category of businesses for the query; ranking, with a processor, local search results responsive to the query based on the reviews of the experts in the reviewed geographic area and reviewed business category; and sending ranked search results to the user.
The local expert reviews patent from Google is this one:
Identifying local experts for local search
Inventors: John Alastair Hawkins and Cristina Stancu-Mara
Assignee: GOOGLE
US Patent: 9,792,330
Granted: October 17, 2017
Filed: April 30, 2013
Abstract
Provided is a process, including obtaining reviews of a reviewed category of businesses in a reviewed geographic area by reviewers; determining that reviews from more than a threshold number of reviewers have been obtained and, in response, identifying at least some of the reviewers as being experts in the reviewed category of businesses in the reviewed geographic area; receiving a query from a user; identifying a geographic area and a category of businesses for the query; ranking, with a processor, local search results responsive to the query based on the reviews of the experts in the reviewed geographic area and reviewed business category; and sending ranked search results to the user.
The patent provides many details on who a local expert might be. It also provides details on how to find them.
Local expert reviews would be checked to make sure that they are long enough to be adequate reviews. Google would use something like ngrams (as seen in a recent patent from Navneet Panda to identify site quality: Using Ngram Phrase Models to Generate Site Quality Scores) to tell whether reviews are spam or not, so reviews that are poorly written and grammatically incorrect may be seen as spam.
Someone who writes 34 reviews about “Italian Restaurants in New York City” might be seen as a local expert in New York City. Someone writing 10 reviews on chocolate stores throughout the United States might be seen as a local expert (in the United States) on stores that sell chocolate.
The patent says that many reviews aren’t trustworthy; which is also a problem it tries to solve:
Local search benefits from using web-hosted consumer reviews to rank businesses (e.g., business websites, websites about businesses, or other local business-related content), but many reviews are not trustworthy, either because the reviewer lacks expertise, or the reviewer has different preferences from the user submitting a local search query. This problem and others are mitigated by some embodiments of a web services system (shown in FIG. 1) that pre-processes reviews to identify experts and, then, uses the experts’ reviews to rank local search results.
The local expert reviews used by this patent appear to be reviews submitted to multiple sites. This way, they aren’t just reviews from Google+. Google would look at reviews, and filter them to remove spam. After spam is removed, the reviews are then categorized using hierarchical categories based upon geographical location and categories of business.
To provide local expert reviews that may be considered more valuable to a searcher, a level of personalization is involved in this process:
Further, some embodiments personalize the selection of experts by identifying those experts whose reviews are similar to those of the user from whom the local search query was received. And, some embodiments calculate an expertise score for each expert, a value which is used to rank reviews about a given business for presentation to a user after a user selects a corresponding local search result.
This patent’s description provides this summary of the process behind the patent in the described here:
For example, several thousand users may submit to a social networking service or a business directory reviews of various businesses in New York State. Embodiments obtain these reviews from the social networking service, business directory, or both, and filter out those reviews that are likely spam and those reviews that are less than a few words in length to identify a set of substantive user reviews. The remaining reviews are grouped by geographic area and category of the business, forming, for instance, a subset of reviews relating to Italian restaurants in New York City. A few users typically submit substantially more reviews in the subset than other users, for instance, one user may have submitted 35 different reviews about 35 different Italian restaurants in New York City, while most users may have submitted a single review. Those users with reviews in the area-category subset are ranked according to the number of reviews they have submitted in the subset, and a number (e.g., the top 10) are designated as experts on, for instance, Italian restaurants in New York City. Later, when a user submits a local search query requesting, for example, Italian restaurants near New York City, the same area-category is detected in the query; the corresponding top experts are retrieved from memory, and reviews by those experts are used to rank local search results identifying New York City Italian restaurants, thereby potentially presenting to the user some of the best Italian restaurants in New York City according to local experts.
There is the complexity involved in identifying local experts. The patent goes into the variations potentially involved, such as looking at starred ratings or votes some reviews are “helpful.” It’s worth reading the patent in detail to know more.
It is interesting seeing the kinds of things that Google might look at to rank local search results. Looking at local expert reviews from people identified as experts and boosting rankings from those people is interesting.
Search News Straight To Your Inbox
*Required
About Bill Slawski
With more than 26 years of SEO experience and a Juris Doctor Degree, Bill Slawski is the foremost expert on Google’s patents as related to SEO. Patent Exploration is one of the quickest and most detailed ways to find new information about SEO. Bill is the Editor of SEO by the Sea, a prominent search engine optimization blog, where he is the author of over 1,300 posts. Bill’s experience includes Fortune 500 brands and some of the largest websites in the world. Bill is a contributing author for Moz, Search Engine Land, and Search Engine Journal. In 2014-2021, he spoke at industry-leading international conferences about topics including search engine algorithms, universal and blended search, personalization in search, search and social, and duplicate content problems, structured data, and schemaJoin thousands of marketers to get the best search news in under 5 minutes. Get resources, tips and more with The Splash newsletter:
