How Google Is Using Reviews to Understand Your Business

Google is trying to understand local businesses better in order to provide valuable content and improve user experience. As the biggest name on the Internet, Google has greater access to information than anyone and more responsibility to use it well. For some time now, Google has been using information from reviews in order to expand their understanding of local content and increase visibility across the search arena. In August of last year, Google applied patent for a patent titled ’Analyzing user reviews to determine entity attributes‘ which details some of the ideas and architecture used in this process.

Local businesses and the knowledge graph

Google Search, Google Maps, and the Google Knowledge Graph are all central to the user experience. While Google’s extensive knowledge base is gathered from multiple sources, local content has always been a fertile testing ground for how Google defines places and promotes business visibility. While websites and local listings still have a key role to play, review information is more important than ever in how a business is seen and understood.

While Knowledge Graph infoboxes were added to Google’s search engine in 2012, they continue to expand their reach and scope. Historically, information from the graph has been limited to high authority sources such as the CIA World Factbook, Wikidata, and Wikipedia. This is all changing, however, with long tail information in reviews and third-party review sites also used to create an accurate picture of local businesses. While this decentralized knowledge base is consistent with Google’s philosophy, it is likely to create challenges for local businesses.

What are entities?

While Google’s index is far from transparent, the recent patent has given us some key insights into how the search engine processes information from reviews. Entity attributes play an integral role, with these key/value pairs defined by Google in order to make sense of the world. For example, one key/value pair may link a product with its creator, and another may link a group of people with a specific attribute such as ‘celebrity’. According to the patent, Google may be collecting attributes of entities directly from reviews. This is significant, with user reviews often containing information not provided by the entities themselves.

In an ongoing effort to make sense of the world around them, Google utilizes a number of techniques to describe, classify, and rank local business content. Individual user reviews are likely to play an increasingly important role in this process, with words, phrases, sentences, and paragraphs scanned in an effort to find key/value pairs related to a particular entity. While we’re unlikely to find out exactly how Google is using this descriptive text, natural language processing is sure to play a role. Depending on the place and language used, Google will attempt to extract, categorize, understand, and index this data.

From reviews to the index

Google may be using review information about entities in order to identify what people are interested in. The process is likely to involve identification strategies that recognize categories of observed user interest, detection strategies that link segments of text to these interests, and indexing in a searchable database based on these categories. Individual queries about an entity may be used to help identify the specific things people are interested in.

The index can count queries and create connections between queries and entities, with both frequency and relevance used to create a searchable database based on descriptive content. This is significant because it creates a connection between businesses and categories of observed user interest. Rather than leaving it up to businesses to define their own strengths and ignore their own weaknesses, this process puts the user and review process front and center of the search experience.

Where does Google look for reviews?

Google has the ability to look pretty much everywhere for user reviews, with no stone left unturned. Along with internal business reviews and those available on Google, user reviews are also likely to be gathered from social media postings, blogs, published online and print articles, and third-party websites. There are more specialized review websites out there than ever, along with online vendors and marketplaces dedicated to specific products and services. According to the patent, Google even considers email to be a valid place to find and dissect reviews.

How does Google categorize this content?

According to the patent, Google may implement different classification mechanisms to make sense of entity types and specific entities based on user descriptions from reviews. Navigational menus are likely to be used as proxies for content categories, which means site architecture is more important than ever before. As well as using website menus to see your content, Google may also use reviews to connect the dots between how your website is structured and what users are interested in.

In an ongoing effort to stay ahead of the curve, Google may also be using categories of observed interests to find out what people are interested in. For example, if a lot of users submit queries containing a certain word or phrase, this content is likely to become a category of observed interest over time. Along with the frequency of these queries, Google may utilize these interest lists based on localization parameters and specific search patterns. Queries deemed interesting to Google will then be used to categorize the review content.

Ranking and beyond

Review content will be used primarily to link particular user descriptions with specific businesses or other entities. For example, if the same words are used by multiple users to describe specific products or services, then strong associations will be set up between these reviews and the entity they are connected with. The frequency of terms is counted, associations are made, and categories are formed to expand the reach of searches linked to a business entity.

Negative reviews won’t punish a business directly, although the attributes connected with the review may become less strongly associated with the business over time. While conventional ranking also won’t be affected, the use of reviews to measure and associate content could have a big effect on relevance and overall visibility. 

As a business owner or marketing professional who wants to improve your brand’s visibility with the world’s biggest search engine, it’s important to get a lot of reviews in order to link your goods and services with multiple attributes. It’s also important to use numerous review services, and place third-party reviews on your website so Google can take advantage of your own navigation system. With Google actively searching for information that entities aren’t sharing themselves, being upfront and honest is perhaps the best advice of all.

This article originally appeared on The Local Lighthouse.