Search predictions come from:
– The terms you’re typing.
– What other people are searching for, including trending searches. Trending searches are popular stories in your area that change throughout the day. Trending searches aren’t related to your search history.
– Relevant searches you’ve done in the past (if you’re signed in to your Google Account and have Web & App Activity turned on).
Note: Search predictions aren’t the answer to your search, and they’re not statements by other people or Google about your search terms.
Visitors to a website may want to perform certain actions related to Entities (specific places or people or things) that are displayed to them on the Web.
For example, at a page for a restaurant (an entity), a person viewing the site may want to create a reservation or get driving directions to the restaurant from their current location. Doing those things may require a person to take a number of steps, such as selecting the name of the restaurant and copying it, pasting that information into a search box, and submitting it as a search query, selecting the site from search results, determining if making a reservation is possible on the site, and then providing information necessary to make a reservation; getting driving directions may also require multiple steps.
Using a touch screen device may potentially be even more difficult because the site would possibly then be limited to touch input.
A patent granted to Google this week describes a way to easily identify an entity such as a restaurant on a touch device, and select it online and take some action associated with that entity based upon the context of a site the entity is found upon. Actions such as booking a reservation at a restaurant found on a website, or procuring driving directions to that site, or other actions could be easily selected by the user of a site.
Google is organizing more and more things in its index based upon entity numbers. I have a couple of examples for you that show how they are being used.
It’s possible that you may have missed a reference to Freebase Entities in a Google Research Blog post from 2013. I missed it myself. The post is
Improving Photo Search: A Step Across the Semantic Gap.
In the post, the author (Chuck Rosenberg) tells us how they improve image searching at Google by labeling images with entities, rather than text strings. The entities they used are entities that you would find at a source such as Freebase. He tells us that they use Freebase Machine ID numbers for those labels:
As in ImageNet, the classes were not text strings, but are entities, in our case we use Freebase entities which form the basis of the Knowledge Graph used in Google search. An entity is a way to uniquely identify something in a language-independent way. In English when we encounter the word “jaguar”, it is hard to determine if it represents the animal or the car manufacturer. Entities assign a unique ID to each, removing that ambiguity, in this case “/m/0449p” for the former and “/m/012×34” for the latter.
A couple of months ago, I wrote about a Google patent that involved rewriting queries, titled Investigating Google RankBrain and Query Term Substitutions. There’s likely a lot more to how Google’s RankBrain approach works, but I came across a patent that seems to be related to the patent I wrote about in that post, and thought it was worth sharing and starting a discussion about. The patent I wrote about in that post was Using concepts as contexts for query term substitutions. The title for this new patent was very similar to that one (Synonym identification based on categorical contexts), and the more recent patent was granted on December 1st of this year.
The new patent starts off describing a scenario that is a good example of how it works. The inventors tell us:
One of the challenges of optimizing an e-commerce site that has lots of filtering and sorting options can be to try to create a click path through the site so that all the pages on the site that you want indexed by a search engine get crawled and indexed. This could require setting up the site so that some URLs are stopped from being crawled and indexed by use of the site’s robots.txt file, the use of parameter handling, with some pages having meta robots elements that are listed as being set as noindex.
If that kind of care isn’t performed on a site, a lot more URLs on the site might be crawled and indexed than there should be. I worked on one e-commerce site that offered around 3,000 products and category pages; and had around 40,000 pages indexed in Google that included versions of URLs from the site that included HTTP and HTTPS protocols, www and non-www subdomains, and many URLs that included sorting and filtering data parameters. After I reduced the site to a number of URLs that was closer to the number of products if offered, those pages ended up ranking better in search results.
I just returned from a few days in Las Vegas and the Pubcon Conference.
I had the chance to see some great presentations and talk to a number of interesting folks, and the company that I am the Director of Search Marketing at, Go Fish Digital won a US Search Award for Best Use of Search for Travel/Leisure, for a campaign we did for Reston Limo.
I wanted to share my presentation from the conference here as well.
A few years ago, I wrote the following about post about Google’s OneBox Patent Application I was brought back to it, with a new Google patent that looks at answering questions within similar answer boxes, and showing rich content, like in the example below:
A patent filed by Google a couple of years ago and granted today takes another look at Oneboxes, and includes this statement early on:
A search engine provider, Google Inc. of Mountain View, Calif., has developed an “answer box” technology, known as OneBox, that has been available for several years. Using this technology, a set of web search features are offered that provide a quick and easy way for a search engine to provide users with information that is relevant to, or that answers, their search query. For example, a search engine may respond to a search query regarding everyday essential information, reference tools, trip planning information, or other information by returning, as the first search result, information responsive to the search query, instead of providing a link and a snippet for each of a number of relevant web pages that may contain information.
Last year I wrote a post titled Google on Finding Entities: A Tale of Two Michael Jacksons. The post was about a Google patent that described how Google might tell different entities apart that shared the same name. The patent in it was filed in 2012 and granted in 2014. Google was also granted a new patent on disambiguating entities this week, which was originally filed in 2006. It is worth looking at this second one, given how important understanding entities is to Google.
It contains a pretty thoughtful approach to understanding and distinguishing between different entities within documents and queries.