What is semantic search? Broadly speaking, it’s a term that refers to a move towards more accurate search results by using various methods to better understand the intent and context behind a search.
Or as Alexis Sanders very eloquently explained it on the Moz Blog,
“The word “semantic” refers to the meaning or essence of something. Applied to search, “semantics” essentially relates to the study of words and their logic. Semantic search seeks to improve search accuracy by understanding a searcher’s intent through contextual meaning. […] Semantic search brings about an enhanced understanding of searcher intent, the ability to extract answers, and delivers more personalized results.”
Google is constantly making tweaks and changes to its documentation and features linked to semantic search. Many of these involve things like structured data and Schema.org, rich results, Knowledge Graph and so on, and the vast majority go unannounced and unnoticed – even though they can make a significant difference to the way we interact with search.
But there are some eagle-eyed members of the search community who keep tabs on changes to semantic search, and let the rest of us know what’s up. To aid in those efforts, I’m rounding up five recent important changes to semantic search on Google that you might not have noticed.
100% of the credit for these observations goes to the Semantic Search Marketing Google+ group (and specifically its founder Aaron Bradley), which is my source for all the latest news and updates on semantic search. If you want to keep in the loop, I highly recommend joining.
Videos and recipes are now accessible via image search
Earlier this week, Google made a telling addition to its documentation for videos, specifying that video rich results will now display in image search on mobile devices, “providing users with useful information about your video.”
A mobile image search for a phrase like “Daily Show Youtube” (okay, that one’s probably not going to happen organically, but I wanted to make the feature work) will fetch video thumbnails in among the grid of regular image results, which when selected, unfold into something like this:
You then need to select “Watch” or the title of the video to be taken to the video itself. (Selecting the image will only bring up the image in fullscreen and won’t redirect you to the video). So far, video rich results from YouTube and Wistia have been spotted in image search.
Google’s documentation for recipes also now features a similar addition: “Rich results can also appear in image search on mobile devices, providing users with useful information about your recipe.” So now you can do more than just stare at a mouthwatering picture of a lasagna in image search – you might be able to find out how it’s made.
Google’s documentation gives instructions on how to mark up your videos and recipes correctly, so that you can make sure your content gets pulled through into image search.
Rich cards are no more
RIP, rich cards. The term introduced by Google in May 2016 to describe the, well, card-style rich results that appear for specific searches have now been removed from Google Developers.
As identified by Aaron Bradley, Google has made changes to its ‘Mark Up Your Content Items’ on Google Developers to remove reference to “rich cards”. In most places, these have been changed to refer to “rich results”, the family of results which includes things like rich cards, rich snippets and featured snippets.
There’s no information as to why Google decided to retire the term; I think it’s usefully descriptive, but maybe Google decided there was no point making an arbitrary distinction between a “card” and a “non-card” rich result.
It may also have been aiming to slim down the number of similar-sounding terms it uses to describe search results with the addition of “enriched search results” to the mix – more on that later.
Google launches structured data-powered job postings in search results
Google has added another item to the list of things that will trigger a rich result in search: job postings.
This change was prefigured by the addition of a Jobs tab to Google’s ‘Early Access and partner-only features’ page, which is another good place to keep an eye out for upcoming developments in search.
Google also hinted at the addition during this year’s Google I/O, when it announced the launch of a new initiative called ‘Google for Jobs’. In a lengthy blog post published on the first day of the conference, Google CEO Sundar Pichai explained the advent of Google for Jobs as forming part of Google’s overall efforts towards “democratizing access to information and surfacing new opportunities”, tying it in with Google’s advances in AI and machine learning.
“For example, almost half of U.S. employers say they still have issues filling open positions. Meanwhile, job seekers often don’t know there’s a job opening just around the corner from them, because the nature of job posts—high turnover, low traffic, inconsistency in job titles—have made them hard for search engines to classify. Through a new initiative, Google for Jobs, we hope to connect companies with potential employees, and help job seekers find new opportunities.”
The new feature, which is U.S.-only for the time being, is being presented as an “enriched search experience”, which is another one of Google’s interesting new additions to semantic search that I’ve explored in full below.
And in a neat tie-in, reviews of employers are now due to be added in schema.org 3.3, including both individual text reviews and aggregate ratings of organizations in their role as employer.
Google introduces new “enriched search results”
Move over rich results – Google’s got an even better experience now. Introducing “enriched search results”, a “more interactive and enhanced class of rich results” being made available across Google.
How long have enriched search results been around? SEO By the Sea blogged about a Google patent for enriched search results as far back as 2014, and followed up with a post in 2015 exploring ‘enriched resources’ in more detail.
However, in the 2014 post Bill Slawski specifically identifies things like airline flights, weather inquiries and sports scores as triggering an enriched result, whereas in its Search Console Help topic on enriched search results, Google specifies that this experience is linked to job postings, recipes and events only.
According to Google:
“Enriched search results often include an immersive popup experience or other advanced interaction feature.”
Google also specifies that “Enriched search enables the user to search across the various properties of a structured data item; for instance, a user might search for chicken soup recipes under 200 calories, or recipes that take less than 1 hour of preparation time.”
Judging by this quote, enriched search results are a continuation of Google’s overall strategy to achieve two things: interpret and respond to more in-depth search queries, and make the SERP more of a one-stop-shop for anything that a searcher could need.
We’ve seen Google increasingly add interactive features to the SERP like new types of rich result, and Google Posts, while also improving its ability to interpret user intent and search context. (Which, as we established earlier, is the goal of semantic search). So in the recipe example given above, a user would be able to search for chicken soup recipes with under 200 calories, then view and follow the recipe in a pop-up, all without needing to click through to a recipe website.
Needless to say, this could be bad news for website traffic and click-throughs – even more than featured snippets, answer boxes, the knowledge graph, quick answers and other rich results already are.
Google makes a whole host of changes to its structured data developer guides
Finally, Google has made a wide-ranging set of changes to its structured data developer guides. I recommend reading Aaron Bradley’s post to Semantic Search Marketing for full details, but here are some highlights:
- Guides are now classified as covering the following topics: structured data, AMP, mobile friendly design
- Structured data has a new definition: it is now defined by Google as “a standardized format for providing information about a page and classifying the page content.” The old definition called it “a text-based organization of data that is included in a file and served from the web.” This one definitely seems a little clearer.
- Twice as many items now listed under “Technical guidelines”, including an explanation of what to do about duplicate content
- There is now less emphasis on the Structured Data Testing Tool, and more on post-publication analysis and testing – perhaps Google is trying to get users to do more of their own work on structured data markup, rather than relying on Google’s tool?
- All content types are now eligible to appear in a carousel.
If you enjoyed this post, don’t miss Clark Boyd’s exploration of what semantic search means today in the wider context of the industry: ‘Semantic Search: What it means for SEO in 2017‘.