UPDATED with Google Authorship retirement & new rich snippets.
Semantic search, search based on context and intent rather than keyword relevancy, is not particularly new but many people are convinced that it will transform the way we search.
Here’s a current example of semantic search in action:
Google isn’t just looking at my keywords and sending me to a bunch of well-optimised weather websites. It is trying to understand the nature or context of my query and deliver the content I need from a variety of sources.
At the same time it is inferring that I may be interested in visiting London so it gives me additional content from Wikipedia as well as various points of interest.
It can do this for two reasons: 1) better AI allowing the search engine to “understand” the meaning of content and the context of the searcher’s intent, 2) developments in the kind of data that we provide to the search engines to help them understand the meaning of webpages, documents or objects. This is called structured data or structured mark-up.
Again, the concept of structured mark-up itself is not particularly new. You already know it as the ubiquitous meta data (meta titles & descriptions) that have long been a staple of basic SEO strategy. Structured mark-up is simply an expanded approach to meta data that provides a richer layer of machine readable context and meaning to a webpage.
This is exciting for its potential transformation (and disruption) on existing search technology and digital travel marketing (more on that below) but there are already a number of highly visible and rewarding applications for travel sites. These are called rich snippets and they’re all made possible through structured mark-up:
Google Authorship The ability to connect content with a particular author and display that author’s details in the search results comes from Authorship markup:
Authorship markup has been demonstrated to improve click through rates, although that may tail off if/when it reaches fuller adoption rates/saturation point. The really exciting prospect though is of AuthorRank becoming a ranking signal, something that is widely anticipated throughout the community.
Update 2.9.14: Google Authorship has since been retired, although the concept of connecting authors to their content and algorithmically understanding their expertise is probably here to stay.
“In Depth” articles
Introduced in 2013, In Depth article results do pretty much what it says on the tin; offer specialised, authoritative and detailed results for a particular query. For example, a query “rainforest” might yield the following block within the main results page:
Currently only the highest authority domains (see the examples above) are featured in these results although Google has offered support for others wishing to achieve In Depth results which suggests in theory it may become more widely available.
Aside from the obvious requirement for genuinely top-rate, definitive and authoritative content (no small task), there are some structured markups needed to facilitate In Depth results:
- Schema.org’s article markup
- A linked Google+ publisher page
- Properly implemented pagination tags for content spread over multiple pages
See this article from Moz for more.
You’ll have seen the gold stars under certain results, usually products or services that generate client reviews:
As with Authorship, this mark-up does not impact rankings but it does provide a much richer user-experience which makes it much more likely that users will click the results.
There are a number of ways to display additional product or offer information as a rich snippet in the search results. Structured mark-up exists for product name, description, price, etc:
There is also a mark-up vocabulary for events, which can be displayed as a rich snippet:
This can also be used for event listings such as concerts and festivals, as well as departure dates for particular tours or cruises.
Finally, there is a rich snippet for video content that will display the video embedded in the search result:
There are many other types of mark-up that are supported as rich snippets, see Google’s guide for the full list. To implement the mark-up you can use the Schema.org vocabulary that is supported by all the major search engines.
So, what’s the bigger picture?
So far these rich snippets are essentially “additions” to your existing search visibility. They don’t improve your rankings, they simply add more information to your existing results (which can have a massive impact on your click through rates.)
It gets really interesting when we look at the bigger picture and try to understand where the potential for transformation/disruption might be.
The underlying concept is that search technology is moving beyond viewing the relevancy of webpages based on the keywords of the content they contain, towards a point that the search engines can recognise items as actual things or objects in their own right, with a rich understanding of their meaning, context and relevancy to different queries.
In all the examples above Google is trying to identify the actual thing, not just the webpage or content that describes it. That includes people (via Authorship), products, events, locations, and many other actual, real world items.
Effectively Google is no longer satisfied with serving content that merely describes things, it wants to feature those things themselves in its search results. This is the key to semantic search – drawing connections between actual objects and understanding them as “things” in their own right.
So going back to our first example, while searching for the weather in London, the search engine would be able to infer the intent behind the search (visiting London) and also show you upcoming events, hotel deals, particular points of interest, activities, and all sorts of other things that could be relevant to your search, perhaps even questions that you haven’t asked yet.
Likewise, if you search for an upcoming event, Google could show you the weather plus driving directions or flight prices. If the event is in the next couple of days it could choose to show you ‘last minute’ hotel rates. Some of this data will be displayed embedded in the search result, but others will come from related searches that link off to other sites.
Think about the difference between this and current SEO strategy which still generally revolves around binary keyword targeting, i.e. making a certain document rank for a particular keyword. In the future, documents (or things) will be able to rank for multiple search queries, all depending on the intent of the user and how well the page has been marked-up.
We’re only really beginning to grasp the full potential of semantic search and it will be a slow, piecemeal process – especially for as long as mark-up adoption rates continue to languish. However, for early adopters there are two big opportunities:
- Short term, the ability to stand your SERP result out from the crowd with rich snippets
- Long term, the ability to exploit semantic search in your content strategy and create layers of content that are thoroughly marked-up to improve their connections to different users with countless different intents.