We are moving towards the semantic web – an internet powered by machines that understand our meaning and our needs. But what does this mean in practise?
Entities and the Semantic Web
Most Google searches are not for words, they are for people, places and things – ‘entities’.
Most Google searches are not for words, they are for people, places and things – ‘entities’. For many years, Google has been building up massive entity databases and knowledge vaults to try to understand the entities behind your search and to know how to answer your query.
Over the last couple of years, they have started to put this into practise. When Google launched their knowledge graph in 2012, they labelled it a move from ‘strings to things’. Put more simply, Google is moving away from keyword search and entity-based search and results are taking over.
Given that Google understands who Alex Ferguson is, they can also link you to other football managers. So by understanding the meaning behind your search, they can not only answer your query more richly but they can also introduce you to other entities you may be interested in. The more Google advances, the more we as users will come to expect search results and navigation that understand and respond to our needs.
Led by Google, the charge to the semantic web is underway and it presents new challenges and opportunities for growth. Now is the time to be shaping every area of your business for a world dominated by entities. In this piece, we explore some simple and effective strategies for the semantic web using 4 case studies of businesses that are already enjoying dynamic growth.
Shaping your Content
To build its vast knowledge base for each entity, Google needs reliable data sources for every entity. So as well as ranking web sites and web pages generically, Google needs to rank web sites on expertise for specific entities.
Google needs to rank web sites on expertise for specific entities.
So it doesn’t matter whether you’re talking ketchup and mustard, Beyoncé and Jay-Z or Jaguar and Chevrolet – you want Google to see you as an expert on the entities that are central to your business.
How do you do this?
In the real world, we label people experts on a subject when they know all the relevant details and they can offer useful insight too. The core audience for any subject defines what is important. Digital is no different. Define your key entities. Then think about the audience for that entity – what must-have trivia doe they expect (like date-of-birth for a person or market cap for a company)? What can you offer them that is unique and valuable, perhaps like an exclusive interview or a forum debate on share price?
Then there’s another critical key step you can take to help Google recognise you as an expert on specific entities – structured data. By using the right mark-ups in your HTML you can label specific pieces of content to make them easy for search engines to recognise. As part of the semantic web collaborative movement, the major search engines have agreed a common vocabulary for these mark ups. This can be found at www.schema.org. More information on how to implement schema.org mark-ups can be found in this guide from Search Engine Land.
The search engines reward web sites that use structured data by injecting additional information directly into their search engine results. These additional data are known as ‘rich snippets’ and they lead to higher click through rates.
For more ideas on how to shape your content, we recommend watching this excellent 13 minute video by Gianluca Fiorelli.
The New York Times recently put together a digital innovation report that was later leaked which included the following quote:
“Recipes were never tagged by ingredients and cooking time. Because of that we floundered about for 15 years trying to figure out how to create a useful recipe database. Just adding structured data, for example, immediately increased traffic to our recipes from search engines by 52 percent.”
The New York Times has a wealth of digital content built up over years, much of which is relevant, in-depth and written by experts in their field. By using schema.org mark-ups they can communicate to search engines which entities they are writing about and the most important information on the page.
Similarly, by clearly marking up the author of the page – in this case, Martha Rose Shulman, who has been writing cookbooks for over 30 years – she too is identified as an important entity in her own right with a dedicated Google knowledge graph panel:
Implementing structured data is an essential SEO tactic.
Although Google frequently changes the kind of snippets it includes (for example it recently removed authorship mark ups), implementing structured data is an essential SEO tactic for any forward thinking digital business.
Search Metrics recently estimated that over a third of Google search results already incorporate Rich Snippets but only 0.3% of websites are currently making use of schema.org structured data.
- Define your most important entities
- For those entities, ask yourself what information is most important to visitors and make sure you are providing it (become an expert!)
- Add microdata mark-ups to your content
Develop your Database – Relationship mapping
In the content section, we already spoke about the need to build up key data for all your most important entities. The best way to do this is via a structured database.
When you’re modelling your database the other vital factor to consider is relationships. Relations are a vital part of semantic development. They help Google understand the difference between ‘Arsenal’ the football club and ‘arsenal’ the noun meaning a collection of weaponry. Entity relationships also help them offer a better experience with Knowledge Graph.
Map relationships between different entities in as many useful ways as possible.
So for a fashion site selling shoes, shoe attributes will include price, size, colour, brand and material. So of course your database needs to be modelled to include all these elements. But it should also be modelled to help you map relationships between different entities in as many useful ways as possible. So you may want to create relationships between different Italian Shoe makers or between different labels for a particular type of boot. By building up this intelligence, you too can create a structure for great user navigation and discovery.
When modelling your entity attributes and your entity relationships, pay attention to the schema.org structure. For example, for a person, the schema.org properties include information fields such as ‘birthDate’, ‘jobTitle’ and ‘gender’ and they also include other properties for relationships such as ‘spouse’, ‘relatedTo’, ‘worksFor’.
By including the schema.org properties in your own schema you will make it easy to include lots of structured data and you will be creating content that you know the search engines are hungry for. This will create fast and sustainable rewards.
Crunchbase is a website dedicated to tracking information on start-up companies and the people involved.
Looking at the example of Airbnb you can see the key attributes of status, founded, founders, etc. You can also see the entity relationships that have been mapped between the organisation entity (Airbnb) and the key related person entities (team members like founder, Brian Chesky).
This allows users to get the information they want on Airbnb, but also to quickly discover related companies and people, such as former employees, companies in similar industries or companies with shared investors.
So it’s a great experience for users and a great experience for search engines. In terms of organic search, you can see below the benefits of this structured data.
Not only do Crunchbase have the top result for ‘Airbnb investors’, they have four further links to profile pages, 5 links in one SERP!
- For each entity, define key attributes within the database
- Map relationships between entities within the database
- When defining entity attributes and relationships, mimic schema.org as closely as possible
Usability – design to encourage engagment and discovery
Users are coming to expect a better experience when searching for information.
As search engines and behemoths like Facebook and Twitter become more sophisticated in their knowledge bases and navigation options, users are coming to expect a better experience when searching for information. Why should they struggle to find content or products on your site when someone else can deliver relevant, useful information so easily?
If you follow the guides already set out for content and database development you will be in a great position to deliver a great user experience.
The design challenges are three-fold. Firstly, you must help users to navigate from one entity to other similar entities of interest. Secondly, you must make it easy to find more great content for any entity you’re already browsing. Finally you must use entities data to encourage new discovery. If you can crack these challenges, you are sure to delight your users and rapidly grow.
In the Artists section of MTV you can search their entity-database of artists. This experience starts with a cool search bar that activates as soon as you start typing.
It features panels for videos, interviews, latest tweets, photos, news, tour dates and a discography.
But where it becomes really exciting is how easy they make it to get more content you love using the ‘similar artists’ and biographical information links. These links take you through to artists that are musically similar and also to the hometown, genre and debut year of the artist.
This allows you to easily browse artists related to the original entity, but also to discover new bands or singers similar to them.
The layout is minimalist, but manages to fit in 100s of relevant internal links on each page by fully exploiting a database that is rich with relationships between entities.
The graph below shows estimated natural search traffic for MTV.com since ‘Artists’ was launched in beta in 2012. MTV has is enjoying almost triple the natural search traffic it was before the launch. A deeper analysis of the data shows that ‘Artists’ landing pages make up 55% of their organic search traffic. (source:Semrush.com)
- Help users quickly find other entities from the same genre
- Make it easy for your users to see more great content for an entity they are already reading about
- Use entity relationships to inspire users to discover new content they are interested in
Improve your Marketing with entity-based strategies
The semantic web also changes how you should think about marketing your web site. Previously you might have tried to boost rankings for certain keywords by attracting links to a particular page.
Semantic search requires you to think not just your domain authority and the authority of particular pages but also about your ‘entity authority’.
This starts with onsite optimisation. Think about how you can group your content to ‘belong’ to a particular entity rather than just a generic site section.
70% of all search queries are long-tail so the benefits of building up your entity authority can be huge.
Then start looking at building relationships with people and sites related to your key entities rather than just your subject matter as a whole. So for a technology site, this might mean building relationship with fans of the iPad to drive them to your iPad content section or bitcoin bloggers to your bitcoin forecasts.
By doing this you can increase your ‘entity authority’.
As semantic search evolves, search engines are giving more accurate long-tail results by recognising that a user is searching for a ‘thing’ not just a combination of keywords. That ‘thing’ is an entity and Google is increasingly rewarding the sites it views as experts on that entity. 70% of all search queries are long-tail so the benefits of building up your entity authority can be huge.
IMDb, the movie website is a perfect example of how entity authority can benefit a website. Movies are clearly defined entities, as are actors, writers and directors and they can all have clear relationships with one another.
On IMDb the key attributes for all these entities are well marked-up such as movie star ratings, reviews, release dates and production companies.
Over time they have also attracted amazing backlinks to dedicated hub pages for each entity.
As a result, its longtail search engine power is devastating. Take for example the movie ‘The Social Network’. All of the increasingly long-tailed searches below return the IMDb page for The Social Network movie, by recognising that users are looking for a particular entity.
So when you think about trying to emulate this success, try to visualise IMDb as an aggregated hub of thousands entities not just a movie database.
In the previous generation of relationship building and marketing, you may have drawn out a generic backlink mindmap for IMDb that looked something like this:
But in a world of entities, their back-link mindmap should something like this:
By visualising and targeting opportunities for specific entities or groups of entities, you can do better marketing and boost your ‘entity authority’. Doing this well leads to dynamic and sustainable growth in traffic.
- Neatly aggregate your content for each entity
- Evolve your marketing strategy to promote dedicated content for different entities
- Improve your reach by developing relationships, backlinks and citations for specific entities
The move towards the semantic web is well underway and gathering pace. For smart businesses it offers opportunities for immediate rewards and to get ahead of the curve a build a platform for sustainable success.