Since attending the seminar, Intelligent Personal Assistant or IPA’s and search were everywhere I looked – on blogs, podcasts, even in my inbox! As Search (SEO as well as PPC) is one of the main services I offer my clients, I felt it was in my best interests to stay one step ahead of the curve and swot up on the effect IPA’s will have on search and as a result search engine optimisation. Please read below my key learnings and how we can utilise insights from this new search landscape in our search marketing.
2016 is the Year of the Intelligent Personal Assistant
In recent years, the world has been introduced to a number of Intelligent Personal Assistants including; Siri, Google Now, Cortana, Facebook M and Hound.
As mobile searches have increased, so have the use of IPA’s. A recent report from GlobalWebIndex confirmed that 20% internet users now use voice search on mobiles.
The rapid and continued rise in voice searches can be seen using Google trends (UK). Searches that do not return a search results page are still a search. Searches for call mum or dad would be almost nonsense in a web search and therefore can predominantly be attributed to the rise in voice search.
Voice Search vs. Keyword Search
The way we search using an IPA differs to how we use traditional search engines. The obvious difference is that voice search queries are spoken and keyword queries are typed. Being spoken, voice search queries tend to be more complex and often include a second refining search query before the desired result is achieved. Over the years, we have been trained to use specific language within search engines, whilst most of us would not think twice about entering “weather London” as a keyword search, very few would use that as a voice search.
Hundreds of different voice search queries featuring a multitude of different keywords could lead a user to the end result of the weather forecast in London. Although the keyword content of these queries will differ, they will all share a similar context. Results for voice search queries are populated by interpreting the context behind a search not just the keywords.
According to Search Engine Watch, while voice searches cannot currently be isolated in results reporting, they saw a 61% YOY rise in appearance of the below keywords in search phrases (Q4 2014). This is interesting as some of the largest growth areas are from keywords which are most likely from Voice searches and can be addressed with a very direct answer eg Where can I buy product X, or How much will it cost me etc.
This demonstrates that with the increased use of IPA’s we are seeing a shift away from web search to direct answer search. IPA’s don’t want to spit out a list of websites, they want to take you to straight to the most relevant answer, be that within a website an app or by carrying out an action. Emphasis is now being placed not only on interpreting the explicit meaning of a search query provided by the keywords but also the implicit context, who is searching, why are they searching for this etc. Because of this shift away from the traditional web search, SEO as we know it is about to change.
Search is moving towards context rather than content
Since the release of Hummingbird and RankBrain, Google has been changing the way they interpret search results. They want to understand the user’s intent when searching in order to populate the best possible answer to their query. By understanding not just the content of the search query but also the context, it becomes much easier to understand exactly what information the user is looking for and what search results would answer their query.
Google found that over half of search queries belonged to contexts which lasted over a month. For example, if you are planning a wedding, it is unlikely that you will perform just one search on one day. Instead you will perform many different searches over a period of time each linking to the context of planning a wedding. When only interpreting the explicit meaning of the keywords, your searches seem completely unrelated e.g. Caterers in London, Petite Wedding Dress, Make-up artist London. However, when the implicit meaning of the searches is interpreted, we can see they all steps in planning a wedding.
By modelling these user contexts using AI, Google hopes to move towards predicative search results whereby they know what you’ll be searching for before you’ve even searched for it! It is this contextual model which us SEO’s need to tap into in order to reach the top of the voice search results.
5 Considerations for Optimising your Content for Voice Search
- Recognise the growing importance of consumer context: While keyword searches still dominate the search landscape, with Google placing more and more emphasis on context and the rise in the use of IPA’s, I believe the focus of SEO will expand from explicit search query to implicit context. So we need to focus on user context and intent rather than on the keyword content of the search.
- Keywords are still very important: As Google’s AI becomes more tuned to language and search queries, SEOs will no longer have to fret about synonyms and long tail keywords as the machine will one day be smart enough to recognise that a webpage about ‘properties for sale’ is relevant to search query about ‘houses for sale’. However, as with all machine learning, there will be some exceptions where the machine just simply can’t grasp the relationship between two entities – this is where keyword focused SEO will come into play.
- Understand your data: Assess which contexts you are performing well in – knowledge of this will be increasingly useful in the future contextually focused world of SEO. When looking at the list of keywords your site is ranking for and the traffic you are getting from these, think about the contexts likely to be associated with these.
- Move away from the buyer persona and onto the context persona: Think about which search contexts rather than keywords you want your website to appear for. Identify user’s intent when visiting a particular area of your website and optimise your content to best serve these.
- Rethink the way you report: As search results become more contextual, keyword rankings will mean less and less and keyword ranking position reports won’t do your SEO work justice. Thin about ways you can report based on the contexts you appear for rather than the keywords.
What’s important to take away from this article is that keyword optimisation is by no means dead it’s just not enough. As search takes on different forms such as voice search through an IPA, search optimisation must also adapt to be inclusive of both the old and new technologies. As Google shift their focus over to context, I believe it is wise that SEO’s follow suit and broaden their optimisation horizons.