Voice search has become increasingly popular and is likely going to continue to grow for the foreseeable future. Voice assistants have been embedded into our daily lives, helping us streamline tasks and processes. While using our voices to get information is a huge convenience, implementing a strategy to optimize the retrieval of information from voice assistants is largely unknown. That’s where SEO (search engine optimization) for voice comes into play. In the early stages of 2019, there has been huge interest in the topic and Google has come forward addressing certain methods into how to make your content eligible for rich results and search features on Google. Google’s developer documentation about voice search SEO was a big breakthrough, as it was some of the first official, concrete advice given to SEOs and site owners about how to optimize for voice search. This article will just scratch the surface and act as a beginner’s guide to SEO service for voice, mostly from the Google search engine.
Importance of SEO for Voice Assistants
Voice interaction and voice search are not a new concept; research in this field dates back to as early as 1990. However, despite great leaps in research and development, there was no widespread commercial adoption of the technology until the last decade. The first iPhone, released in 2007, included a voice dialing feature. In that same year, Google released their iOS app which provided voice search capabilities. But it wasn’t until the release of virtual personal assistants on smartphones such as Siri on the iPhone 4s that the use of voice search and interaction began to take off. In the beginning, voice search was clumsy and often produced inaccurate results. But as accuracy improved and people found it more convenient to type on mobile devices, consumers began to widely adopt voice search and interaction. Moving forward, it is safe to say that voice interaction will become a staple of mobile and home automation SEO. Since the inception of the personal virtual assistants, there has been fierce competition between companies providing them. In addition to Apple’s Siri, Google Now and Microsoft Cortana were released in 2014. Amazon also joined the fray with the release of the Amazon Echo, and its voice assistant Alexa, in the same year. All of these virtual assistants have their own unique features, and different methods of accessing information. But they all share in common the use of voice interaction, and an ability to provide a user with direct answers to their questions. This is in contrast to the traditional search engines which often provide a list of links to webpages which may or may not contain the information sought by a user.
Overview of Siri, Alexa, and Google Assistant
Each of the three voice assistants has unique features; understanding those features is crucial for optimizing for voice search. Siri is a virtual assistant with a voice-controlled natural language interface that uses sequential inference and is capable of learning. It is an intelligent software that is predicted on a user’s specific question or request and it can change over time and customize its behavior. For SEO specifically, Siri is a more direct product of search and SEO strategy as the iPhone market encompasses a large part of the search topic for SEO. Siri uses the default search engine for the corresponding Apple product and it doesn’t have to be always Safari, but there is no way to change the default search engine to anything other than Bing. Siri runs off of structured data from websites and it provides an “answer box” that it sources the information from. This can be seen as a massive opportunity for SEO, but it may also take away organic traffic from a site and keep the user within the interface. Siri is at times considered quite “clunky” as it may not often understand the user’s voice, but with Apple constantly updating its functionality, this is very likely to change. Alexa, an intelligent personal assistant, is a voice service that hosts a plethora of skills and abilities that are activated by specific keywords. It is capable of adapting to the user’s speech patterns, vocabulary and personal preferences and has over time. Alexa has a large potential to take over traditional SEO. It uses SERPs to find its answer and then goes to the featured snippet for a result. Much of what SEOs do now, may have to change in order to match content that is tailored for the spoken word. This will also be heavily backed up by Google’s usage of voice search too and the progression of SEO with structured data and featured snippet optimization. Given that Amazon has no default search engine, it is unclear as to how SEO for Alexa will be measured and which data will be collected.
Understanding Voice Assistant SEO
In addition to this, users are far less likely to go through multiple steps to get the right answer with voice search. If the device does not provide the information the user is looking for, they are more likely to give up as opposed to trying the search again with different wording. Approximately 60% of users will not do another voice search if they do not get the answer they are looking for on the first try. This too can potentially lead to confusion if the device has selected what it believes to be the single number one result, but is actually only one of a number of possible good answers.
With voice search, things are a little bit different. Rather than provide users with a list of options, voice search selects what it believes to be the best single result. Users are then only provided with that number one result. If there are a number of different answers to a particular query, things can get a little bit confusing.
Voice Assistant SEO is very different from traditional search engine optimization. Understanding these differences is essential in achieving a high ranking on all voice search devices. With traditional SEO, web results are displayed in a sequential order much like they are in a list. By improving SEO, you can boost your position on that list, making your website more accessible to viewers.
Differences between Voice Assistant SEO and Traditional SEO
SEO is no longer confined to the screen, and the evolution of voice assistants raises the question of how different types of searches will affect the nature of SEO. In the US, 20% of mobile searches are voice searches, a number that is expected to rise to 50% by 2020. Voice searches have a strong local intent; according to Google, 40% of adults now use voice search once per day, and mobile voice searches are 3 times more likely to be locally based than text searches. 70% of those mobile voice searches are inquiries about store directions or other information. Google voice searches are 6 times more likely to be action queries than type searches, and this pattern is likely to follow for other voice assistants. Location-based businesses should take note of these statistics and make efforts to prepare for voice searches by creating content that answers an implied question, providing comprehensive information about a topic, and ensuring that their business information is complete and accurate. This later point is especially important, since voice searches are likely to source answers from the single best source, which often is drawn from Google My Business data. One of the great differences with voice assistants is the lack of a visual interface. For a voice assistant, a search only produces a single result. Not being the first result is tantamount to being invisible, since they will not read off more than the first result. Devices with screens such as smartphones and smart displays may offer some hope of better visibility in the future, but for now these provide additional ads space above organic results. This can be either good or bad; for a relevant result, showing up first is more valuable when there are no paid ad results, but for any other result, showing up behind 3-4 ads may actually give the user a higher visibility position. This will depend on user patterns and is thus an area that will need future testing and analysis.
Voice Assistant Ranking Factors
Local: Many voice queries have local intent, so serving results that are perceived as accurate but are located far away can be a recipe for losing faith in a voice assistant. If you’ve ever attempted to get Alexa to find your favorite nearby Chinese restaurant, you know what I’m talking about. This means local SEO may be even more important in voice search.
Relevancy: Understandably, Google places a high value on returning correct answers to voice queries, citing this as a reason for their reluctance to allow sponsored results. Duane Forrester, formerly of Bing, stated in an interview that Bing is looking for “skill like results” from SEOs in voice search. This means being the best resource for the given topic.
When discussing the “how?” of voice assistant optimization and improving one’s SERP, it helps to start with a breakdown of the most influential ranking factors, as revealed by various voices of authority at each of the major platforms. As of yet, no platform has come out with an official statement confirming the details of their voice algorithm, but employees of each company have strongly hinted at various ranking factors.
Voice Assistant Optimization Strategies
If the voice search keyword in question that one is targeting has an associated rich or featured snippet in the SERP, it seems that today’s best practice is to create content that has an answer matching or closely mirroring the one in the featured snippet. Backlinko’s report indicates that 40.7% of answer box results comprised of a featured snippet. Featured snippet optimization should help content creators accomplish the new goal of having their content read by voice assistants when delivering search results to answer questions posed by the user.
Content creation and schema markup will be the most effective lever to pull when trying to rank on page results through answers or skills for voice assistants. In terms of content creation, the goal should be to provide quick, direct answers to the questions users are asking. Backlinko’s report suggests that the typical voice search result is only 29 words in length. This might sound short, but it is in line with SEO considerations that the best Google search results are the ones with 40-50 words. Writing voice search answers with the classic inverted pyramid style: in which the broad who, what, where, why, and how questions are all answered in the 1st or 2nd sentence followed by any necessary follow up information is a good idea when targeting Google Assistant. When considering the brevity of Siri’s voice search answers or Amazon Alexa’s flash briefing, this may be a good time to repurpose Twitter’s tip for better tweet performance and consider trying to craft these 29 word answers in the form of a question.
There are a number of ways that digital marketers can optimize for voice assistants. Many of these strategies are similar to traditional SEO strategies: the overarching goal is to have content that is high-quality, easily accessible, and matches what the consumer wants. The difference lies in the execution.
Optimizing for Siri
Take a conversational tone a step further with a Siri-compatible app. Apps hold a large priority in Siri search, as the AI aims to provide more app-based responses than automated web ones. An effective app in the form of a chatbot or virtual assistant, which can provide intelligent answers to a wide range of questions, will require app store SEO but can be very effective at the long tail end of search.
Siri is also a very context-heavy form of search. About a third of queries are follow-ups to previous questions and are often delivered in a conversational tone. This is what separates Siri from other forms of search. It is designed to be a virtual assistant that can hold a casual conversation or respond to commands. Adapting content to resemble the relaxed tone of speech and providing answers in simple/effective language will increase the chance of that content being recommended.
Results are derived from a special set of search algorithms that determine the most intelligent and useful response. With Siri, if the answer is read from the featured snippet, there will be no additional search results. Therefore, on-page SEO such as rank tracking and meta data become less important. Step one to Siri SEO is optimizing for content that will appear in a featured snippet. Using clear and concise language, provide an intelligent solution to a problem or a detailed explanation to a question. Develop rich content that encompasses a wide range of topics in a particular field. Think about what the user will find most useful and the questions people might ask when looking for it.
Each virtual assistant has its own unique functions and features. For the purpose of this report, we will hone in on the specific SEO strategies for each virtual assistant. Starting with Siri. Unlike traditional SEO, voice search relies on a very different set of variables to return results. Therefore, SEO for voice search is less about fetching and more about becoming the default/recommended result.
Voice search is one of the most rapidly expanding technology trends today. In the past year alone, 41% of adults and 55% of teens have used voice search on a daily basis. As voice search continues to increase in popularity and usage, it will be important for businesses to learn to adapt their SEO strategies in order to stay relevant within this medium.
Siri’s Unique Features and Capabilities
Siri’s rise to prominence as the voice assistant for iPhone users is indicative of its strong user interface and usability. Siri uses dialog systems that are able to perform complex tasks for the user and also present a tight integration with third-party apps. However, our discussion in this paper involves how Siri works as a search system and how it is optimized. A key feature of Siri is its ability to allow users to talk to it in natural language. This is a contrast to text-based input where users are restricted by having to type keywords for searches. Siri’s natural language input allows users to speak queries in the same way as they would with another person. This means that users do not have to think about translating a thought into a series of keywords and can instead directly ask Siri what they want. Another interesting thing about Siri is its response to queries. McTear et al. provided an analysis of responses that various voice agents allow and chose the example of asking for the current weather in Tampa, Florida. For text-based search systems, this would typically result in a list of links to weather websites. With conventional search systems, it may take 3 or 4 more clicks and selection of various other links before finding the desired information. Conversely, Siri is able to provide a spoken response with only a few seconds of processing. This allows the user to change the query or request more specific information without having to go through multiple iterations of searches and digging through links. Finally, Siri often retrieves information from a narrow vertical of sources, particularly those associated with the task that it is trying to perform. For example, Siri may retrieve movie reviews from movie-related apps, or the aforementioned weather example may retrieve from a weather site. This stands in contrast to general keyword-based search systems, where the results often cull information from various different sources and assume that the user can filter through and find the best information. By taking data from specific sources, Siri can today make search information tasks more efficient for the user.
Siri-specific SEO Techniques
Eye tracking studies were used as a method of analysis for consumer behavior when interacting with a SERP. The study, measuring time spent on SERPs or “Google Goggles,” gave insight into the cognitive burden of having to pick through results. An average of 2 seconds was given to each result before moving on to the next. This shows that user patience for finding information is relatively low and that a direct result is the most effective way of satisfying the user. This is similar to a Siri query where there is only one chance at providing the right information. A single error could lead to a failed result and no chance of correcting the user to the right information.
SEO is similar to creating the best possible pattern recognition and matching the user’s intent to your content. Siri, compared to search engines such as Google, has a much higher focus on single responses for a given query. Siri also often uses more complex APIs and algorithms compared to conventional search engines. The search engine results page (SERP) is becoming less relevant with many queries leading to a direct answer scraped from a relevant site. Google’s current Hummingbird update is also geared towards providing a user with the best possible answer to their query. This means that Siri and Google are becoming more similar in terms of content delivery. Crafting good content specifically catered for the query is essential for both SEO and Siri optimization.
Understanding how Siri queries and delivers results is fundamental in optimizing for Siri. For a specific query, Siri takes a user’s intent and compares it with a statistical language model. This model allows Siri to map the words in the query to an abstract representation of the meaning. Siri then looks for specific attributes or markers that help give the query some direction. Once the meaning has been derived and a general direction has been established, Siri then tries to narrow down the query to a set of constraints using pattern recognition and match this against the potential results. This will allow for the final step of retrieving the information and generating a response. Understanding this process will give insight into what type of queries Siri will be used for and how to match user intent with content.
Best Practices for Siri Optimization
Relevant and authoritative links: Just like with web search results, having authoritative inbound and outbound links plays a significant role in search rankings. Links from pages with related content and high page ranks are particularly valuable.
Website load time: How fast Siri can retrieve information from a site is a relevant element in search rankings. More so than with standard web searches, page load time will play a role in optimization.
Structured data: A well-defined hierarchy and clear nomenclature should be established for headers, as Siri often looks for answers in headers. Pages should also contain metadata and be rich in content.
Localized content: Siri heavily favors localized search results. Each localized version of a page will score its own search rankings. Therefore, it is a good idea to make locally targeted pages unique.
Clear and concise content: Siri values concise content and searches for direct answers to user inquiries. It is important that the content of your web page is easy to understand and free of spelling errors. The content should also contain answers to possible questions users might have. If a user cannot find out what they are looking for in the first few seconds, they will click the back button and look elsewhere.
The following are best practices for Siri specific optimization, taken from Speech Interaction’s white paper on optimizing for voice controlled applications.
Case Studies: Successful Siri Optimization Examples
Coming in at next to no traffic from Siri on step one and an unbeneficial recitation on step two, this is an example of a cookie recipe owner not wanting to get what he wished for.
Step two of this query brings forth a more shocking result. The user searches for the Nestle Tollhouse Cookie recipe, and according to Fishkin, in the Google environment he would want his website to rank high enough to outdo the high authority, high traffic websites and have the user click through to find his Nestle Tollhouse cookie recipe. With the virtual environment, Siri will read through the result and recite the contents directly from the page without a chance for the website owner to gain traffic.
The first example brought forth by Fishkin was the query “find me a recipe for chocolate chip cookies”. This was used to compare SERPs for a query on both desktop search and virtual search. According to Fishkin, with the desktop search, he would want his website to rank highly enough to get click-through traffic. With the query being in the form of a question, more often than not Siri and competitor virtual assistants will try to answer the question itself. This will not result in any kind of traffic to the site owner’s page.
In the absence of webmaster guidelines from Apple and the fact that Siri’s index is secretive and constantly changing, specific cases and examples of successful optimization in the Siri environment are hard to come by. Despite this, in February of 2016, Jennifer Slegg conducted an interview with the CEO of Moz.com, Rand Fishkin. In the interview, Rand outlines three specific queries that best exemplify how the optimization for virtual assistants can differ from traditional webpage SERPs.
Optimizing for Alexa and Google Assistant
SEO Tactics for Alexa and Google Assistant There has been very little information online with regards to SEO for Alexa and Google Home, despite it becoming much more of a popular topic. An article by Bryson Meunier suggests that the way Google Home will source its results will be based on featured snippets from articles or other spoken results from the website with high authority on the topic. This is similar to the way Google sources its normal search results, placing an emphasis on SEO research and possibly a method of quick link building to ensure that your article is the one that is used as a source. High authority websites may also receive more traffic due to Google Home sourcing info from them. This serves to make SEO in general much more important if it was not already.
Alexa and Google Assistant are voice-activated AI-powered virtual assistants developed by Amazon and Google respectively. Both of these tech giants have a strong presence in the market share and have very similar functions in terms of what they offer to their users. Google has Google Assistant running on all devices that operate on Android 7.0 Nougat or above and also has the AI built into Google Allo, its smart messaging app. Google Home is a smart speaker developed by Google that also operates on Google Assistant. Similarly, Amazon’s Alexa is also a virtual assistant that operates on Amazon Echo, which is a smart speaker. These smart speakers are becoming increasingly popular as of late, resulting in much of their usage boiling down to simply voice-activated commands. This means that SEO for Alexa and Google Home is becoming much more important as it is such devices that these AI will source their spoken results from.
Alexa and Google Assistant’s Role in Voice Search
While Alexa and Google Assistant are active in helping voice search to grow much more convenient, there are some issues that still need to be resolved to make voice search an even more appealing option for consumers. For instance, voice searches do not always match users with the products or services they need. Solutions for search results in relation to specific types of businesses, including optimization for local businesses, have not yet been developed. To further attract local businesses to invest in SEO for voice search, solutions for tracking ROI with respect to voice search are essential. Additionally, many voice searches do not lead to a website visit, and even when they do, there is currently no way to tell if the visit came from a voice search. Results in voice search often provide information in the form of an answer to the user’s question, but neglect to tell the user the source of that information. This results in decreased brand awareness for businesses, since the user is unable to identify the source of their information. Decreased brand awareness has caused some businesses to pull out of SEO for voice search. At present, products and services from Google’s rich search results are required. However, once there is a lot more content in the form of voice-focused answers, it is likely that users will be satisfied with the information received through voice search and may not proceed to using a web-based search. This means that sites providing the information may lose traffic. In spite of this, Google’s data still shows that the number of people who do research on a product or service through voice search is greatly increasing. This suggests that content in the form of voice-focused answers will act as a replacement for the information that comes from web-based search results, even though it does not lead to a site visit. Step by step schema changes have hinted towards compatibility to resolve this and bring users to revisit the website to find the information answered by their question.
SEO Tactics for Alexa and Google Assistant
Differentiating strategy between Google Assistant and Alexa can be seen in their output sources. 82.1% of Google’s answers came from the Knowledge Graph, while 24.5% of Amazon’s responses came from Alexa’s own built-in capabilities. This means it is useful to consider using on-SERP SEO tactics to increase visibility on Google, and on-site tactics to increase visibility within Amazon’s own platform. Targeting specific on-SERP features which are the source of answers for Google queries is a good tactic. A study by Ahrefs has shown that featured snippets, paragraph snippets, list snippets, and table snippets are the most common types of search feature box, and targeting these snippet boxes is most likely to increase traffic to your webpage. Amazon has a similar product called ‘Alexa skills’ which can offer audio responses to customer queries, reading specific info and change tone/wording based on directions. Optimizing for Alexa would mean trying to get your audio replies featured during relevant user interactions. Audio replies are usually adapted from specific pages such as blog posts. Create a page that contains article information you want Alexa to read from, then create a skill within your Amazon developer account, connecting the skill to the webpage using its URL. This would mean using similar on-site SEO tactics to maximize visibility of the page. Although a relatively new concept, it could be speculated that taking ideas from traditional on-site SEO tactics and applying them to maximize page visibility within Amazon’s platform could be the future of SEO when dealing with private AI assistant outputs.
Voice Assistant SEO Tools and Resources
Both of these services work on the same basic principle. Users talk to their devices, give a command or ask a question, and the device responds with what the VPA considers to be the best possible answer or outcome. Usually, this is the top search result from a traditional text-based search. In any case, the endgame of the VPA is to be useful. If it fails often enough at that, users stop engaging with it. With no engagement, there are no voice searches, and no voice searches mean being left out of a rapidly growing search ecosystem. This is why SEOs and site owners should care about the VPA’s intent and how they can secure a spot at the top of its search results.
Siri isn’t the only virtual personal assistant (VPA) game in town. Windows’ Cortana has its own voice search service, and there’s an ever-growing list of third-party services whose primary or sole mission is to give users access to other services using nothing but their voice. But right now, there are only two major players in the third-party voice search market: Amazon with its Alexa-enabled devices, and Google with Google Assistant on Android devices and the Google app on iOS.
