Recommended
Recommended
More Related Content
Similar to VOICE SEARCH: Boosting SEO in the age of Conversation
Similar to VOICE SEARCH: Boosting SEO in the age of Conversation (20)
Recently uploaded
Recently uploaded (20)
VOICE SEARCH: Boosting SEO in the age of Conversation
- 1. NATURAL DISCOVERY IN THE VOICE SEARCH ERA By Alex Laing Head of SEO, the7stars
- 2. WE ARE ON THE CUSP OF MASS ADOPTION Source: Gartner Hype Cycle
- 5. TWO DISTINCT BUT CONNECTED INTERACTIONS
- 6. INTERACTIONS DIFFER BY DEVICE
- 7. 25-34 YEAR OLDS ARE MOST LIKELY AGE TO USE DIGITAL ASSISTANTS 0% 10% 20% 30% 40% 50% 60% 18-24 25-34 35-44 Music Lovers Devices used to access digital assistants Tablet Phone Stand-alone speaker Other device • 29% of 18-24 year olds use a digital assistant through their phone x40% • 41% 25-34 year olds use digital assistant x18% • 13% 25-34 year olds use a stand-alone speaker x56% Source: YouGov Profiles 5th Feb 2018
- 8. SIRI PEAKS FOR 18-24 ALEXA AT 25-34 0 10 20 30 40 50 60 Nat Rep Music Lover 18-24 25-34 35-44 Digital Assistants Used Siri Amazon Alexa Google Assistant Cortana Facebook M Other Source: YouGov Profiles 5th Feb 2018
- 9. OF THOSE WHO USE DIGITAL ASSISTANTS: A QUARTER USE 2+ TIMES A DAY; A THIRD USE DAILY 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 35.00% 40.00% 10+ times a day 6-10 times a day 2-5 times a day Once a day 2-6 times a week Once a week 2-3 times a month Once a month Less than once a month Never
- 10. ALEXA AND SIRI ARE THE MOST HABITUAL 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 35.00% 40.00% 10+ times a day 6-10 times a day 2-5 times a day Once a day 2-6 times a week Once a week 2-3 times a month Once a month Less than once a month Never Frequency of using digital assistants based on users of each assistant Alexa Google Assistant Siri Facebook M
- 11. USAGE DIFFERS BY DEVICE 0% 10% 20% 30% 40% 50% 60% 70% 80% Don’t know To control other media devices (e.g. launch videos on TV) For smart home functions (e.g. turning off the lights,… Launching third-party actions (Alexa Skills, Actions on… To shop Other Making to-do lists Getting the latest news Making calendar appointments/reminders Checking traffic/travel planning (e.g. public transport… To call someone hands free To play music To check the web for an answer to something Checking the weather Speaker Phone
- 12. 52% WILLING TO PAY FOR MUSIC DOWNLOADS x18% DIGITAL ASSISTANT USERS ARE ALSO MUSIC STREAMERS 39% WILLING TO PAY FOR MUSIC VIA ONLINE STREAMING x40% 24% ARE UNLIKELY TO PAY FOR CDS x5% Source: YouGov Profiles 5th Feb 2018
- 13. RADIO STILL DOMINATES TIME SPENT WITH AUDIO ON ECHO Live Radio 71% Podcasts 2% Stremed Music Services 26% Catch-up Radio 1% SHARE OF WEEKLY LISTENING HOURS ON AMAZON ECHO Source: RadioPlayer
- 14. ECHO DRIVES MORE LISTENING HOURS FOR BOTH RADIO AND STREAMING Listen more often to radio, 71% Listen more often to streamed services 73% Source: RadioPlayer
- 15. THE OPPORTUNITY FOR MUSIC IS SIGNIFICANT “One of the primary use cases we had in mind when we invented Echo and Alexa was making the music streaming process in the home completely friction- free. If you make things easier, people do more of it.”
- 16. “Can you play me some sad country music from the 90s…”
- 17. “Play that song that goes ‘I wanna kick it with you all night’”
- 18. Search is Changing has Changed 2008 2018 Eliminate Duplicate Content Text-driven Voice-driven Penalise Spammy content & Links Emphasis Placed on Great Content Keywords Conversations
- 19. 2018 is the year of conversational search Smart home assistants are now responding to users’ needs, not users’ searches… “Italian Restaurant Kensington” “Tell me the best Italian restaurant in Kensington”
- 20. Devices are getting smarter Proactive suggestions Target voice searches through website mark-up Augmented through other sources like maps & YT
- 22. TEN BLUE LINKS CAN BE FOUND BUT THEY ARE A DYING BREED
- 23. Featured Snippets are Google’s “Best Bet”
- 24. Google Suggest learns from your previous searches & helps refine your search
- 25. Still, some valuable searches aren’t competitive…
- 26. SOME CHALLENGES…
- 27. MISTRUST IS BECOMING MORE PERVASIVE
- 28. RECOGNITION IS NOT LANGUAGE UNDERSTANDING
- 29. 1. Disambiguation DISAMBIGUATION "Like short sleeves, I bear arms"
- 30. 2. Query Refinement QUERY REFINEMENT
- 31. ANAPHORIC & CATAPHORIC RESOLUTION
- 32. LIMITED QUERY DATA FROM VOICE PROVIDERS
- 33. Ranking #1 is in vogue again RANKING #1 IS FRESH AGAIN
- 35. Answering questions is key to driving visibility
- 36. We can even control the narrative Target these with FAQ pages
- 37. Identify possible voice search questions
- 38. Get your apps found within search & store App Pack Natural Result
- 40. Leverage Structured Data Adds context & meaning Disambiguates Simplifies dev. Slims code volume
- 41. Create Better content 1. Start articles with a short summary then answer in depth 2. Q&A’s work well, as do lists 3. Mine forums, customer service, chat, email to build word clouds to aid KWR 4. Tables are a no-no for voice bots – instead, create content using well-formed paragraphs 5. GSP!
- 42. Don’t underestimate the importance of co-occurrence “You shall know a word by the company it keeps” - Firth
- 43. Other considerations •PageSpeed is a significant factor; voice search results typically come from faster-loading pages •Google relies heavily on very authoritative domains •Content that ranks well on desktop tends to rank in voice search •Content length - Google sources voice results from long-form content. •HTTPS is critical.
- 44. WHAT DOES THIS ALL MEAN?
- 45. REALITY CHECK: WE DON’T KNOW WHAT THE FUTURE WILL BRING
- 46. SEARCH PROS AREN’T BUYING IT (YET) Of PPC professionals are employing a dedicated voice strategy Of SEOs say voice is a ‘key area of focus’ in 2018
- 47. BUILDING A BASE OF FEATURED SNIPPETS WILL GROW VISIBILITY REGARDLESS OF VOICE
Editor's Notes
- I’m not an expert in the technologies that drive these amazing devices. I am merely interested in how we can leverage voice search to drive visibility for websites.
- Google say that for mass adoption voice recognition needs to be 99% accurate Currently this is looking at 95% accurate Google is better than Alexa at voice recognition However, while voice is being perfected, assistants and FB are honing their “conversation bot” system too – pulling together the background tech so that voice than replace typing eventually. Which will be a natural behaviour because… How Do Hype Cycles Work? Each Hype Cycle drills down into the five key phases of a technology's life cycle. Innovation Trigger: A potential technology breakthrough kicks things off. Early proof-of-concept stories and media interest trigger significant publicity. Often no usable products exist and commercial viability is unproven. Peak of Inflated Expectations: Early publicity produces a number of success stories — often accompanied by scores of failures. Some companies take action; many do not. Trough of Disillusionment: Interest wanes as experiments and implementations fail to deliver. Producers of the technology shake out or fail. Investments continue only if the surviving providers improve their products to the satisfaction of early adopters. Slope of Enlightenment: More instances of how the technology can benefit the enterprise start to crystallize and become more widely understood. Second- and third-generation products appear from technology providers. More enterprises fund pilots; conservative companies remain cautious. Plateau of Productivity: Mainstream adoption starts to take off. Criteria for assessing provider viability are more clearly defined. The technology's broad market applicability and relevance are clearly paying off.
- Currently 20% of mobile searches are via voice in 2 years, 50% of all search will be via voice. (google / comScore) 70% of requests are in natural language (google) – more on that later This is a massive shift in user behaviour, and will have a huge impact on the revenue model for these companies. Especially Google. For context, some figures that were out recently indicated that the global search industry is $92.4 billion (google have 92% share). This is more than: — paper and forest products, $81 billion — construction and engineering, $79 billion — real estate management and development, $76 billion — gas utilities, $58 billion If we are seeing a shift from the screen, what is it that they will be selling, as PPC model doesn’t really translate… *** ComScore forecasts that half of all searches will be voice searches by 2020 http://www.thedrum.com/opinion/2017/10/03/can-the-home-assistant-duopoly-be-broken
- People are Lazy and want to do what is easy. This is true across all marketing, media. Conversations allow us to get served relevant information without having to type or read – less of an effort. Furthermore voice is most natural way for us to communicate – as demonstrated in the Speak Easy study by JWT which showed that using voice required significantly less mental activity. Speech recognition error rates now match human parity at 5%, and are improving all the time, making this more. (Speak Easy) 20% of all mobile search are already via voice (Google)
- But it isn’t just a voice game. We still see that screen is hugely important here… Two types of request / interaction: 1) Ones completed on screen EG directions / news articles / instructions 2) Ones completed through the speaker EG listen to music / get updates et But as an industry we have been moving more and more towards screens, double screens. The rise of video etc seemed set to cement the screen, so how do speakers (only some of which include screens) fit in here?
- Music lovers are significant adopters of voice devices
- We can see Siri is the biggest but Alexa has a big slice of the 25-34 age group – the top adopters.
- Usage is rising. More people use their digital assistant 10+ times per day than people who never do. 1-5 times a day is the sweet spot which doesn’t provide much opportunity for brands to be discovered via voice. We need to be visible.
- Alexa and siri are used daily or more by two thirds of their users (vs 40% for Google and Facebook) But Facebook M seems to have a small loyalist following – although user numbers are low, 10% of usersuse 10+ times a day (!) Google is still used only once a week or less by a third of their users (Siri and Alexa are half this)
- Speakers are ACTUALLY used by more people for more tasks Music / smart home functions in house But also for organising journeys / the day
- Despite all of the other functions available, radio dominates listening on Echo accounting for 72% of all Echo time spent with audio entertainment – aided by the improved user experience delivered through the UK Radioplayer skill (i.e. the Amazon Echo equivalent of an app) Why might this be? Habit Laziness (don’t have to think of something to listen to) Enjoyment / company 25% more likely to list to 36+ hours radio a week Over index on commercial radio stations, rather than BBC.
- Ultimately – more speakers, which are easier to use, leads to more listening. How can we get more of a piece of the pie?
- This isn’t something which will be a surprise to those who designed them – Amazon have seen this as their second-pass and getting into the music industry… Bezos: The next gigantic growth area for the music industry is the home. Boom on discovery: I’ll give you some examples: We saw a couple of customers asking for music by their mood. Like, “Hey, can you play me some happy music? Or some sad music?” Then we saw people getting more micro. Like, “Can you play me sad country music from the ’90s?” Seen in the Echo Dot ad – “Play the song that….”
- This isn’t something which will be a surprise to those who designed them – Amazon have seen this as their second-pass and getting into the music industry… Bezos: The next gigantic growth area for the music industry is the home. Boom on discovery: I’ll give you some examples: We saw a couple of customers asking for music by their mood. Like, “Hey, can you play me some happy music? Or some sad music?” Then we saw people getting more micro. Like, “Can you play me sad country music from the ’90s?” Seen in the Echo Dot ad – “Play the song that….”
- This isn’t something which will be a surprise to those who designed them – Amazon have seen this as their second-pass and getting into the music industry… Bezos: The next gigantic growth area for the music industry is the home. Boom on discovery: I’ll give you some examples: We saw a couple of customers asking for music by their mood. Like, “Hey, can you play me some happy music? Or some sad music?” Then we saw people getting more micro. Like, “Can you play me sad country music from the ’90s?” Seen in the Echo Dot ad – “Play the song that….”
- Featured snippets are google’s attempt to provide the best answer instantly. However, we have seen that voice results are disproportionately answered by snippet results. You want to be visibile for voice? Optimise for s
- At this stage…. We don’t know! Amazon, Google and Facebook are being incredibly guarded. When researching this, it is very difficult to get solid answers for the. They are shrouded in mystery.
- Do people trust their home assistants? Many do. Some don’t. Should they? Maye the Amazon crew can shed some light on that later.
- A chatbot for example will take a query, generate a semantically appropriate response in the correct language. A voice bot takes the query and tries to match it to data to formulate a response. Because of this difference, the results can never be quite as accurate. It’s an issue of context.
- The best home assistants are able to deal with phrases that do not exactly match the ones we entered but are semantically and syntactically similar enough. However, they’re not perfect yet and at times will struggle to decide between various intents.
- Query refinement refers to the process of refining (changing or narrowing down) a search query. Search suggestions can reduce the need for query refinement within traditional search but this isn’t as advanced in voice search yet.
- Refers to the process of connecting future and past subjects in a conversation. User: Find me a place to eat near me. System: I found a nice steakhouse but there are two traffic jams on the way. User: Is it expensive? / How long are they? Bots are getting better but we still see significant issues with response quality when these challenges are encountered
- Google is the worst but there is not a lot of data out there on query response and “I didn’t get that” issues. We also have to be really clever about discovering what people are searching for using voice as there’s no out of the box tool that allows this at present.
- All those challenges sit with the tech providers – people much smarter than me! But this may be the most significant barrier to success using voice search – rankings are everything again.
- Structured data is a system of pairing names with values that helps search engines categorize and index your content. It orders the web better for Google and engines Schema and microdata which is compatible with HTML5, tell search engines “this is an image” or “this is the main headline” or “these are the reviews of this item”
- Co-occurrence is a concept which refers to the common presence, frequency of occurrence, and close proximity of similar keywords present across several websites. Understand this and work it into the content on your web properties. Create entities by getting your brands/stations/artisits associated with certain keywords online
- At this stage…. We don’t know! Amazon, Google and Facebook are being incredibly guarded. When researching this, it is very difficult to get solid answers for the. They are shrouded in mystery.
- What we do know is that these companies are all driven by innovation. They spend a huge amount of time honing and refining their products based on what they think will be more effective. They are also all kings at changing user behaviour through their changing platforms Some interesting introductions give us some clues Google have introduced left to right search results – is this to move us into a world where our ongoing chat with google goes up down? Facebook are experts in shifting user behaviour where they want to. The ALS challenge coincided (?) with the increased push by FB to swap photos for video. This is now common
- What we do know is that these companies are all driven by innovation. They spend a huge amount of time honing and refining their products based on what they think will be more effective. They are also all kings at changing user behaviour through their changing platforms Some interesting introductions give us some clues Google have introduced left to right search results – is this to move us into a world where our ongoing chat with google goes up down? Facebook are experts in shifting user behaviour where they want to. The ALS challenge coincided (?) with the increased push by FB to swap photos for video. This is now common
- For years the advertising industry has face a duopoly between Facebook and Google, but this is the year that everyone is looking on the horizon for the triopoly to emerge. Amazon’s foray into digital assistants has been a business critical move in this space, but also, in the music space too. Getting their Alexa into mainstream homes has given them an opportunity to become serious players in the music streaming space. Media wise, we need to acknowledge this and understand how we work with them to ride the wave of their growth
- Currently Spotify leads music streaming, but in this space, the tech players are squeezing them out. Apple are investing in shazam – what does this mean for their ambitions? As labels, with an investment in Spotify, will the tables turn in the power relationship?
- As mentioned, users are still listening more to radio than streaming on voice devices, and is this in part down to the “name an artist” syndrome . (i.e. when asked what to listen to your mind goes blank) How can we make sure that it is our artists who are the ones that are thought of first? Goes back to a basic marketing principle – what we need to drive is front of mind awareness. IE building the brand of an individual artist. This works easily enough when we are speaking about platinum selling record artists, the Beyoncé and Eds of the world. For development artists this is something we will need to consider when we are launching them. Not just making their song . Sound familiar, but making their name memorable too. But how does this translate to artists who we are used to marketing no a single by single basis? Do we end up having to rely on playlists for their streaming… which leads us onto:
- There has been quite a lot of talk about moving to ask for very niche genre requests. In their interview with Billboard, Boom talked about how they have already started to see a change here Boom: Definitely. When you have nothing to look at, it’s liberating. You’re not constrained by the technology — you’re only constrained by your imagination, and when you talk to Alexa, you ask for music in ways that would be difficult to do in a visual app. I’ll give you some examples: We saw a couple of customers asking for music by their mood. Like, “Hey, can you play me some happy music? Or some sad music?” Then we saw people getting more micro. Like, “Can you play me sad country music from the ’90s?” Now, if you think about how you would do that inside of an app, no one would ever ask that, right? They would go, “OK, I want to listen to U2 from the ’80s, so I’m going to type in U2, get to U2’s artist screen. OK, which albums are from the ’80s? OK, I’m going to create a new playlist, drag the songs...” Five minutes later, you’re listening to music. But this is five seconds. This is really exciting opportunity for users to discover, but also means that from a user POV they will need to know what genre / type of playlist they are looking for, How can we signpost and control these categorisations, rather than being
- Historically, most digital streaming has been done through headphones. This is something that the advertising players in this market have been keen to promote due to the “personal” nature of this experience, and the capabilities with 360 sound and so on. However, as streaming shifts more onto standalone speakers too – does this mean that what was previously a more personal playlist becomes shared. How do the streaming services continue to offer personalised and relevant content if they don’t know which family member is listening. Essentially, going back from the one person device (mobile phone) model to shared device model (i.e. desktop) and all the problems this faced. Netflix have managed this well with their welcome page, so will we see streaming services also need to acknowledge who is selecting music, so they can continue to use this data to personalise playlists. When we know that personalised playlists such as “Listen Weekly” have contributed to growth in listening time going up https://insights.spotify.com/us/2017/11/02/listening-diversity-spotify/
- It is easy and tempting to look at these trends and get pulled in by the “power of new” - what is the new, shiny, media first that we can take. But it is equally important to step back and think about how we can use the insights or changes which have become apparently through the “new” to sway how we use the “old” And specifically, in this space it is worth reflecting on how the dominance of voice assistants, in particular will allow us to use radio in a different way. For example - Radio X know that they get a spike in digital listening at 6.45 AM as they see digital alarms go off. Is it safe to assume that non digital listening also spikes then, and if so, how can we use these insights to plan our radio activity differently? Or, if we know that being able to ask for a specific artists is more important, how can we use our audio activity to signpost this to listeners? On DAX we can target audiences by specific device, so should we make this path to listening even easier for these audiences? This is especially important when you think of Spotify and the importance of the Home / Browse pages at driving discovery / inspiration. Perhaps this is why people continue to default to radio on these devices?