In the age of mobile devices and virtual assistants, the future of SEO requires optimizing content for voice search. New devices such as Apple's Siri, Amazon Echo and Google Home are further accelerating the trend toward voice search. Marketers and merchants must prepare now for a future in which voice and virtual assistants play a much larger role in content discovery and conversions. This session explores how to optimise your content and user experience for a future in which half or more of all queries will be voice-driven.
Being a PMM with a multi-product portfolio - Product Marketing Summit
Optimising Content For Voice Search & Virtual Assistants
1. #SMX #21A3 @PeteCampbell
Pete Campbell, Managing Director, Kaizen
Optimising
Content for
Voice Search &
Virtual Assistants
2. #SMX #21A3 @PeteCampbell
Pete Campbell, Managing Director, Kaizen
Optimising
Content for
Voice Search &
Virtual Assistants
3. #SMX #21A3 @PeteCampbell
PETE CAMPBELL
/ KAIZEN
• Founder / Managing Director of Kaizen
• Content-led visibility powered by in-house tech
• Specialist team of Coders/Designers/Marketeers
• 2 award wins, 5 nominations in 2016/17
• Pete built his first website at 11, first grey hair at 28
• Running his 1st marathon this weekend (#brag)
• Proud owner of an Amazon Echo Dot
16. #SMX #21A3 @PeteCampbell
AMAZON’S BEST SELLING PRODUCT IN 2016
5M+ ECHO DEVICES SOLD SINCE LAUNCH
61%
92% 95%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Siri OK Google Alexa
UserSatisfactionbyDevice
Credit: MarketingLand & Credit Strategies 2016 Survey of 180 Early-Adoption Echo Users
17. #SMX #21A3 @PeteCampbell
User Provides
Query
Voice API
Converts to
Text
AI Performs
Query
Search Results
THE PROBLEM WITH VOICE SEARCH TODAY
18. #SMX #21A3 @PeteCampbell
X X User Types
Query
Search Results
User Provides
Query
Voice API
Converts to
Text
AI Performs
Query
Search Results
IS THAT IT’S NOT MUCH DIFFERENT FROM TYPING
19. #SMX #21A3 @PeteCampbell
User Provides
Query
Voice API
Converts to
Text
AI Performs
Query
Search Results
User Provides
Command
AI Structures
Request to 3rd
Party API
Action &
Response
X
VOICE ACTIONS ARE DIFFERENT. THEY ADD VALUE
X X User Types
Query
Search Results
“Order a Pizza” Brands API #diet
20. #SMX #21A3 @PeteCampbell
HOW OWNERS ARE USING ECHO
Credit: Experian & Credit Strategies 2016 Survey of 180 Early-Adoption Echo Users
22. #SMX #21A3 @PeteCampbell
“The destiny of [Google]
is to become the Star
Trek computer, and
that’s what we are
building”
Amit Singhal – Google SVP Search, 2013
23. #SMX #21A3 @PeteCampbell
“You could ask it a
question and it would tell
you exactly the right
answer, it would tell you
things you need to know
before you could ask it.”
Amit Singhal – Google SVP Search, 2013
41. #SMX #21A3 @PeteCampbellCredit: https://moz.com/blog/what-we-learned-analyzing-featured-snippets
STRUCTURE THE ANSWER TO
MATCH HOW SNIPPETS
APPEAR IN YOUR NICHE
49. #SMX #21A3 @PeteCampbell
1. Add Support for HTTP URLs
2. Use App Indexing API or SDK (iOS)
3. Test in Android Studio or Safari
4. Register in Google Search Console
5. Monitor via Search Analytics Report
HOW TO IMPLEMENT APP INDEXING
https://developers.google.com/app-indexing/android/app
http://apple.co/1UMgEkf
51. #SMX #21A3 @PeteCampbell
“SEARCH FOR [KEYWORD] ON [APP NAME]”
Intent Action Context
Google-Approved API
Commands Only
Credit: https://developers.google.com/voice-actions/system/
52. #SMX #21A3 @PeteCampbell
Use NSUserActivity to build index of App
Content & Search Activities
Eligible Activities Only
Appear on Siri
Credit: http://apple.co/1YilYuO
55. #SMX #21A3 @PeteCampbell
HOW ALEXA’S ‘CUSTOM SKILLS’ WORK
Credit: https://developer.amazon.com/public/solutions/alexa/alexa-skills-kit/getting-started-guide
Built with AWS Lambda or Web Service
56. #SMX #21A3 @PeteCampbell
HOW GOGGLE ASSISTANT’S “CONVERSATION ACTIONS” WORK
Credit: https://developers.google.com/actions/develop/conversation
Built with Node.JS & Conversation API
* PLAY VIDEO UNTIL 2min 37 SECONDS * “I left a message saying you had called” – “OK”
Siri is… terrible
the team trained an AI to play 49 different video games from the Atari 2600, a games console popular in the 1980s. The software wasn’t told the rules of the games, and instead had to watch the screen to come up with its own strategies to get a high score. It was able to beat a top human player in 23 of the games.
the team trained an AI to play 49 different video games from the Atari 2600, a games console popular in the 1980s. The software wasn’t told the rules of the games, and instead had to watch the screen to come up with its own strategies to get a high score. It was able to beat a top human player in 23 of the games.
This approach requires much less computational might. While the previous system required eight days of training on high-end GPUs to play Atari games, the new AI achieved better performance on more modest CPUs in just four days. With Atari well and truly beaten, the team moved on to other games. In a simple 3D racing game (see video below) it achieved 90 per cent of a human tester’s score.
the team trained an AI to play 49 different video games from the Atari 2600, a games console popular in the 1980s. The software wasn’t told the rules of the games, and instead had to watch the screen to come up with its own strategies to get a high score. It was able to beat a top human player in 23 of the games.
the team trained an AI to play 49 different video games from the Atari 2600, a games console popular in the 1980s. The software wasn’t told the rules of the games, and instead had to watch the screen to come up with its own strategies to get a high score. It was able to beat a top human player in 23 of the games.
People don't go shopping for clothing. They shop for clothes that match up with an activity, such as hiking or power walking. Yet, the amount of unstructured data related to adventure sports is scattered all over the globe. Most stores don't keep a mountain biking expert on staff at all times. That's why The North Face uses Watson to guide customers to the right product.
You begin by entering a phrase like "Biking in London in March" to kick things off. You enter whether you are male or female. Watson might ask if you expect rain or snow, and if you prefer any custom options. You'll then see a selection of products, not a laundry list of sizes and colours, but rather clothing that matches up with your upcoming activity. It takes the 'unstructured' database of an e-commerce site and makes it more of a human-centric experience.
That's right, IBM Watson can help you "design" your granola. While it might seem cheeky, the brand uses Watson to analyse thousands and thousands of possible ingredient combinations. Once again, this is a good example of analysing unstructured data.
If you pick an ingredient like Blackberry Powder, Watson will determine that Dried Pomegranate Arils and Red Bean Crisps match up nicely. You can remove an ingredient (like the Blackberry Powder) and Watson will suggest an alternative (like Coriander). The alternative to this incredible intelligence would be to tap an expert chef who already knows which recipes work best for granola.
We can optimise for Narrow AI / the assistants right now
People don't go shopping for clothing. They shop for clothes that match up with an activity, such as hiking or power walking. Yet, the amount of unstructured data related to adventure sports is scattered all over the globe. Most stores don't keep a mountain biking expert on staff at all times. That's why The North Face uses Watson to guide customers to the right product.
You begin by entering a phrase like "Biking in London in March" to kick things off. You enter whether you are male or female. Watson might ask if you expect rain or snow, and if you prefer any custom options. You'll then see a selection of products, not a laundry list of sizes and colours, but rather clothing that matches up with your upcoming activity. It takes the 'unstructured' database of an e-commerce site and makes it more of a human-centric experience.
Took me 7 attempts to get OK Google to understand the question
MusixMatch built the API First, then the App, then the Website, and now only engaging in SEO
Took me 7 attempts to get OK Google to understand the question
Took me 7 attempts to get OK Google to understand the question
Invocation triggers define how users invoke and discover your actions. Once triggered, your action carries out a conversation with users, which is defined by dialogs.
Dialogs define how users converse with your actions and act as the user interface for your actions. They rely on fulfillment code to move the conversation forward.
Fulfillment is the code that processes user input and returns responses and you expose it as a REST endpoint. Fulfillment also typically has the logic that carries out the actual action like retrieving recipes or news to read aloud.
We can optimise for Narrow AI / the assistants right now