Virtual assistants such as SIRI, CORTANA, and OK GOOGLE are now penetrating and taking the corporate world by storm. Powered with natural language processing and personalization capabilities and promoted by the abundance of smart instant messengers, bots can leverage your rich topic-based content in ways that have never been seen before.
In this session you will learn how your taxonomy-based content strategy and semantic-enabled delivery platform can provide the content infrastructure you need to move forward with natural conversations that will give your users the answers they seek.
3. Gartner marks voice-search as a distruptive technology
By 2021, early adopter brands that
redesign their websites to support
voice-search will increase digital
commerce revenue by 30%
Source: Gartner
4. Voice search is picking up
20% of
searches are
vocal
Predicted 50%
by 2020
42% of users
started recently
5. And technology is improving…
90% voice
detection rates
x428 Alexa
skill growth in a
year
15,000 Alexa
3rd party skills
6. The holy grail of voice search
Google Answer
Box
→ AKA “featured snippet block”
→ Saves the user from scrolling
→ Rewards the brands which
provide the best answer
→ Being Read out in response to a
voice search
8. How does that affect us?
LONG
phrases instead
of short
keywords
SINGLE
answer leaving
no room for
second bestUser expects an
ACTION, not
a reference
“Best running
shoes”
“Okay Google, I need a
size 10 and a half running
shoe with a 5-star rating
that’s on sale in-store on
Newbury Street”
9. 3 TIPS TO OPTIMIZE YOUR CONTENT FOR
VOICE SEARCH
10. TIP #1: Think scenarios
Create user FAQ
scenarios
Translate into
questions
Formalize into
short answers
As a driver, I
want to change
my flat tire.
“How do I
change a tire?”
“I need to
replace a tire”
11. → Small topics
→ Include answer summary in the first paragraph
→ Optimize for both short keywords and long tail phrases
→ Create longer, expressive subtitles
→ Use procedural, bulleted checklists for tasks
→ Don’t bother with SEO keyword metatags
→ Use action triggers (“How”, “What”)
→ Use Schema.org markup where possible
TIP #2: Get your content in shape
12. TIP #3: Optimize your gated content for voice search
Public content
handled by
Google
Gated content
handled by
you
13. TIP #3: Optimize your gated content for voice search
Structured, Tagged Content
Sematic
query
Answer
Custome
r
“How do I
switch Ubuntu
log to debug?”
→ INTENT: how to
→ TITLE: switch log to
debug
→ OS: Ubuntu
→ ROLE: Administrator
→ PRODUCT: CF-5500
14. → Voice search is booming
→ Started with marketing, now coming to
techdocs
→ Second best is no longer good enough
→ There is plenty you can do to prepare
→ And if you do it right, you can disrupt the
consumption of your content!
Summary
Hello everyone, both here in the audience and the attendees in the Virtual track.
It’s a pleasure to be here. I’m Hannan Saltzman, and I lead product management for Zoomin Software, a leader in content-driven personalization and engagement.
In the last few years, a lot is happening in Voice search and conversational UI.
Until a few years ago, Siri and its competitors – OK Google, Cortana, Alexa and others – were not much more than a nice game. People used to make fun of them, insult them and watch them make a fool of themselves. And yes, they could also answer very simple question like “What’s the weather in New York”.
But things are different now. The technology is much improved, and the same Siri that made a fool of herself in 2012, have saved a baby’s life in 2016 by calling an ambulance at the right time.
Voice search is available to users today in three different forms: via personal assistants, such as Siri and Alexa, via online search – meaning when you see the microphone icon in Google, and via enterprise search engines.
In it’s strategic predictions this year, Gartner sees voice search as a distruptive technology. One that will re-define how user interact with online information. And If you look at it, there are some very good reasons for why you want to take voice search very seriously.
In last year’s Google I/O conference, Google revealed an aggressive growth in voice searches.
Voice searches grew 35 time between 2008 and 2016, and today 20% of web searches are voice-driven. Further more, google predicts that by 2020, 50% of searches will be voice-activated.
And not only that, but there is a clear acceleration in adoption. Surveys show that roughly 42% of users started using voice search only in the last 6 months.
In addition, the technology is improving.
Google, Apple, Facebook, Microsoft – are all putting their best brains into this.
With machine learning, NLP and big data – results are showing.
If you look at the voice word analysis accuracy – there is a huge improvement, and most platforms now support accuracy of over 90%.
In addition, these platforms keep adding skills: skill is basically another capability which the digital assistant can do – for example, skill of sending a text message, skill of turning the light in your house, or skill of ordering something on the web.
The number of 3rd party skills supported by Amazon Alexa, for example, went from 135 to 15,000 in just under 2 years – that’s a crazy growth.
So let’s talk for a minute about what’s the end game:
That’s where we all want to be: Google answer box, also known as “featured snippet block” is the first result which is shown above all others. It saves the user from scrolling, and more important – if you’re using a voice search, Google will read this answer aloud to you.
This is Google’s way of rewarding the brands which provide what it deems to be the best answer.
So how do you get to be in the Google answer box?
Once the customer asks something – for example – how do I change a tire.
The personal assistant or Google will look into your web site, and if they see content that they think is the best to answer that, and if your content contains the appropriate markup, that content will be surfaced and be read aloud to the user.
So how does voice search affect us?
I believe it affects us in three main ways:
For years, we all were educated to optimize our content for short keyword search. Now, we need to also optimize for long tail searches.
If in a regular search, you don’t have to be best, you just need to be good enough to make it into page 1 of the search results – with Voice search there is only one answer that will typically be surfaced. So it becomes an all or nothing game.
And the last difference, is that unlike a regular search, where user expects to get a reference, in voice search user expects Google to do the work for them and just extract that relevant piece of micro content from within your content.
So let’s get practical, and share three tips on how you can be better prepared for the voice revolution.
Tip #1 – Think scenarios.
To optimize your search for long tail queries, you need to understand how your users think and how they will be searching for the content.
For that, you need to think about the common frequently asked questions by user and the common scenarios where users will look for content. For example, a scenario can be “As a driver, I want to change my flat tire”
Then, you can explore how your users are asking that question. You can use online tools like Google keyword planner, Pagerank checker, Storybase and others, and check empirically what users are asking- for example “how do I change a tire” or “I need to replace a tire”.
Then, you can formalize it into topics that contain the answers and are optimized for the questions.
Tip #2 – get your content in shape.
To effectively support voice search, you need to have:
Small topic-based content
You need to make sure the answer summary is included in the first paragraph, or even better – in the first sentence of that pargraph
You need to remember to optimize for both long tail and short keywords
You can create longer subtitles, which will help you optimize for long tail queries
You as many bulleted checklists as you can; this will increase your chances of making it to the Google answer box
Don’t invest time in keyword meta tags; Google doesn’t do much with most of them anyway.
Start your answers with “How”, “what” and other action triggers. They get better SEO recognition and increase your chances of making it to the answer box.
And last, but not least; there is a standard called Schema.org, which is used for semantic markup of web elements and greatly increases your chances of making it into the search box. At zoomin, we are working with our customers to ensure their structured content can effectively be leveraged to Schema.org markup.
Tip #3 – is don’t forget that it’s not all about Google.
For most companies – there is much more than public content. There is also a lot of gated content, which is not visible to Google, which users will expect to find over time using voice search.
For that, you would need to create a strong taxonomy and classify your content.
So, if a user asks “how do I switch ubuntu log to debug”, your corporate search engine or digital assistant can then use NLP to break that down into intents, and use systems Like Zoomin that can provide a semantic API that takes that intent and returns the most appropriate answer to the user based on your content.
So, to summarize:
There is a huge growth in voice-activated search, and voice searches will be 50% of all searches by 2020
Voice search may first affect with your marketing organization, but very soon it will come to haunt you, so you’re better be prepared.
What voice search means is that there is only one answer surfaced to users – so second best is no longer an option – either you’re first or you’re not relevant.
And the good news is – there is plenty you can do to prepare!
So – go do some voice search!