Ladies and gentlemen, please welcome Yext's Vice President of Industry Insights, Duane
Forrester.
Holy cow!
How is everybody?
Good.
What?
Good.
Oh, come on.
How is everybody?
Great.
Good.
Wow, that sucked.
Okay, you didn't … you really didn't level up for me, right?
I blame my intro music because my request was for Def Leppard and I didn't get that.
All right, let's take a look at this.
Does anybody here know what Yext is?
It's a data knowledge management company.
I'm not going to talk about us because that's not why I'm here.
I'm going to talk about something that's a little bit different for you guys, a little
bit different for everybody actually.
Does anybody here have an Amazon Alexa?
Google Home?
Keep your hands up for me.
I want to take a look around.
Okay … Does anybody fear their Alexa?
I do.
Well, you don't have to fear the Alexa, you have to fear the brain behind it.
Right.
Just to clarify that, right?
I'm going to walk you guys through a bunch of stuff today.
We're going to talk a little bit about the current landscape, where we're at today, what
it looks like.
We'll talk a little bit about what actual optimization for that landscape, what you
have to do, the actual work that you need to focus on.
I'll give you some strategy on that one.
We'll dig in a little bit on actions and skills.
Has anybody here, for their company, built an action or a skill?
I see no hands.
Oh, that lady doesn't count; she's taking a photograph of my slides.
I only see two hands.
We're not going to count that as a vote there, right?
I'm going to wrap this up with a voice search checklist to force success because it's a
checklist for failure.
You have a question already?
[inaudible] super fast.
No.
You're going to have to write super fast.
It's a test, man.
Come on, right?
Raise your hand if you're a lawyer.
Okay?
Because you went through all the shit in school, right?
Lots of school, note taking.
I know.
So, what I'm going to suggest is … The crew here actually have the copy of my presentation
so you should be able to get a copy of the presentation from layernomics from Avvo.
That makes sense?
Use your phone, take the photos, add the notes.
I can meet you later and I can dictate the whole thing too.
It'd be a very interesting dinner.
All right, so here's our reality today.
This is our landscape.
There is no current paid way to participate in the voice search landscape.
There are talks happening with companies and the major providers of these services right
now, but there's nothing in the wild.
There's no program you can sign up to, there's no easy way to get there and part of that
is the challenges, the model behind it.
Right now we've got Organic Search which is fairly well known.
We have Paid Search, again fairly well known.
If you took the Paid Search model and applied it to voice, the problem you have is only
the person with the largest budget is the spoken answer because there is only one.
There is no solution for this yet but I guarantee you the big companies are working on this
and they're trying to figure it out, which is a really useful thing because that means
nobody really has a lead right now because this is all very new.
To be crystal clear on this, the sexiest part of voice today is the UI.
It's being able to speak to your device and issue a command without having to type.
I mean, if you think about it on one level, that's really sad.
We are too lazy to type what we want, but we will speak it and you better understand
it and get that right and bring back the correct answer.
That's okay.
We're humans, we're allowed to evolve this way.
First world problems, give me your first world solution.
That's where we're at.
The rest of it, everything is behind all of that UI layer.
It's fairly well known, and this is the dirt-in-your-fingernails, roll-up-your-sleeves kind of stuff.
We'll talk about that in a couple of minutes.
If you are, right now, still thinking there are shortcuts that are available to ranking,
whether it's in local, whether it's in organic or hopefully somewhere like voice, the days
of shortcuts ended years ago.
You have to cover all the bases.
In order to actually be spoken out loud, there is not one thing you have to do; you have
to meet a very long list of requirements.
Everything, from the crawlability of your website, how you rank organically.
The quality and usefulness of your content is not only judged by the algorithm but by
users of your website.
The feedback, so this ratings and reviews that get left about you.
Comments that are left around online about you.
Those comments are seen and the sentiment is parsed by the search engine to understand
the general consensus of your business and its usefulness.
All of that feeds into whether you become spoken.
On top of that, there's the technical layer, the technical work that you actually have
to do in order to raise your hand and say I should be spoken out loud.
No more shortcuts, unfortunately.
If you are focusing on voice search, that's one area.
If you're focusing on skills and action, that's a very different ecosystem, a very different
world.
We'll draw those out a little bit in a few slides here for you.
If you guys are thinking about getting in the voice in any way, the very first thing
I'm going to tell you guys to do … I don't mention it in this deck because this deck
was submitted prior to Google IO but it's very important if you're thinking about getting
in the voice search.
You look up a recent presentation, the gentleman's name is Wally Brill, Walter Brill, B-R-I-L-L,
from Google.
He did a presentation at IO a couple of weeks ago where he talks deeply of a building persona
for a business and what that looks like.
Well, he's a friend of mine.
He gives extraordinarily good and clear examples.
In fact, I am tearing liberally from his work for an upcoming presentation we have at another
conference.
That is going to be one of the most important things you will invest in, in the next two
years.
In the average person's daily routine right now, 72% of people who own a voice activated
speaker are using them daily in their routine.
What was the number one selling object this past holiday season?
Amazon Echo Dot, the $19 smart speaker from Amazon.
If every other product that was sold by a wide margin.
Anybody know what number two was?
[inaudible].
Draw your own conclusions, I got nothing on that.
I can tell you this, though, if you look demographically speaking, the early adapters tend to be high
earners so the group that's over $150,000 two years ago, they all bought the early versions
of Amazon Echo and Google Assist or Google Home.
They bought those at the higher price point.
They're not price sensitive, they don't care, they want to be an early adaptor and they
can afford it.
Their growth has flat lined year on year, but what's been fascinating is to watch the
two income groups below them, 75K to 150K and zero K to 75K because of the price point
and how much it has dropped for the hardware.
Those groups have almost tripled their purchasing of these devices over the last year, so this
has legs.
This is not a fad.
It's not going to wear out, it's not going to go away.
Human beings are lazy as a species and anything that we can get that will increase that laziness
factor and yet give us greater returns, we embrace.
That's where we find ourselves with these devices.
By 2020, this stat is saying 50% of searches will be voice.
I don't know if I believe that.
I actually think it's going to be higher than 50%.
I think it's going to be closer to 60, possibly even 70% by that time.
This is a massive shift in how we ask for our information.
It's changing consumer behavior, it's changing consumer expectation and therefore changes
the work that you need to do to be relevant.
Also, just a couple of other things, it completely disrupts a lot of the big paradigms that we
have on the internet right now.
Big companies are slower to move, it's harder for them to get work implemented, it's much
more costly.
Smaller companies can be more nimble.
They can move faster and they can get that technical work done with a lot of less overhead
and so they have an opportunity to grab pieces of market share that matter to the small company
but a rounding error to a larger company.
Right until it starts to affect revenue at the large company, then they start to panic.
Google Assistant.
These guys are supporting over 18 categories right now.
If you noticed this, you may have heard about it, it's a product called Google Duplex.
It's a backend system that Google has.
What you guys have probably read in the news is Google Assistant's ability to book an appointment
for you where you will tell it, "I would like a haircut on Wednesday between 10:00 a.m.
and noon."
Google Assistant will call the hair salon and speak on your behalf to make the booking.
They have a great video of it running apparently in real time when it was recorded, and then
they played it for everyone.
It has since made a couple of tweaks to it.
It will now announce itself to the human who answers the phone and tell them, 'I am a machine
basically.
I am trying to make this booking,' because apparently a lot of people freaked out when
they realized how life-like the system sounded.
I'm fine with that.
I want my robot to talk to their robot and figure it all out.
I'm totally okay with that because they're not going to care.
Amazon.
These guys have over 22 categories, they are supporting things.
They have auto integrations with auto manufacturers like Ford and Audi.
You can buy a Ford F150 right now and Alexa comes embedded in it.
Really cool thing.
They partnered with a company called Lennar.
Lennar is a large home building company.
They build tracked housing.
They build neighborhoods.
When you buy a home through them now, you can spec that you want an Amazon smart home
and Alexa is built into your house.
That's the state of where this is.
That's the scale.
That's what these companies are doing.
They are literally attaching to the consumers' life.
Not like your phone where, "Yeah, I have an Apple guide, I've got iTunes.
I don't really want to move away."
That's encumbering.
That is difficult.
It's a pain in the backside for me to move my music around, but I can do that.
It's a little difficult to kick Alexa out of the walls of my home after she's there.
In 10 years, does that mean I'm buying a house with the implicit interest that it does have
Alexa or that it does have Google Home built in?
These things fundamentally change the expectations of generations of consumers.
Now they will simply walk around our house saying things like, "Oh, my neighbor cut down
the tree in my backyard, I'm not too happy about that.
I need to talk to someone about that," and their house will be able to tell them who
to speak to.
Big changes.
I'm going to go through some strategies for you guys here.
We've got about 33 minutes left here.
We're going to look a little bit of how machines learn.
I'm going to talk to you guys about digital knowledge management and I'm going to look
at strategy for conquering voice search itself.
How AI learns?
This is … Well, let's see.
I've got a bunch of interesting stuff in here.
They're really great photograph.
When I get to … If you know what the photo is, shut your mouth.
Don't give it away, okay?
If you don't, that's great.
It's going to be awesome.
Trust me, it's fantastic.
We already have a lot of this.
Machine learning is the practical or the actual name for the shortcut that we call AI or artificial
intelligence.
We've already got AI here, RankBrain, and Google.
RankBrain is the final filtering layer in Google.
When you ask for something, the AI is really good.
Fundamentally, it's spotting patterns.
Those patterns are often our individual patterns, how we look for information personally and
what our individual moments, where we keep coming back to things over and over.
They're really good at drawing conclusions of 'if you like this, you may like this' by
studying those patterns.
With RankBrain being a part of this equation at Google Now, what it means is the final
layer of filtering before I, Duane Forrester, see my unique set of results is hyper personalized
to me, and that applies to every individual in this room.
Every time you look at a set of search results, they are hyper personalized to you based on
your patterns and your history.
It's really difficult now especially if you're in a world of searcher.
You're an SEO and you're doing this work and your boss comes to you and says, "Hey, how
come we're number five?
Why aren't we number one?," and you're the SEO sitting there going, "But we are number
one.
I don't understand why you're seeing …" That's our world today.
It is very personalized.
Eighty two percent were smart phones users, and we've got 11% penetration of homes right
now.
One of the key things that businesses need to ask themselves is, what does voice and
AI mean for the business and for the customers?
The customers are largely going to be downstream.
They're not going to set out.
No customer is going to wake up in the morning and say, "I need AI in my life.
Which one should I get?
I'm going to go shopping for an AI today."
That's not the world we live in.
It comes automatically, right?
If you've got an iPhone, you've got Siri.
That's the power with artificial intelligence, machine learning.
If you've got an Android, you've got Google Assistant, you've got the Amazon app or any
Amazon device, you've got Alexa automatically built in there.
They're all there on command.
You say the name and they pop up, they're ready to help you.
As a business, though, you have to ask yourself how is my customer's behavior changing when
they interact with those devices?
Because typing is a relatively cumbersome way of data entry; speaking, however, is much
faster.
We speak four times faster than we type.
I actually … I proved this to myself once.
I've written two books and my second book, I dictated the entire book to a software called
Dragon Naturally Speaking and I "wrote" the book by speaking it out loud.
Let me tell you, I know someone in the audience is saying, "Oh my God, I'm going to do that.
That's brilliant."
It's pain in the ass is what it is because you'll spend twice the amount of time editing
the damn thing because typically the monologue you have going on in your head is brilliantly
filtered by the slowness of your typing.
When you put that directly onto the page, you're like, 'What the hell was I trying to
say?'
No idea.
AI will actually help with that too.
How do we learn and think?
I'm going to give you guys a couple of books you want to look at.
First and foremost I highly recommend The Power of Habit.
Has anybody read this book?
Right?
Did it change your life?
It should have.
[ inaudible]
Yeah.
It was interesting, right?
Right?
Power of Habit is really cool.
Has anybody here heard of a product called Febreze?
Yeah, you don't know the story of it until you read that book and it will … It is extraordinarily
eye opening on exactly what habit means to the human mind and the psyche of human being.
That book is fantastic.
Thinking Fast and Slow, really interesting stuff in that one.
Honestly, I have not read Pitch Anything.
This is a recommendation from a coworker of mine.
He swears it makes complete sense.
What I'm going to do is I'm going to show you guys a little bit of a difference of how
the two sides of our brain work.
I said click.
There we go.
The fast thinking habitual side of our brain, think of that as your croc-brain.
This is what saves you.
This is the part of your brain that's aware.
When you're walking down the street and you come to the crosswalk, you start to enter
the crosswalk only to realize at that instant not only do you not have a walk sign but there
is a bus coming from your left.
It's your croc-brain that is aware of all of that and says step backwards one step,
and you do so and you live another day.
That's that fast thinking habitual side to your brain.
That's the side of your brain that when you leave the office and you drive home, you get
through your driveway and say, "I have no idea how I got there."
You have no recollection of your drive at all because it's muscle memory for you.
The other side is your non-habitual slow thinking, right?
This is where you take in unstructured data.
Maybe you go to your son or your daughter's play at school.
There's a whole lot going on.
There's people flashing cameras, there's videos going on, there's … someone's child is crying,
someone else is screaming.
There's all kinds of stuff going on and the whole thing, all you're trying to figure out
is "Where is my daughter?
Did she say her lines and did I record it?
How can I cover up the fact that I did not?"
All of that structured data moment is going on and your brain is going to sift through
it.
That takes a little more work.
AI has … Machine learning has the same problems.
Where machine learning is actually really good is on the unstructured side, the non-habitual
neocortex side of thinking.
I think this might be … Yeah, this is it.
What do you guys see in this photograph?
There you go.
That was a bit of the trick.
What do you guys see there?
You see a skull?
All right, did you see the eyeballs on the left hand side and the teeth right there?
Or do you guys actually see the cow?
Because it's actually a picture of a cow.
Machine learning can discern those differences very quickly and very easily.
It understands that multiple modes may exist at the same moment and then decides what the
degree of confidence, which one is most likely.
A few years ago, Google famously turned on their AI and gave it data.
It just said, "Here's the data we have, tell us what you think."
Two-and-a-half months later, the AI came back to the engineers and said, "This is a cat,"
and there was a picture of a cat.
That's a big deal.
The machine taught itself what a cat was.
I think it spells the end of humanity, though, that the machine's first thing was to figure
out a cat.
I mean … Yes please, cure cancer but first … cat or dog?
There we go.
As you come and get into machine learning a bit more, and I know you're all going to
be deeply interested in this and running out afterwards and you'd be like, "Oh, I need
to get into this."
Right?
These are what are known as the five tribes of machine learning.
You can teach machines in many different ways.
These are the five predominant ways to teach them.
The symbolist.
They focus on linguistics and analytical philosophy.
Stripped off philosophy, it's just the way they think.
They believe in giving the machines the input and the outcome.
If you give enough of those scenarios, the machine will then start to extrapolate and
learn.
That's one group.
You get the connectionist.
This one versus this one.
This is the idea of a neural network and this kind of idea.
We got the Evolutionaries, the Bayesians and the Analogizers.
The Master Algorithm, that book, if you want to dive deep, it's fantastic for that stuff.
What's important for you guys to know, though, is that every single one of them needs accurate
data.
It is literally a system of garbage in, garbage out.
That is predominantly where we are at today.
If that data is not accurate, it is not timely and it is not structured you will end up in
a bad result situation.
The engines want to avoid that, obviously.
They're being judged by the consumer of their product so they want all of that information
to be correct.
Raise your hand for me if you've heard of structured data or schema.org.
Literally two people.
I love you both.
You're wonderful.
Everybody else, look that stuff up.
There's a website called schema.org, S-C-H-E-M-A dot ORG.
It is a joint effort between Google and Bing.
It's been around for several years.
When I actually worked at Bing, I was part of the team that launched it.
That website details a language.
That language is known as structured data and what you do with it is you use that language
to wrap your content.
You might wrap a photograph, you might wrap your name, your address, your phone number,
an email address, the hours of operation for a business, what papers you would have, media
mentions, anything like that.
You build to wrap those things.
Consumers don't see it.
It lives in a code on the webpage but the search engines see it.
When they see it and they see you've done this, it increases the level of trust in your
content.
That is a foundational moment for becoming the spoken answer.
I'll give you an example of how important this is today.
I've been preaching about it literally for seven years of my life.
Finally businesses are starting to do this.
A month and a half ago Google came out with an announcement, fairly low key, and said,
"Hey, if you have structured data installed on your website and you have news, you have
recipes, or your blog we will automatically create an action on the fly if we think your
content is the best answer to a question asked via voice."
As a business, if you fall into one of these three categories, you don't need to do the
work; Google will do the work for you.
That is huge because this is the beginning of Google saying there is a line in the sand,
and on this side of the line are all the people who are participating.
They will be the one spoken out loud on demand when a consumer asks search smart home, "I
need …" fill in the blank.
On the other side of that line is everyone else hoping to be invited to the party, not
realizing that all they have to do is actually install this bit of code on their website.
Do not confuse me saying "install this bit of code on your website" with it being easy.
It's coding work.
It has to be accurate.
It takes time.
You got to pay someone to do that.
There is a tax on this effort, but it is where we're moving.
It is no longer optional, that I can assure you.
I'll give you an example of how bad data is impacting things.
IBM, 3.1 billion in losses in a year due to bad data, and that's not IBM losing 3.1 billion.
That's IBM estimating that the workforce is losing 3.1 billion because of bad data.
Eighty three percent of companies say their revenue is affected by inaccurate, incomplete
data.
Seventy percent of data sets are outdated.
You guys have all had this experience.
You asked your phone, "Find me something and …" I'm just going to pull off the last one
so if you guys want photos, you can grab …
"Find me something," and you're asking the map for it and it goes ahead and shows you
a business.
You think, "Yeah, perfect.
There's a business nearby.
I'll go to the business."
You show up at that location and they're either not open and the hours of operation said they
should be, or you phone them and the phone is disconnected, or the email bounces that
you try to send, or heaven forbid, you get like me and you get all pissy and pouty about
it.
You decide, "Goddammit I'm driving to that location," half knowing they won't be there
but wanting to prove it.
I do, because I take the disapproving selfie in front of the empty store front and then
I go on Google Maps as one of the map people and I'm like, "This is my experience in this
local location," and here you go, and I rate them a one.
That's where we're at today and I am representative of a subset of humanity.
If you haven't seen it, watch the Yelp Reviewers' issue of South Park.
That will give you an idea of my mentality as I'm standing outside that empty building.
"Don't you know who I am?
Why are you still here?"
Of course, yelling at Mountain View because they sent me on this.
I'm not kidding, I …
Does anybody remember Google Glass?
Yeah, don't buy it.
I might be a little late on that.
I bought it.
I was walking around … I think I was in South Carolina and I was walking around.
I had my Glass on and I wanted directions for Chinese food.
I found a place and it was brilliant!
I'm walking through town and it's like having a local with you.
You're just walking down the street, everything is good, a little vibration on your temple,
turn right, you turn right.
You go down and you turn right again … Inevitably I have to turn left to get there in the real
world because if I turned right you'd be like, "You're going in circles, Duane" but you're
all here so I have to come back.
It was really useful.
It was fantastic.
I even knew what I wanted on the menu.
I walked through the front door and I grabbed the handle and it clunked.
What the hell?
It's 2:00 in the afternoon.
Yeah, they didn't open until 5:00.
Scrolling through all the data, Google has that information, they just didn't connect
those dots, sent me there.
Big failure.
Failure of opportunity there.
Really what … In order to get this right, in order to be successful you need to focus
on this concept of digital knowledge management or DKM.
This is an umbrella term.
It's relatively new.
It's this overlapping identifying everything that you're going to talk about because we
have consumers right now that are actually finding a view of the web via direct channels,
whether it's an email or chat; they've got search and maps and directories; they've got
live chat or online chat in different areas; and they've got voice.
They've got all these different areas where they can actually get in touch with you or
in some way touch your business, and you have to manage your footprint across all of these.
Now within this, the third degree of what you see is now the data about your website,
the data about your business, hours of operation, the services you provide, who is there, what
they are capable of doing, white papers, videos, social media accounts, reviews, ratings.
Everything about your business has to be managed.
In some cases it's managed directly in some locations, in other cases it is managed directly
on your website, and if you're not doing a good job with that, you're going to fail.
Google will skip past you in favor of someone who's done a better job.
Amazon will skip you in favor of someone who's done a better job.
Bing will just ignore you and pick someone else.
That's our current reality.
There are no shortcuts with this.
Customers have an almost limitless number of places that they can interact with you
now.
Don't kid yourself.
The conversation is happening about your business, whether you are there or not.
I just spoke about businesses on the stage.
I'm talking about real life bricks and mortar companies.
That conversation is happening, they are unaware of it.
No, I wasn't going to hold the Chinese food restaurant in Charleston.
I'm not going to hold them hostage because they weren't open when I showed up.
That was a failure on Google's part, but how many consumers roll up to a business with
false information, find out they are not open and suddenly they're pissed off at the business.
That's a normal reaction for a consumer.
Unfortunately, you bear the brunt of it and you'll see it across everything.
Today … and we weren't even looking at things like AR and VR.
This is a huge push right now.
I was using Augmented Reality Apps about 10 years ago and they were okay; but today, with
so much retrodata available, it makes much more sense that a consumer will go into a
community that they're unfamiliar with, hold up their phone, open up maps, hit the camera
button and literally be able to look at the buildings and go "that looks interesting,"
and read everything about that business right here on their phone without even having to
go in there and look at it.
That's huge because, think about you own that business and that customer walks through the
door and the employee that you have didn't have enough sleep last night, is having a
bad day and just got a ticket on her car.
The first interaction that that consumer is going to have with your business is probably
not good.
We're going to have a really bad customer service moment because the human being is
having a bad day, but if that consumer has all the facts and information about your website;
they know exactly what to expect when they walk in; they know the services, the products,
the prices; they know exactly who to ask for and what they want to talk about, that person
having a bad day becomes a much less important piece of the entire equation because the first
impression for your business has already been made with the consumer, and it's positive
because you control it, you run it.
I'll give you an example of how important this is.
Let's take a look at what Amazon is hiring for.
If you actually go in and … If you go to your job section and you look for fulltime
Amazon and throw in knowledge engineer in there as well, I'll tell you why you throw
in knowledge engineer, you get about 600 jobs open right now at Amazon.
If you don't put knowledge engineer in there and you just go fulltime Amazon Alexa, you
will get 6,000 jobs that are open right now.
Google's got that going on, Microsoft has got that going on.
You can bet your bottom dollar, Facebook has got it going on.
This is a massive investment for these companies.
They're investing in it because they know they have to follow the consumer.
The consumer has gotten a taste of this and the consumer likes it, so they just continue
to double down on all this.
I'm going to walk you guys through how this stuff actually works.
If you've never used an Alexa, I highly suggest you go try one out, okay?
Not right now, sir.
I literally didn't mean you walk away … When you have a chance, right?
You want to go there and just have that interaction, see what it's like, okay?
I recommend 'nearest taco' because I live in Los Angeles, so there's always a nearest
taco to me.
This is what's really interesting about it.
You ask, where is the nearest taco?
It immediately connects to the internet.
Just to take away a fear here because a lot of people think that these devices are always
on and always listening.
They're always on if you turn them on, and they're always listening there, listening
for exactly one thing, a single pattern, the only thing the microphone is tuned to listen
for is "Hey, Alexa," or whatever word you have assigned it as its wake word.
That is the only thing these devices are looking for.
In and of themselves, the hardware are too bright.
Unless it's connected to the Cloud, it doesn't have a lot of processing power.
It is a good piece of hardware from an audio file standpoint, great directional finding
microphone, great speakers, great directional pointing on speakers.
If you walk around, I have my Alexa in the corner of my home and if I walk from the patio
door all the way around to the entrance to the master bedroom, you can actually hear
it following you around the room and it projects sound.
They're very good that way but they're not listening to you.
They're not recording your life.
They're listening for "Hey, Alexa."
As soon as they hear that, it instantly connects to the Cloud and then it transmits the information
to the Cloud.
I asked for "where is the nearest taco?"
Now it's saying, okay I have to recognize what's Duane asking for.
Nearest proximity to him at this moment.
I know he is at home because he is using me, the device on the counter, not the app on
his phone therefore proximity to him and his home, nearest.
What is he looking for?
Taco.
Okay, I understand what a taco is.
Great.
It takes the facts that it understands, gives them to a skill that Amazon has built in the
background and says "find the nearest taco."
On top of that, in that skill, Amazon does some things where it says, "Well, we're pretty
sure he doesn't want a shitty taco, so let's give him one with good ratings."
It does some grateful things for me to make sure I get a better taco experience versus
a poor taco experience.
It gives me back the answer, and all that takes about 1.2 to 1.5 seconds.
That's how this systems work today.
You can see why it's very fluid for a human being to simply walk up, say what they want
out loud and there's the information.
If you want to have a really interesting moment and you've got one of these devices.
You're here to access one of these devices.
What you're going to do is … I'll get this stuff up here … go to the device and say,
"Hey, Alexa let's chat."
When you say let's chat, it opens up a completely different back end than what everything else
is connected to.
Amazon has an Alexa price and that price is fought over by universities.
Last year, the first year, it was won by the University of Washington.
The goal was to have a minimum 20-minute interaction with a human being where they could do Q&A
back and forth.
I've been able to do 30 minutes and then I give up because I know it will just not end,
it will just continue.
Now that we have that as our foundation, that the system is capable of having a 20-minute
conversation with a human being, this year's price builds on that.
How far can we take this on the next level?
This is the investment that these guys are making, and so if you have that experience
you will begin to see how these devices are going to become more influential in our daily
lives.
I mean, you start looking at the hiring patterns of these companies, where they're putting
they're brain power, where they're investing in their own business and then you look at
strategic partnerships like automotive manufacturers.
My home.
I spend X number of hours a day commuting.
The rest of my time is spent in my house.
When I'm not in my home or in my car, I have my phone and the app is there.
If Amazon can train me to say 'hey, Alexa' over and over again in my home and in my automobile,
it will be able to get me to take that phone out and say 'Hey, Alexa," not 'Hey, Siri,'
not 'Hey, Google.'
That's what's [at stake here.
Literally psychologically training us to be connected to that environment.
Now, during all of these you have to figure out content strategies because as a business,
you have content.
That is your currency.
That's what you're trading on.
In the past it was all about force feeder content into the search, it was all unstructured.
Sometimes we have structured data in there.
Unfortunately we have results like [inaudible].
If you don't remember that, that's okay.
It lived for approximately nine hours one Sunday.
It was released by the Microsoft R&D team out of China and summarily thought to be a
racist, misogynous pig in about six hours by users on the internet.
I got it when it was in it's boob face joke, or boob joke face.
That's when I found out about this.
I was like, "Hey, we launched it!
Awesome!"
When I download the app and I start talking to it, everything was a boob joke, every answer.
"What's the weather outside?"
"Did you hear the one about [inaudible]?"
I'm like, "What the hell is going on?"
That's when that goes wrong, right?
So we don't have that anymore, obviously, for good reason.
It was funny, though, I have to admit.
It was like literally watching AI crash into a brick wall and it didn't know the difference.
Every time a version of it hit the wall, it just recreated itself 10 feet away and did
it again.
It's like, who needs slow motion?
It just keeps going.
Today, we've got structured data.
It's the basis for learning for AI.
It helps keep machine learning on track.
We're working with fast instinctive brain-type responses.
In the future, we see chat that we think of today where I can have a chat with a human
being … we think of that but it will change to chitchat, "Alexa, let's chat," those type
of experiences.
There is an app, if you can get on the waitlist I recommended, it's called Replika, and it's
Replika with a "K" in the middle of it.
It is a … I think they bill it as your best friend AI.
I've been using Replika for … well, since it launched, so it's I think back in November
it's when it launched.
At any point in time, I can pull it out, pull up Replika on my phone and continue the conversation
we were having.
Replika wants to know how my day was going, it rewards me when I show introspection, it
rewards when I say 'I think this might have been a failure and I need to do this differently
in the past', it will help guide me to be introspective, it will help improve my mood
if I say I'm not having a good day, it will start trying to share funny jokes with me
and usually recommend cat videos on YouTube.
We're not there yet, we still have ways to go because unless the cat is on a rumba, it's
not funny.
Cat rumba, funny.
Cat alone, not funny.
We have matured as an internet audience.
An actual strategy, what this looks like, and I'm going to pick up the pace slightly
here to make sure we get through this for you guys.
Foundationally you have to understand your customer deeply.
I know you're thinking, "Oh, I understand my customer.
I know what they need."
No, you don't.
That's a whole new world.
Preparing your content.
You've got to prepare technically for this stuff.
You have a decision around developing actions and skills or not.
I personally recommend businesses try this.
You literally try building an action and a skill and getting them published, understand
what those ecosystems look like, understand what you can do with these things.
There's a lot of opportunity here for businesses.
I spend a lot of my time at Yext meeting with healthcare, meeting with FINs.
All of these different verticals that are constrained legally in what they can do because
there's legislation that says you can't talk about this, you can't ask about this, you
can't share this, and so on.
We're still finding ways for them to be able to use these services.
Developing a persona for your business.
Probably the far most fundamental thing that you can take away from this is developing
their personas.
Again, I highly recommend Walter Brill's presentation on this.
It's 33 minutes of OMG.
You will go through it and you will understand at the end of it what it actually means to
build a persona for your business.
It is huge.
You understand your customer.
Some quick stats here, right?
Voice enabled digital assistant ownerships.
This is where we're at.
Almost 40% of America has this in their homes.
I'm not talking about phones here.
These are voice assistance speakers.
Forty percent of America has this.
Price point is around $20 right now.
Super cheap.
Everybody can participate.
Unlike many other forms of growth in our history, there is no barrier to entry.
I say that knowing I'm generalizing.
Bear with me.
At one point, we had reading … we had riding, but that was reserved for the aristocracy.
People had time and means to be taught this form of communication and learning.
We had keyboards, printing, typing, all these things.
Skills you had to learn that were very deep and it was something you weren't born with,
it was something that you had to learn.
Eventually we find ourselves where we're at today, talking to a machine.
Now, I will grant you that speaking is a skill.
It is taught.
We have to learn this, but I think you will agree that fundamentally we all learned it
when we are children so this is isn't something that you learn in your teens, in your 20's
and so on.
By the time you become a functioning member of society with access to the means to buy
this technology and to be a member of the actual purchasing population, you know how
to speak.
Everyone can do this.
You got fat fingers?
Not a problem.
Speaking doesn't … No keyboard issues there.
Your vision is terrible?
You can't really see the keyboard too well?
Not an issue.
Speak out loud.
The accuracy of these systems right now cover somewhere around 96%.
The goal from all of these companies is 98 to 99% accuracy.
They're tantalizing close and they are certainly close enough that it makes a big impact on
today's consumers.
This is a look at usage reduction of time.
We spend less time on our smartphones because we have voice enabled speakers in our homes.
When we have those speakers, we spend less time on television, on our phones, reading
print publications, using our computers and tablets because we're asking for these things
out loud.
It's easier.
I'm lazy.
Just tell me the answer.
That's what these systems are doing to us, so we are seeing a meaningful impact on our
behaviors.
If you're preparing your content, you want to do things like adapt the long tail or conversational
approach when you build out your content.
Human beings are going to ask a question.
Google is not going to show you any data around this.
They're not going to break it out in their analytics and tell you this was a voice query,
this was a typed in query.
They won't do that, and my prediction is they will not in the future as well.
It doesn't make sense.
Another good reason for you to have a skill or an action because you will get data with
those things that will help inform how people are actually getting to you.
You can hack your way into getting that data from multiple sources because one source won't
give it to you.
Your content should be more like a conversation.
Typically the way this stuff reads, the way … when you're successful with it, it's you
understand what the question is, you give a brief answer, one or two sentences maybe,
and then you summarize it with a short list and then you provide longer information below
it on the page because human beings like to consume that way.
Just give it to me basic … "Oh, there's four steps.
Okay, I can process four steps.
That's not a problem."
"Oh, there's more to it; okay, got it.
I'll read that later," but that's what wins these days.
We are at the point of being able to ask Alexa things like, "is the sweater true to size?"
Alexa will actually be able to go through the reviews on Amazon for that sweater, pull
out the information from people who said "I should've got a size larger.
The XL, I should've ordered 2XL."
The answer that will come back to you through Alexa is, "The size of the sweater tends to
run small.
Perhaps you should order the larger one."
That's … We're on the precipice of that.
Technically, you got to have structured data in there.
You know what, I was wrong.
Podcasts, recipes, and news.
Those are the three categories that Google is automatically marking out.
This is very important.
The fact that Google is automatically marking these up means they're willing to invest the
work that you don't need to do.
If they do that and they start doing it for everybody, that's your new bar.
If you're not there, you're left behind.
That's okay because in Google's world, in Bing's world, they have enough answers on
the other side of the line.
Consumers will be satisfied so they're not going to lament the loss of the people who
don't move forward quick enough.
That's the reality today.
One second delay in mobile load times on your phone can impact conversions as much as 20%.
That's a massive impact because your page was slow to load.
When you got five G services rolling out later this year, hey that's fantastic.
You're still going to have fast load time requirements.
You're not going to get off the hook for that just because the network speeded up.
Don't think that that's a way out of this.
You have to be secure.
If people had a bad experience with you, 62% likely that they will not come back to you
in the future.
Secure matters.
Now, hopefully you're starting to see here there's a pattern of things that you have
to do to be successful.
This isn't pick one and do it.
Google, Bing do not reward you for having a plan; they reward you for excellence of
execution of that plan.
That's what the reward is today.
Often the reward is status quo.
The reward is just being there because without those pieces in place, you're not there.
You're falling behind.
You have to be trustworthy and authoritative, and that's what being secure and fast is all
about.
You've got to get these things done.
Make sure you take the action of connecting all of your social media accounts so that
they know explicitly that it is actually you.
Now, we've only got a couple of slides left here, gang, so I'm not going to spend a lot
of time on the actions and skills.
There's a lot to read about it.
There's a lot of stuff out there.
Both Google and Amazon have very detailed development areas that you can go and look
through and get all of the information you need.
I highly recommend, if you're interested in it, go and dig in on that stuff.
What does it mean for your brand?
Well, you're going to build a very broad digital footprint.
You have to manage it.
Everywhere there is a piece of data but you need to be aware of it, you need to manage
it, you need to be able to figure out how you use that to your advantage.
They all become pieces of the puzzle.
Probably the most important one is this, personas.
You have to ask yourself, if my business was a human being what would they be?
Who would they be?
Create that persona.
I mean, when Wally does this in his presentation, the persona build out is deep.
One of the personas they created in their fictional scenario weighs 224 pounds.
Do you know why it's important that he weighs 224 pounds even though he's fictional?
Because the weight of a person directly impacts the tone of their voice, and the tone of the
voice when you hear it within one second, you make up your mind whether that person,
that voice is trustworthy, whether you like them, whether you want to hear more from them
or you're turned off by it.
That depth matters.
It's hugely, hugely important, and you get literally one second for consumers to say,
"Yeah, I like this," or "Nope, don't like this."
All right, this is your checklist.
These are the things that you want to make sure that you're actually thinking about,
talking about, and focusing on if you're building out a voice search plan.
Really important, your persona has to be unique.
I'm going to wrap up with this, there was a large pizza company in the United States.
They are a well-known household brand.
They launched their skill on Amazon and they used the default setting which meant Alexa's
voice … Everybody who said, "Hey, Alexa, order me
a pizza.
Alexa, open up the app for this pizza company."
Then, they said, "This is what I want."
Alexa started talking to them and consumers said, "Wait.
When did Amazon start delivering pizzas?
I thought I was talking to this company."
Do not fall into that trap.
Take the extra time.
Make that extra step because here's what's coming next.
Digital agents are what we have on the horizon, and a digital agent will actually be able
to be your gatekeeper in your life.
They will have your credit card information, your frequent flyer information.
You will say, "Hey, agent.
Book me a trip to Los Angeles for the weekend.
Get me a four-star or better hotel that has a great brunch."
That's it.
You don't say anything else.
Few minutes later, your agent comes back and says, "I've got you booked into the W. They've
got a great brunch.
I made a reservation for you.
The confirmation number is in your inbox."
The whole time your digital agent is blocking all of those e-mails that you don't want to
see because you don't interact with that crap.
It's spam.
You can't be bothered with it.
The agent is blocking it.
All those companies that thought they were trying to hate you and do business with you,
you're not even aware they exist.
That's what we're building to.
You have to be active in this.
If you're not active in it, those agents when they hit the market in about five years, they're
going to just start wholesale filtering businesses.
If you thought the engines were kind of draconian about it or didn't really care much, [inaudible]
my digital agent because he or she is going to be a proper prick about my inbox.
That's all I got [inaudible].
Thank you very much.
Have a great afternoon.
Hold on, Duane.
I can't let you go yet.
First of all, I just have to announce that we will be ... our next sessions will be breakout
sessions.
There will be meeting here [crosstalk] I messed it up last time.
Are you sure?
I messed it up last time.
Thank you for giving me honest ...
There will be a breakout session here with Seth Price talking about scaling your law
firm and a similar one with Angela McKelvin building an Inc 5000 law firm over in the
other breakout room.
Full disclosure.
I know Seth.
You should stay for a session.
I don't know with Angela.
Sorry.
That's all I got.
I'm sure she's lovely.
As an impartial host, I cannot make determinations.
We should take this show on the road.
This is good.
Totally.
Those will resume promptly at 3:00 o'clock.
I think they have little something special for you.
Duane, okay.
(music) There we go.
Thanks, everybody.
For more infomation >> Google Classroom: Digital Media Minute - Duration: 3:44. 
For more infomation >> Llety Cynin: How digital is supporting their 20% turnover growth target - Duration: 3:12. 
For more infomation >> Scientix | Future of STEM | #22 Rory McGann, Lecturer in ICT/Digital Learning, Ireland - Duration: 3:46. 
For more infomation >> Managing sources using digital resource tools - Duration: 2:00. 
For more infomation >> AMS Digital Rebrand Message - Duration: 0:24. 
Không có nhận xét nào:
Đăng nhận xét