ACCESS SF: Yahoo’s Award-Winning Captioning Initiatives

ACCESS SF: Yahoo’s Award-Winning Captioning Initiatives


[MUSIC PLAYING] LILY BOND: We are 3Play Media. We are hosting the event. And we are a video
captioning, transcription, and audio description company. And our goal is, really, to
make accessible video easy. So we have a lot of
user-friendly parts of our company, like
our account system. And we have interactive tools. And all of these things
really make the process of captioning and audio
description much easier for everyone. But one thing that
we really love to do is post free
educational resources. So this is an example
of one of those. We really like to get
people who are invested in video and digital
accessibility together and use that expertise
to help grow this space. So we have a lot
of free resources on our website, all
devoted to education. And we also have been involved
with a lot of industry research in closed captioning. And you can download
free research studies that we conducted with Oregon
State University on our website as well. And finally, one
of the other things that’s really important
to us is giving a voice to people in the
communities that we serve. So we have a project called
“Faces Behind the Screen,” which is a series of
interviews with people who are deaf, hard of hearing,
and blind telling their stories and really focusing on
all of the great things about being deaf
and being blind. And so with that,
I’m going to hand off to our first speaker,
Mike Shebanek, from Oath– formally, Yahoo– who has done
a lot of really great things with video accessibility. And he has a great
presentation for you. MIKE SHEBANEK: Hello, everyone. Great to be here. My name is Mike Shebanek. And I head the
accessibility team at Oath. I thought I would start
with a little video to give you a little
background, what we do in the company
around accessibility and around captioning. So if you don’t mind, I’ll
just play that right now. I’ll be right back. And most of you
probably know Yahoo for Yahoo News, Yahoo
Finance, Yahoo Sports– a bunch of different
[INAUDIBLE] websites. But recently, we went
through a transition. [INAUDIBLE],, so I thought I
would introduce my new company, Oath, O-A-T-H. It’s
the combination of all of the products and
services you know from AOL, plus all the products and
services you know from Yahoo. So the grand total of that
is something called Oath. And Oath represents
a huge number of things you probably have
heard of, like Huffington Post, Tumblr, Autoblog. Gosh, there’s a whole
bunch of those in there. What else do we have up there? We have Yahoo Finance,
Riot AR and VR, if you’re into Augmented
Reality and Virtual Reality. So my team is responsible
for the accessibility of all those different
brands and products. And so, if you run
into one of those and you see access enabled
for any of those products, that will lead, ultimately,
back to my team. So what we do is we are
really a service organisation. There’s only five of
us that do this work. So we’re not the ones
actually doing the code or doing the design work. We’re the ones
helping those teams understand how to do better
design, more inclusive design. How to take the best
examples of code techniques and implement those in a
way that’s going to help them make their products better. But ultimately, we’re
an ad-based company. So it’s really important
for us to make sure that we’re not dis-including
the over 1 billion people in the world who
have a disability. We can allow our advertisers
to reach more and more people if we make our products
and services– even the ads– more accessible. We are also focused internally
on making our company a great culture
in which to work. So it’s important
to us to make sure that people with disabilities
can have access to the jobs and compete in our company. So we’re working with our
Diversity Inclusion Office, our HR, our talent
organisation to bring people into the company and make
it a great place as well. But ultimately, we are supposed
to provide an expertise. So the team of people at our
company is pretty amazing. They’re pretty well
renowned in their fields. And it covers everything
from mobile app accessibility to [INAUDIBLE]
accessibility to VA accessibility, or captions
we’ll talk about today. And then, of course,
our team is also responsible for making
sure our company is compliant with all
the regulations and laws and things. But I want to just make it
really clear our job is not for compliance. So we are definitely not
a compliance organisation. We’re aiming way
higher than that. As I say in that first
bullet up on the screen, our goal is to make
things that users want, that they want
to use, that they like using, they want to share
with people, so they come back and do more. And so that’s what we aim for. And so, in our minds,
usability and accessibility are really lofting
way up at the top. If we make something
that’s successful but it’s too hard to
use or too many steps, we haven’t done our job right. It really needs to be something
that people can do with joy and that they can
[INAUDIBLE] easily. And of course, that last
goal– it’s an interesting one. We are working hard to
try and invent solutions where none exist. So if there’s a standard
or a solution, we’ll do it. But in cases where
those don’t exist, it’s our team that’s trying
to be as creative as possible to solve those
kinds of problems. So I put up this pyramid here
in case you’re the situation. You might not have an
accessibility team, might not know what it is
you’re supposed to be doing. But for us, compliance
is the baseline. That’s where you start. Most companies are working
hard to get to that. And that’s good. And that’s right. But when you can
elevate yourself to providing
accessibility that’s very high-quality
and very consistent, you’ve reached
that middle level. And that’s what most
people are trying to [INAUDIBLE], including us. But when we’re doing it really
well and doing it consistently, then we can build innovation
and create new solutions. But it’s really hard to
be inventing at the top, if you’re not doing the basics. So I want to encourage
you with that. Everybody starts at the bottom
and builds their way up. But building a new building
foundation [INAUDIBLE].. So we’re here to
talk about accessible media, time-based media. Most of us think of it as video. And when we think of
video and accessibility, we tend to think of
closed captioning. But I wanted to– since this
is our first presentation for the day– just talk a little
bit about where captioning comes from. If you’re not in
this space, it’s actually kind of interesting. Captioning is often called
subtitling overseas. Here we call it open
captions or closed captions. But there’s also this new thing
called video descriptions. You may see this
symbol on occasion. And this is where
verbal description is added into the quiet portions
of the script or performance to describe what might not
be otherwise understandable. And I think that there might
be some more discussion on this later in the day, right? Yep. So I’ll queue that one up. And then the last one here
is kind of interesting, which we often forget,
which is there’s now rules and requirements around
making things like TV interface systems, open-top boxes, like
Rokus and Apple TVs and things, have talking menus,
so that people who need that–
those are provided provided for the
new technologies. But I wanted to ask
a quick question because I don’t
know how many of you have been involved in
captioning or subtitling or if you’ve been
involved at all. Does anybody remember
where this started? On TV, right? Famous programme? I thought I would bring this
out of the wayback machine OK, famous food person– food shows
are all the rage now, right? Who was on TV that started
the whole idea of, like, I’m going to cook for you? AUDIENCE: Julia Child. AUDIENCE: Julia Child. MIKE SHEBANEK: Julia Childs,
exactly– great answer. So way, way, way back in the
TV days, WGBH of Boston– public TV– approached
Julia Childs. And they approached
her and said, we want to do this new
thing that no one’s ever done before called
closed captioning, or in this case
subtitled captioning. Would you be OK with it? And she was like, absolutely. This is great. Let’s do it. So if you happen
to be old enough and you were watching
that show, that would have been your
first exposure ever to this idea that something
in the video space should have captions to it. And of course, that begat
news shows, other things that then had captions. And at the time, of course,
something that was what we call “immediate turnaround”–
like, instantly– was, like, five hours later. So we’ve changed
a lot since then. And thankfully, we have,
with companies like 3Play, the ability to
get captions much, much faster than the
five-hour turnaround. And of course, then captions
became pretty widespread and certainly now
required on broadcast terrestrial television. But the problem
is, what happened when the internet showed up? There weren’t any rules,
any laws for that. But thankfully,
you can now watch Will Farrell, with
captions, on our phones, anywhere we want to be. So the world has definitely
improved since those old days. So before I jump into what
we’re doing for captioning, captioning, I
thought it might be interesting to get you a little
peek into what we do overall at Oath. Because captioning by
itself, for us, isn’t enough. We want to think
about, holistically, every person and
every experience and how that’s transferred
to every user that uses our product. So we’ve created
something that’s really rather unique in the industry. It’s called the
Accessibility Lab. It’s a dedicated space up
in our Sunnyvale campus. Some of the people in the
audience have been there. So if you want to talk
with them about it or ask them– maybe you
guys could raise your hand or just make yourself
known somehow if you’ve been to this lab. [INAUDIBLE] A couple
people– good. So you can pin them down,
too, if you want to know. But the idea here
is that we look across all of the
different disabilities, including hearing loss. So for us, it’s
just as important to make something available
to the blind or people with dexterity or
mobility issues, all those sorts of things. So in this lab, on
the back of the wall, there is glass, if you’re
not able to see this. We got a little library there,
with all kinds of topics, whether they’re technical,
political, social, around disability. And we bring Team [? 10. ?]
We bring designers in from all over our company. We sit them down
and say, what do you know about accessibility? And we ask a really
interesting question. And I could ask you
guys the same thing. If you have a four-year
degree, we just want you to raise
your hand if you’ve heard the word “accessibility”
as part of your curriculum. All right, one or
two, three hands. That’s normal. It’s not something
that’s taught. It’s not part of a traditional–
even a technical education. So we spend a huge amount
of time teaching people how assistive technology works,
who uses it, what to expect, what users expect. Because without that
basic knowledge, we don’t know what
we’re aiming for. So we have a couple
of stations here for hearing loss, where we do
things like closed captions and video player controls. We have other stations that
have screen readers for people who are blind or low-vision. And we try and cover
the complete gamut. But this is a place
where people can come in the safe environment
and ask hard, sometimes silly, sometimes embarrassing
questions because we’re just not often trained or used
to people with disabilities and how successful they can
be given the right tools. So I did a little panaroma
around the corner here. And you can see the big
screen TV at the top and put captioned websites
and things up there. You can take a look at those. This is a typical
day in the lab. So it’s pretty cool. I think that’s
[INAUDIBLE] and my team giving a presentation, in
this case to a bunch of guests that have come through. We’ve had over 1,300 people
come in in six months. So it’s a busy, busy place. But the result is
a different culture and a different attitude about
what accessibility means. So if you’re in
this field, these are the common issues
you’ll hear and experience from other people–
things like just making sure text is re-sizable,
things like making sure you have high contrast. If you’re looking at the center
picture here on the right, intentionally, you
can’t see the clock, because it’s a dark grey clock
on on a dark grey background. And that clock was
something you were supposed to click to tell you to remind
you when the meeting’s going to start [INAUDIBLE]. So guess what? We called the designer and
said, hey, what’s going on here? They changed it to white. Everybody started
using the [INAUDIBLE].. And that’s a lesson I
think you’re going to hear, once I get towards the
end of my presentation. If we do accessibility right– closed captioning,
in particular– it’s useful for everyone. And that’s really the
take-home message here. Yes, you’ll be serving a
community of people who are deaf or hard of hearing. But you’re going to be serving
a much bigger community as well. And I’ll talk about
some of the things we learned when we implemented
closed captioning in a much bigger way in [INAUDIBLE]. The other things that
are simple things– things like keyboard access, for
people who don’t use a trackpad or [? use ?] [? gestures ?]
[? on the phone. ?] And then, also, a screen reader
support for certain. And then, of course, the last
one is closed captioning. And so, for us,
it’s really critical that the entire experience,
from end to end, be accessible. So we opened up another
facility, very specifically, for closed captioning,
in the Boston area. And this is when we opened. It was covered by
Boston Business Journal. And the first
question they asked is, why are you doing this? Why are you spending
all this time and effort doing closed caption work? And of course–
because there’s a need. People need this
information access. And we needed a space
where we could [INAUDIBLE] We needed a space
where we could bring in all the different devices
and platforms [INAUDIBLE].. AUDIENCE: [INAUDIBLE] AUDIENCE: Sorry. MIKE SHEBANEK: That’s OK. AUDIENCE: [INAUDIBLE] MIKE SHEBANEK: Everybody
wants to get [INAUDIBLE].. AUDIENCE: [INAUDIBLE] MIKE SHEBANEK: So this
is the Boston lab. And in it, we have pretty much
every mobile and over-the-top and desktop and TV device. And this is, sort of,
our central command post, if you will
for, [INAUDIBLE] and closed captioning plus
all our [INAUDIBLE] platforms. So we have everything
from personal computers to tablets, Android
and iOS phones. We talked about Rokus and
[INAUDIBLE] game systems that we have products on for
people to stream our video. So the goals for us, in
terms of captioning– and I hope that they’re
your goals as well– is really to make those captions
available as immediately as possible. You can imagine, if
we’re doing Yahoo Finance and the captions come a day
later, how important is that? Probably not so important. Other things that
are evergreen, we can put the captions on
at some appropriate time. But our goal is
always to get them as immediately as
we possibly can when those videos are available. We also want to make sure that
the captions are high quality. And this was, for me,
a really big lesson. When I first came
to this, I thought, captions are either
there or they’re not. What does it even mean
to have quality captions? We’re going to talk about
that in just a minute. But it turns out,
there’s actually quite a number of
characteristics you’re going to want to care about. And [INAUDIBLE]. The third one is ubiquitous. People want captions wherever
they want to watch the video. And you can’t always
predict which video people are going to want to watch. But it’s part of
the difficult part of the issue you want to track
is, which videos do we caption? Do we do them all? Do we do some of them? Do we do only the most watched? And that’s part of
the art of doing this. The fourth, of course, is you
want them to be compliant. And there are
regulatory requirements around how you make
those captions available, particularly around a law
called CVAA in the US. So this was part of my lesson. We started captioning
long before we met 3Play. And we started to hear things
back from our consumers. They said, something’s
not right here. And so this was a funny SNL,
Saturday Night Live, skit. And it was Benjamin
Netanyahu on the left. And I think it was the Hulk on
the right, played by the Rock. But the captions didn’t
read quite right. It says, “Do you invite net
on Yahoo without asking?” because the company
thought, since we’re Yahoo, that must have
been what it meant. They didn’t get Netanyahu. And so, wow! That’s bad. That’s just bad for a
whole bunch of reasons. So that woke us up to, wow. We really need to think
harder about the quality of our captioning, not just
whether it’s done or not. This is a show called
Community, which was on television
[INAUDIBLE] the internet. And we were running that. And there’s a scene where he
says, “Oo, you’re a raconteur.” Very edgy, right? Except the captioner said,
“Oh, you’re a rock on tour.” That’s enough. So it can be a
little embarrassing. We all make our mistakes. But we want to learn
from those, correct those as soon as possible. So that’s when we started
reaching out and thinking, we’ve got to do a
better job at this. We need to be thinking harder
about this, what the real user experience is going to be. So that’s when we
came across 3Play. And they helped us
understand those things and really helped our game,
in terms of the quality. So now, if you go to
something like the Yahoo– and this is Yahoo Finance– there’s actually some green on
black captions across the top here. And it’s much more accurate. It’s much more timely. In fact, this particular feature
is one that’s pretty rare. Most of the time,
captions are going up here across the bottom of the screen. But if you have the
lower third, as it’s called, where you have a text
or a ticker running across, don’t overlay information
that the users want. So we built in a feature
that’s not required. It’s not even requested. We just did it because
we thought it was right. And it’s the ability to move the
captions to a different portion of the screen. So we can place them on top. So if people prefer those, they
won’t overwrite the information that might be occurring at
the bottom of the screen– emergency broadcast information,
ticker tools, whatever it happens to be. I also have another
one of these. We work with a lot of partners. So we’re doing a lot of this
captioning for ourselves. But we also bring captioning
in-house from others. This is an NFL football game
that was captioned recently. I don’t know if we
have audio on this one. We’ll give it a try. But if not, the
captions are there. So you can watch this way. This is a helmet-cam view of
the game, which is kind of fun. This is using a different
technology for live captioning, if you will. So we talked a little
about the quality issue. And this, as I just described,
is truly really important to us. Being timely, as I
said, for finance is obviously a big deal. But even for news stories, being
as timely as you can possibly be, providing that video and
the captioning at the same time is really, really critical. Accuracy– I think
we already covered. And you got a [INAUDIBLE] that. Synchronise is another one. So sometimes, you’ll
have perfect captions, but they’ll be appearing
after the scene. Or they might appear
before the scene. Now, if you’re watching a
drama or a mystery, or a joke, and the caption, the punchline,
appears before the joke happens and everybody’s laughing,
it doesn’t really work. So making sure that
that’s synchronised turns out to be really,
really important. Similarly, sometimes
captions start late. So the first part, the
introductions, aren’t there. Sometimes, they end early. Sometimes, they drop
out in the middle. So it’s actually
really important that, as you design
your systems and you’re doing [INAUDIBLE]
captions, if you’re watching for these events
throughout the entire presentation, then,
of course, make sure that they’re complete. So as I mentioned,
we do a lot of our own original programming. And so for everything
that we make in-house– Yahoo News, Finance, Sports,
everything– we ensure 100% have captions [INAUDIBLE]. So you play [INAUDIBLE],,
you get those. As I also mentioned, we
work with a lot of partners. And these are some amazing
name-brand companies that produce a lot of video. So we’ll work with them to bring
the captions they provide in. In some cases, we’re
actually captioning videos that they give us that
they don’t caption. So we’re, at the
moment, the only place that we know of where you
can get a New York Times video with captioning. The New York Times
doesn’t caption that. And so we’re talking with them
about how to get that in place and what that might look like. The other interesting
thing, too, is we asked a lot of users,
what would you like to see? What would you want
to have captions? You know what they told us? It’s an amazing thing– movie trailers. It’s kind of amazing that movie
trailers don’t have captions. You can go to a theatre and ask
for assisted listening devices. You can ask for
all kind of things that will present captions. But if you want
to see a trailer, you got to come to Yahoo. That’s the only place you’ll
be able to actually preview the movie you’re going to
see with captions enabled, which is kind of cool. I always get asked,
so how big is this? Is it just a little bit of work? Is it a lot of work? Like, how many videos
are you actually doing? So I put up a recent order. I won’t say which one. I got to be a little bit
careful about disclosure on some of these things. But in one quarter, in
90 days, to give you a sense of the scale, we
were doing 40,000 videos– 40,000 videos. So yes, it can be done. I don’t know if any of you have
40,000 videos you post online. But trust me, it’s possible. 120,000 video-on-demand
minutes, and over 200 live hours of streaming– so we’re really
at a scale that’s colossal. And it’s been interesting
because the initial conversations we had, where
it’s going to cost too much, it’s going to be too hard
to do, and there’s just too much volume, we can’t do it– and slowly but surely,
we’ve proven that case that you know what? Start small. We’ll get more. We’ll improve our quality. We’ll expand our coverage. And now, we’ve reached a place
that we’re pretty happy with. But we still have
more we want to do. So we’re not done yet. But I think we’re at a good
[INAUDIBLE] along the way. I mentioned this
law called CVAA. It’s relatively new. It’s the 21st Century
Communications and Video Accessibility Act. That’s a mouthful. So we just call it CVAA. And it has requirements around
things that were previously broadcast on television
that come to the internet and that, within 15 days, you
need to, if you screen those, make sure captions
are also provided. So there’s some teeth
in that you want to make sure you’re covering. But I think this is a
really salient point. And I hope this
isn’t irking to you. If we only applied the
compliance guidelines and said, what’s the minimum we have
to do to not get sued– 1% is what’s required. 99% of what we’re doing, we’re
doing because we want to. And I’ll tell you why
we want to in a minute. But that’s a really
important thing. Aim high. You will be well rewarded. So nicely, one of
the advocacy groups called The Hearing Loss
Association of America saw what we were doing, was
tracking all these things and the changes that
were happening– was gracious enough to give us
an award in 2016 when we were Yahoo really acknowledging
all the work that was going on in the closed captioning. But I also want to make
it clear we’re also doing work in Video Player. Because it’s just as important
important to make sure that that captioning button is
easy to find and that it works and that all the settings
are there and that people can get to it, not just
that the captions exist. But for us, the
return of the reward was way bigger than the
reward we were aiming for, which is to enable
people with hearing loss to be able to enjoy all the
videos the rest of us enjoy. So the first one, certainly,
is to serve that community. But the other one
that’s very interesting is, when you put text on screen,
that video gets really sticky. And that’s an insider
term for people watch it. And if you’re in the
internet business and you have stuff
that you’re streaming and you want me to watch,
because that keeps them on your site longer or that
allows you to sell an ad and generate more traffic,
that’s a big deal. It’s a really big deal. And this is not just sticky
with people with hearing loss, this is sticky across the board. So it lengthened
engagement and the number of people that engaged with
the video and their scrolling through a screen or
clicking on their phone is really, really high. And that was something
I don’t think we were quite prepared for,
but something that has really, really helped us overall. We’re looking at some numbers
in the industry that tell us that something like– and
maybe Josh or [INAUDIBLE] could confirm for us if you’ve
seen something different. But like, 85% of the
news, when people are going through their
website or a mobile app, those videos are on mute. And if you don’t have
text on that screen to entice someone to know what
that video is talking about, the chances of
holding that person and conveying that
information is a lot lower. So in the given environment
we’re in, again, having captions that are
available on-screen– playing mute allows people to
enjoy those in environments where they don’t have headphones
or they’re in a public place and can’t use the speakers. Again, you’re
increasing engagement across the board for everybody,
which is really cool. The other one that
was really interesting is we have, as you can imagine,
a huge number of editors who sift through all this
video and all these articles for Yahoo Finance, Yahoo News,
Sports, all these things. And we didn’t realise how much
work they were going through to capture the text of every
video so they could go back and find it later. So when they found out we were
doing all this captioning, their first call was to our
team, going, how do I get this? Because now, I have a
complete record of everything that happened in that video. And I can either
copy and paste it into a press release
or marketing piece or share with people what
it is we have on our videos and cross-reference it later
and archive it and find it. That turned out to be a
humongous savings for that team and, again, something
that we hadn’t really anticipated that makes– now,
when we look back at it– perfect sense. So some huge returns
on that for us. If you have it in you to
translate the captions– usually, there’s a
baseline of English. But you can translate
to other languages. [INAUDIBLE] to 3Play. You have the option
now to open up markets you didn’t have before. Instead of re-shooting an
entire video– it’s a part of it you might not even be able
to– but if you can provide it in multiple languages,
suddenly that video is useful in a lot more countries. And you want to be more
global or international– [INAUDIBLE] multi-lingual
speakers in the States– this is a great way to start
getting that total [INAUDIBLE] worry about your getting
different production for different videos. And the last one, for us, is
a really, really good deal. As I mentioned, other
editors– once we have a text, it’s incredible how
we can search on that and personalise for
people [INAUDIBLE].. But what’s even more
interesting for us is we have a time code to
go with the text. So we know where in
the video someone wanted to share something
or stopped to look or find the information that
they cared about or they’re sharing
with a friend. And there’s some amazing
opportunities to, if you have time-code
based text of your videos, to go back and
personalise for people, to do searches for people, and
to create a context for what people are watching
that are really unexplored at this point. So more to share with you in
the future, I hope, about that. But that, for us, was
really, really exciting. So we have a set of values
at our new Oath company that are very explicit for us. And one of those stated
values– that every new hire comes in– when they come to
our company, they see this. And it’s why, on
their first day, new-hires come and get their
laptop, they get their badge, and they visit that
accessibility lab. That’s one of the
first things they hear when they come to our company. So when we say we’re committed
to inclusion as a value, we really do mean it. And we’re heavily invested. And we’re so thankful to
have companies like 3Play help us in that [INAUDIBLE]. Thank you. [APPLAUSE]

Author:

Leave a Reply

Your email address will not be published. Required fields are marked *