The Future of the Web: SEO Mythbusting
DION ALMAER: I feel like
there's this weird, interesting spectrum on the web where,
on one end is content, and the other is app. But it blends at
a certain point. MARTIN SPLITT: It
does blend, it does. DION ALMAER: And it's that
blending part that I find is always interesting,
where I'd hate for us to be building
something in a way that, then, search can't
understand that piece. MARTIN SPLITT: Right. Because you want to
find the content that lives inside the applications,
and you want to still find the applications, right? DION ALMAER: Right. [MUSIC PLAYING] MARTIN SPLITT: Hello, and
welcome to "SEO Mythbusting." Today, I have Dion
Almaer with me.
What do you work on? DION ALMAER: I work in the
Web Developer Ecosystem group here at Google. And so that's within
the Chrome organization, but we work across Chrome,
and search, and ads, and all of the different
areas of Google that care about the web. Personally, my passion
is about the ecosystem. Hence the name. So it's about looking at
the health of the web, and what are we doing to
really help developers, but even beyond developers–
hence the SEO world and the like– to make sure that people
are being successful. So that could be successful
in building amazing user experiences, but it's also
successful in actually being able to pay the
bills, and so they can keep making those
amazing user experiences.
MARTIN SPLITT: That is great. So when I have you here today,
I would love to talk about SEO and the web platform
going forward. So we want to look into it,
like, what are the new use cases, what are the things
that are coming up next, and what are the things
that we from the search side should be preparing
for, let's say. Do you have any highlights
that you want to talk about? DION ALMAER: Yes. So I gave a keynote
at the Chrome Dev Summit. in November 2018. And it's kind of
going through sharing all of these great, amazing web
platform updates and tooling updates.
In the back of my mind, you're
always thinking about, like, are we getting closer to
search and helping the word? Or are we causing more
problems going on? MARTIN SPLITT: Yeah, it's– DION ALMAER: I'm actually
really curious to get some of your thoughts
on some of these items. MARTIN SPLITT: Amazing. Do you have any
specific things that you would like to address there? Like is there any
showcases, or anything that we should look into
from a search perspective? DION ALMAER: Yes.
So one case study, actually,
that we showed was Pinterest. And it really got me thinking,
because they came along, and their growth team
built this new PWA that– MARTIN SPLITT: That's fantastic. DION ALMAER: –was–
it's amazing. Like the performance
is top notch. It's a great experience. So they've got this great
PWA that's really exciting. Then they've got their kind
of regular desktop site. They've got– they're using
an AMP for different things. And so if I'm a team like
that, how do I, like– I've got all of these
kind of different sites, in a way, which is almost like
the opposite of the promise of the web at times. Like, build it once,
and it was everywhere. So, like, which of
these– how do– what's the one that I
want to give to search? Like how do I think about that? MARTIN SPLITT: That's
a really good question. So I think, generally
speaking, if you have the same content
and multiple ways that you present
them to the user, find out the one that
is most likely to apply to as many people as possible.
So if you have
one specific thing for, let's say, like, sign up
users, to actually remix it or something, that
might not be the thing that you want to
surface to search. Because me coming into it
from outside and not having ever used it, and
I might not even be able to use the
feature, and then it's like why am I getting here? So try to think
of what are people typing into a search bar? What is the purpose
they are coming from? I want to see this image
board that Dion has created for a Chrome Dev Summit.
So I might type in, like,
image board dev summit. And then I want to
get the PWA probably, because it loads fastest,
and it is available even if I'm offline,
and stuff like that. So give me that. But also link to
the end version, because then we can also show
the end version next to it. So it's not really
a ranking signal, so we might see, like,
different rankings there depending on the purpose and
the user characteristics, what they type into
the search bar. But if you link them, then we
can make the decision for you so you do not have to worry
about that kind of stuff.
If you then have a desktop site
as well, that is great as well. If it's responsive,
that's even better. If it's not responsive, then
give us an alternate link so that we can decide, OK, so
this is the Google crawler that uses mobile first,
so we're probably going to highlight
the PWA experience over the desktop site. Versus, if someone's
using the desktop site, we might give them the other
version rather than that. But we are trying to make the
right decisions for your user so that you don't have to
worry about these things. But, strategically
speaking, give us a canonical that makes sense
to as many people as possible so that we can surface the
related information based on what people are looking for,
rather than any differences in the presentation.
DION ALMAER: Got it. Cool. The other thing
that comes up a lot is integrating
our tools together instead of having it in silos. So we've already been working– we've got Lighthouse over here. We have PageSpeed Insights. We've got search. We've got all of these
different pieces. We announced at
Chrome Dev Summit the unification
of some of these. The PageSpeed Insights now
uses Lighthouse and the like, and that's super exciting. Wayfair, another company,
built this performance portal to kind of help the team make
sure that they don't regress. Because we see that, a large
percentage of the times, people will, like, do an
investment, have a quick burst, get their performance good, and
then it will slowly go down. So they have this awesome
performance culture that's trying to fix that. And I'm just looking
at that, and thinking about it's yet another example. Like here's all the
performance stuff. Couldn't we be surfacing
the kind of SEO pieces into that too? How do we bridge these
worlds together– MARTIN SPLITT: That
is [INAUDIBLE]..
DION ALMAER: –so
both sides can kind of see the other things
that are going on? MARTIN SPLITT: So Lighthouse
gets more SEO audits, which is great, I think. That probably, at some point,
feeds into PageSpeed Insights as well. Search Console integrates
some of these metrics as well, because performance
is also important for search's core ability. We want to, like,
reward the content that gets you– that gets the
user quicker into the content rather than having to wait
forever until something loads, especially on mobile networks.
But for external tools
and external companies who want to dig into
the data, I think Search Console is a
great way of already doing that in a packaged way. But, but we are working
on ways of integrating external parties and external
content providers and platforms to get the data that we
are collecting already for Search Console. So, for instance,
if you know one of the many larger content
management systems or platforms that exist out there where
people are creating content, we don't want them to
have to specifically go into Search
Console, and now, like, deal with the interface
they are used to, and then deal with
something new.
That's why we are bringing
the data into these platforms. And, eventually,
hopefully, once we have gathered enough
information to understand how the data is
used and what data is necessary and meaningful
to external parties, we will open these interfaces. There will literally be an
API to integrate the data from Search Console, which
gives you information like click through rates, how
many impressions you get from search, how many
clicks you get from search, all this kind of thing, how
many pages of your pages are indexed actually. Because not all of
it might be indexed. There might be an issue. You might have a markup problem.
We want to surface this
where people already are. That can be an IDE integration. That can be something
like Wayfair does. But, at this point in
time, we are too early on to open that up to
the general public, but eventually that's
going to happen. DION ALMAER: That's so exciting. And some people were talking
about how, like, they want to understand
the unknown things. Like they wonder
where like, what's– if I do this,
that, or the other, how does it affect
my customer lifetime value, and these things? And we keep running into
something on the performance side where they'll do
this work, and they'll make something
that's a lot better, but their metrics actually show
in the field that it's slower. And we're like, what's going on? Like, in the lab it's faster.
It's obviously faster. I can see it's so much better. How is it slower? And then it's because
the reach that they got just totally changed. So these people are
on, like, 2G networks in emerging markets that can
actually really use this now. MARTIN SPLITT: That's
moving targets. Then you're comparing
apples to oranges. DION ALMAER: Right. So you've got all of
these new customers. And so the mean metric isn't
the thing that matters there. Now you've reached
this new audience. You get more revenue,
and the like. So, like, how do we know,
like, with the search side and discovery, like,
what are the things– the next things I should do? Which shall I
prioritize, and the like? MARTIN SPLITT: Right, right. I think a bunch of stuff
that Search Console gives you already is
kind of handy for that. You see how certain
content works, and how certain other content
might not work so great. And that can have technical
difficulty– sorry– technical reasons. Or it can be actually
content reasons. So what we are
trying to do is we're trying to also broaden the
perspective a little bit.
I mean, you said that. Like, oh, yes, we're in
this metric right now. We want to improve performance,
but actually it has degraded. And it's like, well, it has,
but it's not a bad thing, because what actually happened
is we opened up to more people. That happens with
search as well if you look at it from a more
holistic perspective. So if you're not just looking
at technology but also at your content strategy. And what I would
love to see happening is that, if we bring this data
to where people already are, that people become
more aware of it, that maybe you should also look
at your content and the way you present it, and
the way you make sure that you're talking to the
person on the other end who's trying to solve a purpose. You're looking for something
specific rather than necessarily a specific product. It's like I need this task to
be gone and done and dusted, and how do I do this? So I think bringing the data
to other people will help.
And I think what
developers should look at is how their content
is performing in search engines as well. So how often do I show
up in search results? Where in the search
results do I rank? How many clicks do I
get from search results if the content looks like
this versus like that? So that's something that I
hope to bring to more people as well. DION ALMAER: Cool. Yes, that makes sense. So web components has
been another big topic. MARTIN SPLITT: Oh, right, yes. DION ALMAER: And we released
an early version of a new one that I'm super excited about
called virtual scroller.
And, basically, for
years, web developers have been saying, like,
give me a UI table view on the web
that can perform, and scales with the DOM. And so we feel like we're going
to be able to deliver that. But then what's that going
to mean in the SEO world? Is search going to be
able to understand this if it's all virtual? And, like, as we kind
of push the bar here for performance and
user experience, well, how do we make sure we
don't leave it behind? MARTIN SPLITT: I think
what a lot of people outside of our
sphere don't realize is that we are
working hand in hand.
It's not like SEO– or search versus
the browser teams. We are trying to figure out how
we can work with a web platform rather than against it. We are using the web platform. We are basically running
a browser, if you wish. So we are actually
looking into things like components and virtual
scroller, specifically. And I think it's a
fantastic thing, because we have literally just released
documentation for lazy loading, or infinite scroll, whatever
you want to call it. Basically– you know the drill– I can have a page with a
bazillion images or product listings. You can't have them all on
the page right away, right? You want to make sure
that, as the user goes through the content,
they come back.
Well, how often did it happen
to you that someone sends you a link, says like, have
you seen this product? And then you go to the link
and you see a lot of things but not that product. And then they go, oh, yeah,
you have to scroll down for a few minutes. That's not good, right? Web components are another
thing that we are, like, really careful about. And we have some guidance
out there already, and we expect to give
more guidance as well. Because if you build new
web components from scratch, they look like they're semantic,
but they are not necessarily semantic in the strong sense. If it's a button that just
happens to look like a button but really is a
diff underneath, you lose a lot of the
accessibility features, and so does search
also loses a bunch of the semantics behind it.
So, right now, we
are recommending putting the content, the
meaty bits of your content, like the actual, well,
what it is about, put that into the
light DOM, and then to use the shadow DOM and your
components for presentation. Virtual scroller is a specific
interesting case where not only is
presentation but also the way of how the content gets
loaded and the content behaves. So we are working
on figuring out how to make that
consistent with search so that you don't have
to worry about using these new technologies, because
we're really excited about them as well. DION ALMAER: Yes. No, that's great. How do you feel about– I keep longing for a
world in the future where we add, with
web components, these different semantics,
kind of high-level semantics that search can understand but
also assistants can understand, so we can get further into
transactions and the like, with the notion of
I really want to be able to go onto a
commerce experience since I want to buy
a T-shirt for Martin.
And everything is brokered in a
way that it knows your T-shirt size even if I don't. You know what I mean? MARTIN SPLITT: Yes. DION ALMAER: And
there's a person tag, and all these different things. And then you get
a great UI from it too, because it's standardized. But then search could use
that, and assistant could do full transactions if we do it. And it feels like that would
be a way for assistants to kind of reach the long tail,
versus trying to go to everyone who's building a website.
It's got a lot of
work to do as it is, and we're asking them to
modernize this, and fix your performance, and
do this for discovery, and also build this
thing for the assistant. MARTIN SPLITT:
Yes, that is true. DION ALMAER: Do you feel
like we'll ever get there? MARTIN SPLITT: I think we
get there probably quicker than we think, because,
conveniently, we have already pushed the idea of using
structured data to expose, like, a bunch of information
on your pages in a way that not only our crawlers
can understand them, but also anyone
else who basically parses this information. And we are using
schema.org for that. So we use, like, a
standardized format rather than just a proprietary thing. So I hope that
gets more adoption and we see it both
picked up by developers and by these assistant systems.
Because, then, you can use
the web really powerfully because you have all the
semantic data that you can just pull together, and then
package in such an experience as you explained, like the
full end to end commerce, for instance. DION ALMAER: Cool. Nice. MARTIN SPLITT: Man,
thank you so much. DION ALMAER: Thank you, Martin. MARTIN SPLITT: Thank you. See? Development and SEO don't have
to be contradicting each other. Also, the iced tea
is pretty not good. [LAUGHS] Pretty watery. But it has the
fantastic right color. Like, you picked the
exact, correct prop. [MUSIC PLAYING].
