All questions have Show Video links that will fast forward to the appropriate place in the video.
Transcript Of The Office Hours Hangout
Click on any line of text to go to that point in the video
JOHN MUELLER: OK.Welcome everyone to today's
Google Webmaster Central OfficeHours Hangout.My name is John Mueller.I'm a Webmaster Trends Analyst
from the Zurich office,and I'm currently in the
Mountain View office,together with two outreach
ERIC: Hey, guys.
BARUCH LABUNSKI: Hi, I
think we're not live.
We're not live yet?No, we're live.It says Live on--
BARUCH LABUNSKI: Here.
JOHN MUELLER: I
think we're live.We should be.
BARUCH LABUNSKI: OK.
JOHN MUELLER: And Mary.
MARY: Hey, guys.
JOHN MUELLER: I have one thing
I wanted to announce to you guysand get ready, but I actually
didn't get it quite finishedyet.But we're going to be doing an
alpha preview of a WebmasterTools feature for a
search query datum.And I'll post the
link to a forumwhen you can sign up
if you're interested.We'll probably have the
forum open for a week or two.So if you're
watching this later,you have to be pretty quick.Add your name there.It'll be something neat that
we're playing around with,that we're trying
out, and we'd loveto have as much early
feedback as possible.So it' d be great if you guys
find time to check it out,sign up.I'll probably have the
link there maybe an houror two after the Hangout.Anyone interested?Yeah.Gotcha.
JOSHUA BERG: Yes, I'd
like to try it out.
ARTHUR RADULESCU: Me, too, John.Can you drop the link
in the chat also?
JOHN MUELLER: Well, at first,
I have to finish the forum.Then I'm--[LAUGHTER]But that should be ready soon.I am guessing that'll be open
for testing around next week.So if you sign up, we should be
able to set you up with that.And this is a very
early alpha version,so things are going to change.And it's really useful to
have your feedback early sothat we can change
based on your feedback.All right.Does anyone want to ask a first
question before we get started?
BARUCH LABUNSKI: [INAUDIBLE].
ARTHUR RADULESCU: It's only me,
or does Baruch sound robotic?[LAUGHTER]
ROBB YOUNG: Stephen Hawking.
JOHN MUELLER: Try
it again, Baruch.
BARUCH LABUNSKI: [INAUDIBLE].
ROBB YOUNG: No.
MARY: Not going to work.
JOHN MUELLER: Yeah, maybe we
can get back to you afterwards.Anyone else with a first
question before we get started?
MALE SPEAKER 1: Yeah,
I have a question.So can you guys hear me OK?
JOHN MUELLER: Yes.
MALE SPEAKER 1: So I work at
a company called redfin.com.We do real estate.And traditionally, we built
guys made in this crawler.Some of our pages are
being crawled by Googleand indexed perfectly,
it that some of our pagesare being indexed by a
JS crawler perfectly,and then others are not at all?
JOHN MUELLER: What I'd do there
is try it out with WebmasterTools, with the Fetch
and Render feature,to see if there's anything
systematic with those pages--if maybe individual requests are
timing out, if it's taking toolong to load all the embedded
content, the scripts,those kind of things,
you should seesome of that in Webmaster Tools.
MALE SPEAKER 1: But
the strange thingis when we use the
Fetch and Render,the page renders perfectly.But when we try to search for
find any of the content.So we don't feel like
what Fetch and Render shows.
JOHN MUELLER: OK.It might be that we're crawling
slightly differently and kindof timing out on some
of those requests.Maybe from a technical point
of view, we could crawl them,so they're not roboted.But maybe we're timing out
with individual requestswhen we try to get too much.It might also just
be that this is stillkind of early days
that these kindof subtle differences, they'll
play a role a little bitin that we might be able to get
some of the content in properlyand render it really
well for a search,and other content we
can't completely get in.Maybe we don't have the
capacity to crawl your websiteas much as we might want to.These kind of things
could be happening there.So if you're patient,
that's somethingthat I assume will get
a lot better over time.If you feel that there's
something systematic that we'rejust not picking up
certain kinds of pagesthat you can look at
in Webmaster Toolsbut which don't really show in
Search, then if you can send mesome sample URLs,
that would be great.And we can take a
look with the team.
MALE SPEAKER 1: Yeah.How do I actually
send you those URLs?
JOHN MUELLER: Just send
me a note on Google+,a private note.
MALE SPEAKER 1: Just
message it on Google+?OK, sounds good.Thank you.
JOSH BACHYNSKI: Hey, John, I
have a quick question for you.
JOHN MUELLER: OK,
go for it, Josh.
Very recently, Barryhas mentioned that he is
seeing an uptick in changesin the Google
[? Search, ?] as if therewas some kind of
algorithm release.I'm just wondering
if you could--I know if it's Panda or
whatnot, you cannot confirm.But I was just wondering
if you could confirmif it was Penguin,
and if Google stillhas plans to run Penguin
on a monthly basis,or if those releases
will be announced.
JOHN MUELLER: We make a lot
of changes in our algorithms,so that's really hard to nail
down to something specific.I don't think it's related
to Penguin or a Panda updatethere at the moment.
JOSH BACHYNSKI: But it's some
kind of algorithm release?I'm just wondering
if you could--I don't know if it's Panda or--
JOHN MUELLER: We make tons of
changes, so it's really tricky.I can take a look at
what Barry wrote up,and maybe he has some sample
URLs that we can look atand sample queries.But we make hundreds
of changes every year,so you're bound to
see some fluctuations.Some sites are bound to
see these fluctuationsa little bit
stronger than others.And maybe that's just, like, one
of the many updates that we'vedone this year is something
that these specific sites areseeing.But it's really hard to say.
JOSH BACHYNSKI: Of course.Can you confirm
whether or not Googleis committed to announce
when Penguin will be releasedin the future, or
are you moving tothe monthly
JOHN MUELLER: We're
generally movingto a little bit of a
faster update cycle there,but I can't say that
it'll be monthly.There might be some changes
there, some subtle differenceswith the timing.It should be faster
than the last one.So not as juicy as you probably
were looking for in the answerthere, but we're working on
kind of moving that a little bitforward.
JOSH BACHYNSKI: OK.And so just to be clear, it will
be announced when it releases,or it will not be announced
when it releases in the future?
JOHN MUELLER: I can't
promise either one.I think if it takes longer,
if there are bigger changes,we'll definitely announce them.If this is just subtle
update that happen over time,then we probably won't be
announcing them individually.
JOSH BACHYNSKI: OK, thank you.
JOHN MUELLER: All right.Let's go through some of
these questions on the side.Any here that you
guys want to answer?
ERIC: I saw one-- sorry.I saw one question about
someone's site who got hacked,and they had malware found.And then there were a lot
of back-links appearingfor the hack sites,
and they're wonderingif they should disavow.They're talking a little bit
about ranking change becauseof the hack content.I think that's a
really good question.I think a lot of people,
when they are hacked,they're trying to
figure out what to do.And getting hacked is
very disheartening.So the first thing is
definitely make surethat everything's cleaned up.And so if you've got
everything cleaned upand you file a
reconsideration requestor after we recall
and re-index, youdon't see a Hacked
script label anymore,then you're probably good to go.The second thing
would definitelybe to secure your site.You don't want to
be re-hacked again,because we see a lot of
people get re-hacked.And if you are seeing a lot of
these back-links, in general,we're really good at knowing
that these back-links arefrom hacked sites or bad sites.So in general, you don't
really have to worry about it.But if you do identify it,
you should just probablythrow in the Disavow
file just in case.I think that's just
general good practice.So yeah, if you do identify
a lot of those links,go ahead, throw in the
Disavow file just to be safe.And in terms of ranking
changes, any time somethingchanges on your site, it's going
to take a little bit of timefor your ranking and everything
to get updated again.So it might just
take a little time.Just be patient with that.But in general, if the
hack content is gone,you should be good to go.
MARY: Yeah.I can take the next one.
JOHN MUELLER: All right.
MARY: So the next
question is, do wesocial signals as
a ranking pattern?So social signals are really
good for friends or peoplesearching online for certain
things, like restaurants.But they're always changing,
and they're consistentlybeing updated, with
people hustlingto find things all the time.And it's changing constantly,
daily, or by the minute.So we don't use it
as a ranking factor,but it's still very
useful for your usersor for visitors to your
restaurant, to your site,to your blog.
JOHN MUELLER: All right.Here's an important one."Is there any technical
border for thin content?For example, 300 words
is thin, but 305 isn't.How can we recognize that
content is a problem?"Any of you want to handle that?Are we OK with 307?
ERIC: 303.125.No.So I've worked on the
Search Quality team before.So we do see a lot
of these sites.It's not about a specific
number that you're looking at.It's in general just looking
at your user experience.If a user finds that
duplicate content is-- it'snot good to look at.If it's very similar
to something else,a user probably doesn't want
to see duplicate results.So in general, I wouldn't
be focusing too muchon a hard number or,
like, a specific technicalspecification.If you're really concerned, run
it by a couple of your friendsand be like, hey, I
have this article,and this article, I
guess, is similar.Would you take a look and
see if it's a good article?Is it significantly different?If you're taking one article
and kind of re-spinningit or rewriting it,
I think in general,maybe then you have a
little bit of a problem.I don't think you should be
doing that in the first place.Just make sure that it's content
that your users want to see.I wouldn't focus too much
on a specific, hard number.
MARY: Yeah, the
core of the issueis, are you making the site so
that people can just find iton a Google search, or
are you just making itfor search engines, or are you
actually wanting your usersto find valuable information?And if that valuable
information from the online sitewill allow you to have separate
thin content, that's better.
JOHN MUELLER: Yeah,
so there's no number.Sorry.But I think that's also
useful, because then youcan focus on your content
itself instead of tryingto count the words,
because that's not reallya good use of your
time probably.All right.What do we have?"Is external link data
in Webmaster Toolsdependent on whether the links
have been clicked or not?In other words, do
old links suddenlyget surface in Webmaster Tools
because first-time clicked,and therefore added
to the visible pool?Just chasing a
logical explanation."No, the link data
in Webmaster Toolsis really based on the links
that we find by crawling.We don't look at
whether people actuallyuse these links to
go to your website.That's something you could
look at in Analytics,for example, to see if these
links are actually useful.So if you have an
advertisement somewhere,and that's maybe a
no-follow link somewhere,and you're wondering if
it's worthwhile to keepthis kind of
advertisement there,then you could look
in your Analyticsand see if that's
actually somethingwhere users are
going to your site.And perhaps even taking
it a step furtherand seeing if users are
actually doing somethinguseful on your site, if they're
converting, if they're buyingsomething from you, if
they're coming back-- well,maybe not if
they're coming back,but how they're going
through your website.So from that point
of view, I thinkwhether users go through a
link to your site is somethingthat's useful for
you, as a webmaster,but not necessarily
something that weuse for finding
links on the web.We're pretty good
at crawling the web.We've been able to follow links
for a really long time now,so we have a lot of experience
with looking at HTML pages,pulling the links
out, and showing them.OK."When will we hear if
any of these suggestionswill be added to Webmaster
Tools and linked to the GoogleModerator page that we put up?"
MARY: We've been looking
at it pretty often.
ERIC: Yeah, we've all
been looking at this.There's a lot of really
MARY: And it's not just
for the English market.We got a lot of
suggestions and feedbackfrom our other markets,
too, so we've been looking.
JOHN MUELLER: I think
this is something that'sreally useful for us to do
the longer term planning,but it's not something where
you'd see us going in thereand saying, oh, this is
top-wished item here.Therefore, we'll
implement it next week.We really have to plan the
work with our engineers,with the back-ends
that we have to run.So it's not something where
you'd see immediate changes,but it's definitely
useful for usto understand where
you guys see the issuesand where you guys think that we
could improve Webmaster Tools.And like Mary mentioned,
it's really interestingto look at this also
across different languages,because some languages have
completely different wishes,where they say, well,
maybe in English,this is, like, the
number one item.But in French or in
German or in Russian,it's on page two of all
the wishes that they have.So it's going to be interesting
to balance all of those thingsand work with the
Webmaster Tools teamto find a way to get
everything working there.
ERIC: Is that Moderator
page still open,or have you closed that?
JOHN MUELLER: Maybe.I don't know.
ERIC: If it's still open,
keep on throwing some stuffin there.I've seen some
really good ideas,so if you haven't
seen that one yet.
JOHN MUELLER: All right.Question-- "if a home page
without a trailing slashhas a canonical with
a trailing slashbut the canonical is redirecting
301 to the home page again,can that be a problem?"So I think in general,
in a case like this,you're giving us
inconsistent signals.On the one hand, you're
saying that, actually, the URLis like this, with a slash.On the other hand, when
we try to crawl that page,you're saying, well, actually,
it's a URL without the slashthat you want to index.So that doesn't cause
any problems in the sensethat we won't index the site
at all, that we'll demote itin rankings, but it'll make
it harder for our algorithmsto pick the right URL.So on the one hand,
we'll see this URL.On the other hand,
we'll see the other one.Maybe we'll fluctuate
between the two,depending on when we crawl.But that doesn't change how
this content actually ranks.So it's kind of inconsistent
in that you don't tell uswhich URL you want
to have shown,but it's not going to cause
any problems in the sensethat your site's going
to drop in rankingsor that our algorithms
are going to think, well,this webmaster doesn't
know what he's doing.Therefore, we're going to
ignore his site completely.It's essentially
just a choice of URLsthat you're giving us
confusing information about.It won't cause the
rankings to drop.All right.
JOHN MUELLER: You
want to take one?
MARY: Yeah, I can
answer this one.So we're very good
about understandingwhether your
subdomain is relatedto the domain-- like
if you have contextthat's specifically different.But at the same time, if
you are on a free hostand the free host has a lot
of spammy sites, for example,then we may take manual action
on the entire free host.So it depends on what kind
of free host you're at.And if you see that the free
host has a manual actionand it's affecting your site,
then you should also probablyreach out to the free host,
the webmaster, or their contactlist to ask them to
just clean up the siteor clean up the
spam on their end.So it could affect you.But at the same time, we're
also good at understandingsubdomains and that they're
related to as separate sites.
ERIC: Yeah.I don't know if this
perspective from someone runningthe free host or just
part of the free host.If you're running the free host
MARY: I think
they're part of it.
ERIC: OK.If you're running
the free host, youshould be trying to get rid of
that spam as soon as possible,because there's
probably just like-- wetry to be really
granular in our actions.So if there's something
wrong and it'saffecting the
entire site, there'sprobably just an
egregious amount.And if you're part
of the free host,I would definitely
tell your free hostthat, really point
that out to them.Just be like, there's
a lot of bad content,and the people
using your serviceis really being
JOHN MUELLER: All right.There is a very
page-specific question."Would Google consider this
to be a quality article?"I'd have to take a
look at that, but Idon't know if that's
something that we'dbe able to really tell you
that this page is fantasticor 50% OK.That's probably not
the level of detailwe'd be able to give
you information on.I don't know what URL that is.I'll take a look
at that later justto make sure we're not
missing anything exciting.All right.How about this one?"I have a domain.I have domain.com/uk/ie/us, and
I want to move to country codetop of the domain.70% of my users are from the
UK with domain.co.uk for UKand domain.com for the US.I lose all rankings for the UK.Should I use a 301 redirect
from .com to UK and a subdomainfor the US?"That sounds complicated.So essentially, if
you're moving a sitefrom different subdirectories
to new domains,then you should really
have the 301 redirecton a page-by-page basis
from those individual partsof your site.If you're using
something like .com,so generic top-level domain
for individual countries likethe US, like you mentioned
here, then make sure you useWebmaster Tools to set the
geo-targeting for that domain.If you're using the .co.uk
domain and hosting the UScontent in a subdirectory,
then keep in mind that wecan't assign geo-targeting.For that.So we'll see [INAUDIBLE] domain.And we won't actually see that
the /us or the US subdomainwould be geo-targeted
for the US.So that'd probably
be a bad combination.But moving it to country code
top-level domains is fine.Using subdomains is fine.Using folders on a generic
top-level domain is also fine.Essentially, any of those work,
and the one that you chooseis probably more
defined by what youwant to do with your website,
how you want to present it,than something that I'd say
you need to do for web search.
ROBB YOUNG: John, does
hreflang also work on those?So you can have .com in
the UK and hreflang there.So that's UK and .com.They say US and hreflang
there, and that youshould get the best
of both worlds?
JOHN MUELLER: Sure.You can definitely do
that The hreflang markupyou'd have to do on a
per page basis, though.So it's not that you
can do the home pageand it'll apply
for the whole site.You really have to do
that on a per page basis.
ERIC: There's an
international targeting tool.I think [? you'll have to ?]
watch it six or seven monthsago in Webmaster
Tools. [INAUDIBLE].Especially if you're going
to do the hreflang markup,it'll tell you if you've
made any mistakes.It's pretty easy to make
mistakes on that sometimes,so I'd definitely use that
international targeting tool.
BARUCH LABUNSKI: Can
I ask a question?
JOHN MUELLER: Just a second.
MARY: Oh, yeah, let me
just finish this up.If you do switch your
domain, for example,just make sure we
understand whatthe change of address
to the 301 is.Any other question
about ranking there,we pick it up pretty fast.There might be some fluctuations
as you're moving your site,but we do understand
it really fastif you set it up correctly.Yeah.So did someone have a question?
JOHN MUELLER: Baruch.
BARUCH LABUNSKI: Yeah.Hi.Can you hear me?
JOHN MUELLER: Yes.
BARUCH LABUNSKI: OK,
so I have a questionabout how do you discount
links and when do you do that?When does the
algorithm decide, hey,you know what, we're going to
discount this basically outof the portfolio of
the site's links?In what case, for instance,
would you do that?When?
JOHN MUELLER: So
you mean the linksto the site or the links--
BARUCH LABUNSKI: Yeah.So if a site is
behaving really bad,when do you decide to go
ahead and discount the linksor whatever you call it?It's just it hasn't
been answered,and I just wanted to
know when does it happen?
ERIC: We pick it up as-- if we
can identify it as a bad link,we definitely try to take action
on it, whether it be manualaction or algorithmically.When does it happen?Probably as soon as we can
find it and identify it as bad,I guess.
BARUCH LABUNSKI: OK.If sites are really behaving
as they're not supposed to,I just wanted to know.Maybe it could take a month,
two months, I don't know.That's why I'm asking.
JOHN MUELLER: So I
guess, like Eric said,when we pick it up, essentially,
that's when that happens.And if there's something that
we can pick up algorithmically,then we might be
able to pick that upas we crawl those pages,
as we find those links.If it's something that is
based on a manual action, then,of course, when someone
submits a spam report,when we run across
that spam issueinternally for other reasons and
someone from the web spam teamlooks at that, then they can
take manual action on that.And then that'll be usually,
I guess maybe a coupleof days or something like that.
But the spam reportsare read on a daily basis?I mean, I know you
have a large team,but it's not as large for so
many websites around the world.
ERIC: So, I mean, we try to
look at as many spam reports.I think we do look at most spam
reports when they come through,and I can't say that's daily.It's really based off of, like
you were saying, resources.But there's also a lot of other
ways that we detect badness,so it's not necessarily just
dependent on spam reports.So there's a variety
of ways we can pick upsomeone that's
trying to manipulateour system through linking.
BARUCH LABUNSKI: Right.OK.
JOSH BACHYNSKI: When are we
going to be done with links?I mean, for example, I just did
a really successful campaignfor a client, some viral
marketing, and good Googleshares, a few hundred tweets,
maybe 100 Facebook Likes.And they didn't get-- they
got three links out of it.And the links that
they got were justaggregator sites,
social aggregator sites,like tweetthis.com or whatever.So they didn't get
any links out of doingwhat I would consider a fairly
successful viral marketingcampaign.When are we going to
be done with links?It's getting pretty
difficult to find links,and I think it's well
time, in my opinion,that Google is
tracking somethingelse other than links,
in my personal opinion.
JOHN MUELLER: We have
a lot of experiencewith [INAUDIBLE] crawling,
indexing, and ranking.So it's not like you
need to have links.And for example,
one of our friendsback home recently set up a new
website for the neighborhood,and they don't have
any links at all.And they have, I think,
over 300 pages are indexed.They're getting nice rankings.They're getting lots of
traffic through Search,enough for them, at least.And they don't have
any links at all.So nobody has linked
to that site ever.They're submitting site maps.They have an RSS feed,
those kind of things.But this website
does fairly wellwithout getting links at all.So it's not the case that you
absolutely need to have links.Obviously, it's still a
part of our ranking factors,but it's not something where I'd
say you don't get any links outof something that you
do that you failed.Like these viral campaigns
that you mentioned,that might be a fantastic way
to drive traffic to a website.That might be something where
we pick up other signalsand use those as well.So from that point
of view, I wouldn'tsay that it's like [INAUDIBLE]
to absolutely get the likes,not something that Google
always completely relies on.
BARUCH LABUNSKI: But a
different search enginesomewhere around the
world has tried that,and they went ahead
and removed it.And I personally don't think
the results are that great,based upon they tried it, right?You still need links, right?
JOHN MUELLER: I think links are
really important for crawlingthe web, because
that's a great wayto discover content that
people are pointing at,to discover related pages
that are linked around.That's a great way
to go from one pageto lots of other pages.So from that point
of view, I thinklinks is definitely an
important part of the web.I believe Yandex removed links
from their ranking factorsin some specific
parts of their search,so it's not the case that
it's completely removed.But I think it's important
to try these kind of things.And I imagine these
are experimentsthat we run from time to time,
as well, to see what would itbe like if we didn't
look at page rank at all?How would the search
results look like?What other factors
could we look at?And I think that's something
that any search engine hasto keep asking itself and
keep challenging itself,and saying, hey, just because
this is something we've alwaysdone doesn't mean it'll
always be like this.Maybe at some point, people
won't have any links at all.Maybe the web won't
be in HTML anymore,but it'll all be in Flash or
in apps or something like that.And these things happen.And it's not the case that
we can make a decisionand say, oh, the web should
move to apps or to Flashor whatever.The web does what
it wants, and wehave to make sure that we use
what's happening on the weband provide useful
search results to people,so that people still find
our services useful, as well.
BARUCH LABUNSKI: But it's not
going to happen in a long time,right, like in SMX.Matt Cutts was saying it
won't happen for a long time.
JOHN MUELLER: I have no idea.A long time is a
very flexible time.If you're looking at
the web, somethingthat happens within a year
is like a really long time.But if you look at the
bigger picture of computers,then maybe a year is
a really short time.But I'm not saying that in a
year, we'll stop using links,but I know that we're
constantly tryingto challenge ourselves
and figure outwhat's best for the user, how
best to pick this content up.And you're going to
see changes over time.So if you're an SEO or
webmaster and you're saying,well, I have it all figured out.My website is ranking
number one now.I won't touch a thing, and
it'll stay ranking number one.Then probably that's
not going to happen.Things always change on the web.
ARTHUR RADULESCU: So John,
then what other factorsyou pick up if you
don't use social mediaas a ranking factor, as
in your previous comment?
JOHN MUELLER: I don't know.[LAUGHTER]I don't have anything offhand
that I'd be comfortablesharing.There are lots of
things-- when youlook at the content
on the web, howit interacts with each other.I don't have anything
specific that I'dpoint at where I'd say,
this is a ranking factor,because then you'd all run off
and try to gain that figure.And then we wouldn't be able
to use it as a ranking factoranymore, because
essentially, peopleare just playing games with it.
ARTHUR RADULESCU: No, no.I didn't want you to
specifically say something.It's just for the reason
you just mentioned.But generally speaking, that all
the traffic from your exampleearlier-- I mean,
from Josh's example--came from social media.So everything came from social
media apart from the links.And you're saying that
you're using some factors,and you're looking
at other factors.Then it's not links.Then what it is?I mean, just a general
idea, if you can.Of course.
JOHN MUELLER: I'd say there are
lots of indirect aspects there,too, where maybe you don't get
explicit links from this kindof activity, but maybe you get
indirect links, where peopleare saying, well, I'll
Bookmark this for later.And then they take that
page and they recommend itto their friends on Facebook.And it's not like
a link, where you'dsay, well, this
passes page rank.Therefore, it'll be
used for ranking.But your friends will look at
the recommendations in Facebookor wherever they're active,
and they will follow that,go to that website.And maybe one of these
people indirectlywill put a normal
link on their blog.Maybe a newspaper
will write about this.And these are all effects
that maybe indirectly wemight be able to pick up on.Obviously, we can't
track everything,and we can't use all kinds
of social media signalsfor ranking.But there are lots
of indirect effectsout there that we could try to
pick up on and use for ranking.
ARTHUR RADULESCU: Thank you.
ROBB YOUNG: If people are
coming from social mediaand they're getting
traffic from social media--visitors, sales--
then it seems--[INAUDIBLE] wood for the trees.Then why are you
looking to Googleto help you build
links out of that?Why not just get sales,
traffic, and customersfrom social media?Why not treat it
as it is, which isanother source of
traffic and sales?
Because Google is stillthe biggest game in town
and we have no choice.
JOHN MUELLER: No, I disagree.I think there are
lots of chancesto get other traffic
to your website,and that's not
something where peoplehave to put everything
into a search.I've seen sites that
are really popular thatare roboted out completely.They're services that don't
even have a real website, thatessentially just have an
app that maybe we don't evenshow in Search at all.And they might get
a lot of traffic.They might get a
lot of customers.It's not the case
that everythinghas to lead to search.
ROBB YOUNG: It totally depends
on your business model.There are companies like Zynga
that have just built gamesthat people don't
search for those.They just find them on
Facebook or they share it.I'm not saying my
business could rely on it.It absolutely couldn't.But there are
plenty of things nowthat appear on social media
which admittedly are generallyjust clickbait.But they make their money from
finding you on social media,getting you onto their site,
and getting advertising revenue.It's no good if you're
e-commerce necessarilyor if you don't
have a good idea.But I'm not saying if
you run an SEO companyor you sell marketing
services, that'sgoing to be perfect
for social media.But other things certainly are.So it's industry-based.
Oh, I don't know.I don't know if I
agree with you guys.
JOHN MUELLER: That's fine.That's fine.We'll still send traffic
from Search to your site.
BARUCH LABUNSKI: No,
it's just that peopleare very used to Google, period.They do whatever they need to
do, and they do it on Google.It's just that's how it is.
JOHN MUELLER: I'd say the
web evolves very quickly.So people are used
to Google now.That doesn't mean that
we can just lay backand say, well, they're always
going to come to our site.Just like if you had a
website that you don't changefor a couple of
years, then peoplearen't always going to come
to your site just by default.Things have to evolve over time.I guess a lot of
the practices werecommend also apply to
any other kind of usersthat you drive to your site.If you have a really
fantastic website that'suseful for your
users, then that willwork regardless of the kind of
traffic that you're getting.It's not the case
that you have to builda specific website
for social mediaand a different one for Search.You can make a fantastic website
that just works for your users,regardless of how they come.
ERIC: I'm going to sound
like a broken record,and I'm sure everybody's
heard about this before.But content, quality,
and relevanceare what's important to us.As a search engine,
we want our usersto have good content
and relevant results.So when we're talking
about things like links,they are a signal of good
content and relevance.So don't think of
links as an end all,be all, like I have to get
links, I have to get links.Or don't think of
social media, Ihave to get a whole bunch
of clicks and everything.Just think about what your
users are looking for.And if you're thinking about,
from the perspective of Google,what people are Googling
and trying to searchfor when they're looking
for your website.And so going back to signals,
I just think about it that way.Don't think about
specific signalsthat you can do
or change somehow.Think about, how can
I make my contentmore relevant and
the quality better?So I know, broken record.I'm sorry.
JOSH BACHYNSKI: No,
no, no, that's good.
BARUCH LABUNSKI: Well--
JOSH BACHYNSKI: Let's
ask it this way.Hold on, Baruch.Let's ask it this way then.
BARUCH LABUNSKI: I just--
JOSH BACHYNSKI: If
you have to choose,what is the better signal
for Google-- relevance,directly tracking relevance, or
indirectly tracking relevancethrough [? quality ?] signals?Would Google rather
read relevance directlyand quality directly,
or would they rathertell there is quality and
relevance through indirect usersignals, whatever they may be?
ERIC: Why not have both?
JOSH BACHYNSKI: Both, of
course, is the answer.How good is Google tracking the
former opposed to the latterthese days?Because it used to be-- the
reason why you guys startedtracking links in
the first placeis because that was
a really great wayof telling what was
relevant online,because everyone
was linking to it.Now with the technology you
might have at your disposal,how comfortable is Google
directly looking at relevanceand directly looking at
quality signals, as opposedto having to rely on
JOHN MUELLER: I don't think we
have the magic answer for that.
JOSH BACHYNSKI: I'm asking
a philosophical question.I'm asking for a
philosophical answer.I'm not looking for
JOHN MUELLER: I
think in general, wetry to find things that give
us this kind of information,and we try not to rely
on just one signal,because sometimes that can
be a little bit misleading.And that kind of applies
across the board to everythingthat we're doing here.So if you're saying either
this or that, like Eric said,why not have both?I think that's kind
of what we'd aim for,to confirm that the signal is
the correct thing, that we'repicking the right things up.And there's a lot
involved there.And these things
change over time.It's not really that trivial.
BARUCH LABUNSKI: We just want
to basically-- not for Googleto be more transparent.We just want to take
care of our clients,not to make them number one.It's just to make the web
clean at the end of the day.Like, our clients don't have the
time to spend on their website.I mean, there's so
many things to do.So that's why I
asked that question.I just wanted to
know, because there'sa lot-- when it comes
down to competitionareas, either contractors
or something else,whatever it is, a
lot of businessesare trying to manipulate, right?So that's why I
asked that question.
JOHN MUELLER: Yeah.That's always tricky,
to find the right one.
JOSHUA BERG: John, I have a
different question on the topicof quality, content quality.We know from the--
I mean, we've heardabout the above-the-fold
layout algorithmsthat ads above the fold are
not a good quality signal.Is there a limit that would be--
or even a range that we couldsay that this is
definitely not safe,because I'm thinking
of a site thathas been quite successful
as a forum-type site thatgets millions of visitors.But they do have
1/3 of their spaceabove the fold as
advertisements,like sometimes running Google
ads and stuff like that.Do you think that's
too much, or do youthink that because they're
getting enough other qualitysignals that that
would be overlooked?
JOHN MUELLER: I really kind
of take a step back thereand look at it from a
user's point of view,and not just count pixels
and say, 1/3 of this pageis covered by this, 1/3 is this.But really try to look
at how users would feelwhen they land on this page
from the search results.Do they feel that this
answers their query?Do they feel that this
is relevant to them?Is this high-quality
content that theyfind when they open that page?Or do they just get
lost in a sea of ads?And that's something where
it's really not [INAUDIBLE],kind of like with the number
of words on a page, to say,well, this specific number of
ads is good and anything abovethat is bad.But you really need to take
a step back and look at itfrom a user's point of view.
JOSHUA BERG: OK.So the user's reaction
to that action-reactionmay be more important.
JOHN MUELLER: I think that's
something you could look at.I mean, if this is
your website, youcould look at things
like in Analytics,how people are going
through your website.You have that information.That's something you can
interpret really well,because you know your website.You know what you
expect users to do.And that doesn't
mean that we're goingto look at your Analytics,
because we're notgoing to use that
for ranking anyway.But that's something
that you, as a webmaster,can use as a tool to figure
out, am I doing the right thing?Other things you can
do there are just,like, traditional user studies.Take a bunch of
people that you thinkmight be interested
in your website,invite them to your office,
and have them complete taskson your website.And ask them questions about how
they respond to your website.And I think that's
always really useful.That's not like
one-to-one SEO technique,but it really helps
you figure outwhich parts of your site people
are responding positively to,which parts of your site
they basically get lost in.
BARUCH LABUNSKI: John, to
Joshua's question, basically,when quality raters
come to the site,just like they do
on Google Maps,is there a way that
they can add commentsin an auto-generated
comment in Webmaster Tools,once they were there?I mean, they were there--
JOHN MUELLER: It's not the
case that we go to your websiteand we're like,
oh, this is 70% OK.Therefore, we're going
to drop it in rankings.These kind of raters, when
they look at the search resultsand they analyze the changes
in the search results,they give us general
information about the algorithmsthat we use for them.They're not going to have
feedback specificallyfor you, as a webmaster.So I doubt that they'd have
time to submit a feedbackform for your website.But this is something
where I wouldn'tfocus on that part of Google,
because that's not somethingwhere people are going to
your website and saying,oh, well, we have to review
all websites on the internetso that we can make sure that
the search results are OK,because that obviously
would never be possible.So really focus on your users.You know your users best.You know your website best.You know what people
should be doing there.And that's something where,
if you make a fantasticwebsite that works really
well for your users,then it would be an
error on our sideif we don't reflect that in
Search in the relevant areas.
JOSHUA BERG: OK.Yeah, thanks.And so one other question about
forums that are well-used.If a site occasionally gets
some adult content in postsbut maybe the site
as a whole doesn'twant to be tagged or labeled
as adult content, then a forumlike that, should they need to
be very cautious about allowinganything like that, because the
whole site or larger portionsof it may get affected by that?I mean, the pages
can't be individually--or would you suggest,
if some contentneeds to be individually labeled
as such, that it could be?
JOHN MUELLER: I
guess first of all,it's important to keep in
mind that even if you haveuser-generated content, it's
essentially your websiteand you're publishing it.And if people are
submitting somethingthat you don't want
to have published,then you need to find a way to
catch that at the right time.So that's essentially
not something,from our point of view,
where we would say,well, some random
person on the internetsubmitted all this
spam to this website,and therefore we shouldn't
count it against that website,because some random person on
the internet actually did that.But rather, you, as the
webmaster, as the site owner,are essentially
responsible for the contentthat you're publishing there.And if we look at
your site overalland we see there's
a lot of junk here--maybe there's a lot of
adult content on this site.Then that's something
our algorithms might say,well, overall, this site here
is kind of in this category.And that's regardless of
who generated that contentand put it on your site.It's essentially you're
the person republishing it.So if you don't want
that content associatedwith your website,
then make surethat it doesn't get
published on your website.
JOSHUA BERG: OK.But I was thinking about
a site that maybe itis pretty well-moderated, but
there are questions or topicsthat may be related to
medical or medicine thathave to do with adult content
or something like that.Then, is that something they
should still keep a handle on?
JOHN MUELLER: That's
kind of up to you.I mean, as a webmaster, you
decide how far you want to go,what kind of content you
want to have published.In Google, we try to be pretty
granular with our algorithms,but if we can't easily
recognize which parts of a siteare kind of adult content,
which parts of a siteare not adult content, then it's
possible that at some point,our algorithms will
say, well, thereis some adult content on here.Therefore, we have to be
cautious with this website,showing it in Safe Search.So that's kind of
a decision that youhave to make as a
webmaster and say,well, I'm OK with my
website being like this.I'll accept what people
are publishing there.Or maybe you'll pick a
separate part of your websitethat has the adult content.And you'll say,
well, everyone whohas questions on
these topics shouldgo to this part of the site,
something that's eitherblocked from
crawling or indexingor just clearly
separated from the rest.
JOSHUA BERG: OK.Yeah.And I was thinking it as a
relatively small portion--
JOHN MUELLER: Sure.
JOSHUA BERG: --of the site.
JOHN MUELLER: That's
essentially up to you.Let's grab some of the
questions that were left here.Do you see any that
you want to jump in?
MARY: Yeah.I like the question about
the large franchise.
JOHN MUELLER: Which one?This one?
MARY: Yeah.So what are best practices?So ideally, when you're
making localized websites,you probably have
a product or youhave a service in
that region, right?So the hours, a map
of the location,phone number, just
anything that'srelevant to the local place.And if they are in
different languages,you use the hreflang
option, too.And to make sure that
it's not duplicate,like if you're making it for
local regions, local languages,then you have different hours.You have different locations.You have different
phone numbers.People are going
to understand that.And I see a lot of good examples
of that, a lot of cars sharingstart-ups in the Bay Area that's
expanding to other countries.They're making a lot
of localized websitesbut on the same domain.But they have
different information.They have different prices, too.And if you just look at
some good examples online,then that should also give you
some good ideas on what to put.Any other suggestions?
JOHN MUELLER: Yeah.Sounds good.Do you have one, Eric?
ERIC: We can--
JOHN MUELLER: Let's
see what we have.
MALE SPEAKER 2: John, can
I just talk real quickon Josh's earlier statement?
JOHN MUELLER: We'll try.
MALE SPEAKER 2: [LAUGHS]
All right, thanks.It's more of a general
concept of links.I was unable to join last
week's last Webmaster Hangout.However, I watched
the whole thing,and someone-- they weren't
part of the actual Hangout.They were asking the
questions on the side,and they asked about getting
editorial links for their site.And you seemed to
immediately jump onthe, well, that sounds
like you're doing somethingthat you shouldn't
be doing, whenisn't editorial links more going
out and basically getting linksfrom the value of our content?I mean, isn't that the best
way to go about doing it?
JOHN MUELLER: I guess it depends
a lot on what specificallyyou're doing there, because what
I see a lot is people sayingthat they're going to
build editorial links.But essentially,
what they do is justspam webmasters
that happen to havewebsites with similar topics.And hey, you linked to me.Please, this is
really cool stuff.You should link to me.And that very quickly turns into
kind of almost like unnaturallinks, that they
wouldn't be thereif they're not doing some kind
of link exchange or somethinglike that.So that's something where
I'd be really cautious.And a lot of times when we see
people talk about their linksand say, well, I know
I have some bad links,but I'm going to build a
whole bunch of good links.And essentially, they just
go out and spam a lot more.Then you're just
digging a deeper holerather than actually
fixing your problems.And if your content isn't
attracting those linksby itself, then
that's something thatmight be something worth
thinking about a little bitmore to figure out why this is
happening, why people are goingto your site but
actually they don'twant to recommend it to anyone.
MALE SPEAKER 2: So
it's not OK emailingother relevant bloggers who
are talking about your brandor brands that you're
involved in, asking for links?
JOHN MUELLER: I'm not going
to make any general statementand say this is always
good or it's always bad.I can see it
definitely makes senseto sometimes draw attention
to something really fantasticthat you're doing, if people
weren't aware of that.But if you're essentially
just emailing random peopleto ask for links, then
that sounds pretty lame.I don't know.
Basically, if we'regoing through an
election-- the States,the US is going
through an election.And if there's a fake vote,
it's not counted, right?So it's all about getting
real votes, right, John?
JOHN MUELLER: I don't
know how the elections arehandled in the US.I have no idea.
BARUCH LABUNSKI: It's
all about real votes.
JOHN MUELLER: Real votes?I don't know.I don't know how that works.
JOHN MUELLER: I'm sure that
there are good parts there.Let me see.Here's one question
about mobile-friendlythat I thought we'd get
through very briefly."You said in the future,
mobile-friendly factorscan affect ranking
of mobile Searchbut maybe not of desktop.Currently, as Google is
using many desktop signalsfrom mobile Search,
as well, is Googlegoing to index desktop
mobile sites separately?"So to some extent, we are
picking them up separatelyalready.With rankings, that's
something that'sbeen the case since a while now.So I think mid-last year
or almost two years agonow, we did a blog post
about different factorsthat could affect your
mobile search rankings.And that's something that's
been a case for a while.So if you do some
of those thingsthat we mentioned on
the blog post there,then chances are that users
searching on smartphoneswill see your site on
a different rankingthan they would on desktop.And that's something that I
imagine will be moved forward,as well.In the recent mobile-friendly
label blog post that we did,we also said that we're going
to experiment on ranking changesthere, as well.So if we see that an equivalent
site is in a search resultequivalent, be
relevant, for example,and we notice that the
user's on a smartphoneand this other site works a
lot better on smartphones,then maybe we'll show
that one higher in Search.So I definitely
expect more changesin that direction going forward.
BARUCH LABUNSKI: Is
mobile and Search--is the mobile in
Search different rightnow, because I'm seeing
that it's not the same.Like for instance, a site
will be number one in desktopand then number three in mobile.I mean, before, it was
all the time, it was--
JOHN MUELLER: It's hard to say.I mean, the search
results in generalhave been different since
almost the beginning, wherethe composition of the
search results will change.Sometimes it makes more sense
to show big images in desktop,but then it won't in
mobile, for example.The ranking generally
would be fairly similar.But even on desktop,
you'll see changes from dayto day, where you'll
check on one computer,it'll be ranked like this.You check on a
different computer,it'll be ranked different.And it's not that
we're specificallytargeting that
individual computer,but we have experiments that
are running all the time.We take personalisation into
account as much as possible,and all of that can also
play a role on mobile.
ERIC: And if you allow
a location on mobileor something, there's a
lot of different thingsthat we can look at on mobile.
BARUCH LABUNSKI: OK.
JOHN MUELLER: All right."We think we might have been hit
by a duplicate content penalty.Once we've resolved
the issue, how longwill it take for us to
regain our rankings?"We have a duplicate content
penalty, like a manual action?
ERIC: We do have a manual
action that's specificallytargeting thin content.So if you file a
reconsideration request,after the
reconsideration request,it may take some time again with
the indexing and re-crawling.And in general, if you
have those types of issuesand you've identified
it, same thing.It's re-crawling, re-indexing.It's going to take us some
time to pick up these changes.
MARY: And keep in mind
that your ranking beforemight be inflated
because of any behaviorsthat you did that
inflated your ranking.So it's going to
allow us to understandwhere you fall
naturally, and thatmight be lower or
higher than before.So work on the content and
get rid of that manual action.
JOHN MUELLER: All right.
ROBB YOUNG: Will we have time
today to answer my question,or do you want to-- is
your email enough to meor do I leave it
until next week?
JOHN MUELLER: I think that--
JOSH BACHYNSKI: Oh,
the standard question.
ROBB YOUNG: I think
it'd be a good learningcurve for the other
people around,since this seems to be
such a complex issue.
JOHN MUELLER: Yeah.I'd probably just leave it to
the Google+ post that you send.
ROBB YOUNG: Because I
just want to understandhow far to distance
ourselves, whether you'retalking about removing the
domain from our ownership,removing any links
whatsoever, removingany hosting
association, removingany hreflang association.I mean, totally
distancing ourselves,or whether we're just
talking about ditchingthe links between this?We're on the hreflang
at the moment.We don't 301 or anything.And you told us--
JOHN MUELLER: If you're not
redirecting from the old domainto the new one,
then that's probablywhat we're looking for.
ROBB YOUNG: Because we do
hreflang to tell peoplethat-- but that was a
workaround, in all honesty.It wasn't a-- they're
both in the same country,but we were just
trying our luck.But then, you know that.We're not--
JOHN MUELLER: I don't
think the hreflangthere would make a
lot of sense, but itshouldn't cause any problems.
ROBB YOUNG: No,
but it did give usa short-term boost over a
period of six to eight weeksor something, which
then seemed to-- youthen caught up with
us, effectively,and thought, actually,
this is the same site.Go away again.
JOHN MUELLER: I don't
think that you'dsee any problems of [INAUDIBLE].But essentially, the
redirects, like in your case,where you have one domain
that has a lot of problemsand you set up a
new site, then theredirects kind of forward all
those problems to the new site.So if you [INAUDIBLE]
place, then that'sreally helpful for
us to understand thatstarting with this new domain.But at the same
time, it also meansthat you have to build
up from that domain.It's not something
where you can say, well,I'll take the previous
standing and build from there.But rather, you're starting
over with that new site.
ROBB YOUNG: Right.See, previously, you
said 301 would be OK,because it's not a link
penalty and there'sno penalty associated with that.So feel free to 301.But now you're saying, actually,
the domain is so toxic,don't 301.
JOHN MUELLER: I think
in your specific case,I'd avoid redirecting now.Yeah.
ROBB YOUNG: All right.So we're basically starting a
new 10-year-old business withno--
JOHN MUELLER: I don't know.You still have a lot of
things that essentially peopleknow about.People are going to your sites.I think from that
point of view, you'renot starting exactly at zero.But it's definitely
not an easy situation.
ROBB YOUNG: I know.And we have to keep
that live, that site.We have no choice, because we're
a gift certificate company.So people have that domain to
visit within printed literatureand have for the last 10 years.And in the US, gift
certificates don't expire,so they have to visit it.So we can't not have it there.But how far do we
distance ourselves from itin order for it to no
longer ever be a problem?
JOHN MUELLER: In general, if
you're setting up a new site,if you're not redirecting,
then that's a good sign for usthat these sites are
ROBB YOUNG: So the same
owners, same hosting,and same everything
else should be fine, aslong as there's no links?
JOHN MUELLER: Yeah.
ROBB YOUNG: OK.
JOHN MUELLER: All right.OK.So with that, I think
we're out of time.Thank you all for joining.Next week, we'll be-- or
maybe two weeks, I guess,we'll be back at the
usual times, whichdon't work so well
for California.Sorry.But maybe we'll
do some later onesover the course of
the year, as well.So thank you all for joining.Thanks for all your
questions and comments.And hope to see you guys in
one of the future Hangouts.
ARTHUR RADULESCU: Hey hey.
MALE SPEAKER 2: Thank you.
JOSH BACHYNSKI: Thanks, John.
ARTHUR RADULESCU: Thank
you guys for coming.
JOSH BACHYNSKI: As always,
we really appreciate it.