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: All right.Welcome everyone to today's
Google Webmaster Central OfficeHours Hangouts.This will be, I guess, the
last one for this year.But I'll set up new
ones for next yearas well so we can
keep rolling on these.Looks like we have
a bunch of peoplehere signed in already,
hanging out with us.A bunch of questions
were submitted.We can probably make
it through them.We have a bit more time today.So maybe we'll have time for
discussions and questionsfrom you all as well.As always, does anyone want
to get started with a questionor comment or something?
BARUCH LABUNSKI: Yeah,
I wanted to starta question about HTTPS.So I changed a website
with 45,000 pages.Of course, it was
me and someone else.And basically, everything
was changed to HTTPS.The exact certificate
that you wantedand everything moved correctly
in Webmaster Tools from HTTPto HTTPS.And then, of course, you
watch the traffic go downto zero in the HTTP account.So that was moved to the HTTPS.But then, like always, traffic
started going a bit down.How normal?How many positions usually?Because it's not the
first website I moved.And normally, there
wasn't any fluctuations.But how normal is it
for the fluctuations,because it looks like it's
significantly going down?But then it went
up again normally.And I moved the analytics
as well property.So it's everything-- I'm
watching things closely.It's a lot of [INAUDIBLE]
JOHN MUELLER: Yeah,
it's really hard to say.In general, you wouldn't see
much of a big change overall.You might see some
small fluctuationsas things kind of shuffle over.But it should settle down
to the previous state.And I think sometimes
what happensis when people do the move
during an off-season time,then they might see those
effects as well in the traffic.So that's something
worth looking at, maybelooking at the traffic
patterns you hadlast year throughout this time.Did it change during that time?Did it go up or down?And try to see if
that matches whatyou're seeing with HTTPS there.
BARUCH LABUNSKI: No, last
year was the same, of course,like always December.Yeah, it's a bit slow.
JOHN MUELLER: Yeah, but it
probably depends on the site.So that's something maybe we
can take a look afterwardsif we run through the questions
and have a bit of time.But in general, these things
just take a bit of timeto settle down and be
reprocessed and re-indexedin the new way.
OK, sounds good.And then also
regarding that email,did you see that one I sent you?
JOHN MUELLER: Probably.
BARUCH LABUNSKI: The latest one?OK.
JOHN MUELLER: I'd
have to double check.
BARUCH LABUNSKI: OK, thank you.
ROBB YOUNG: John, for the
secure move, it's something thatdid-- one of the things we did.But then I'm wound again.Is the current advice still
you're moving the whole thingor you just give Google
the option of bothand let Google choose to
prioritize secure over not?Because when we did the whole
site move and canonical,we had to unwind it all
because it just died.But this was in the early
days of the first two or threemonths of when other people
were reporting the same.But what's current advice?Move or give Google the option?
JOHN MUELLER: With
changes like this,what I would do is just make
it available in both versions,first of all.So in your position,
it would be give Googlethe option in the beginning
so that you can be surethat the HTTPS site is working
as it should be-- that youdon't have these mixed content
issues-- those kind of problemsthere.And once you're sure that
the site is working the wayit should be, that the
performance is the waythat you expect
it to be, then I'dset up a 301 redirect
from the HTTP versionto the HTTPS version.So do it in a two-step process.Obviously, if you're sure that
the HTTPS version is workingreally well because you've
tested it properly internally,then just setting it up and
doing a site move directlywould be just as fine.
BARUCH LABUNSKI: But
what about the defaults?Does that keep in the defaults?Because you said you were
ranking things by defaults.
JOHN MUELLER: How do
you mean by default?
ROBB YOUNG: My default domain?
BARUCH LABUNSKI: Yeah, you're
starting to rank HTTPS.
JOHN MUELLER: Oh, OK, with
the recent blog posts.So, essentially, if we
discover a URL on HTTPand we discover the same URL
on HTTPS, then what will happenis we will try to prioritize
the HTTPS version over the HTTPversion.So that's kind of what we
mean by default in the sensethat if we have both of
them, and they're essentiallyequivalent for us
because we don't havemuch more information
about them-- you don't havea redirect set up-- you have
around canonical set up-- thenwe'll try to pick
the HTTPS version,provided that all
the criteria are met.
ROBB YOUNG: So you
don't really need a 301.The risk is that
in the past you'veconfirmed that a
301 doesn't give you100% percent of the authority
of the previous domain.
JOHN MUELLER: I wouldn't
worry about that with HTTPS.
ROBB YOUNG: Well,
but we have to.
JOHN MUELLER: But you're not
really doing a site move.You're staying on the same URLs.You're just switching
the protocol.So we know it's the same domain.It's essentially just
switching the URLs over.So that's the kind
of situation whereI wouldn't see those
URLs as losing any pagerank from a redirect.
ROBB YOUNG: But
there's no real riskin just letting both
URLs out there--seeing what happens over time.And if it shows that your
always shown the HTTPS,then just leave it.
JOHN MUELLER: Sure.Risk-- it's something
where essentially wehope that you guys move to
HTTPS because of your users,because of the
connection to the userthat the connection is secure,
that they're aware of goingto the right site.So that's something where if
you leave both of the versionsup there, then users
might not move.Whereas if you set up a clean
redirect, then of course,you're sending all your
users to the HTTPS version.And they stay on that
because maybe your site islinking to those pages anyway.So that's something
where I wouldn'tsay a good long-term
solution is to justleave both versions open.I try to pick the
version that's bestand make sure that everyone
is routed to that version.
ROBB YOUNG: And is that
dependent on audience?Our audience is 90%-plus
new anyway because people--normal e-commerce scenarios--
there's lots of new customersout there.So we don't have a forum like
some of these other guys.So they're coming back
every week anyway.So eventually over time, most
of them will end up on the HTTPSanyway, I would imagine.
JOHN MUELLER: Probably, yeah.At that point, you can set up
the redirect anyway, right?
ROBB YOUNG: Assuming the
traffic is right, yeah.You know our sites are
risky-- any kind of change.
JOHN MUELLER: I think
the average site--if you're just moving to
HTTPS, then that's reallynot something I'd worry about.That's the kind of situation
where I'd just say, well,you want to move to
HTTPS because youhave a bunch of good
reasons to do that.And as a part of
the move to HTTPS,you set up those redirects.That's something that we're
doing as well on Google.We're essentially going through
the whole site, everythingwe have on google.com and
making sure it works on HTTPS.And when we're sure it works on
HTTPS, we set up the redirects.Obviously, it's not
something that'sdone overnight that we
would switch redirect on.And if anything breaks, we'll
try to fix it as we go along.
ROBB YOUNG: Does Google
rely on Google for allof its organic traffic though?
JOHN MUELLER: Well, for
some of it sure, yeah.We have to show up
in search as well.And we regularly have the
same kinds of SEO problemsthat all of you have
where someone will go outand robot the homepage
or robot the login page,and everything go
through the login page,and you can't see
the content anymore.So we can't for it.And all of these things
are normal issuesthat every site
has to work with.
BARUCH LABUNSKI: Well, John,
regarding the disavow in HTTPand HTTPS, does it
matter if you firstadded disavow to the HTTPS and
then add it to the non-HTTP?
JOHN MUELLER: I wouldn't worry
too much about the order there.That's something where I'd
essentially just uploadthe same file to both of them.
BARUCH LABUNSKI: Yeah, so
now you get two messages.
JOHN MUELLER: Yeah, both of
them in the same account, yeah.Over time, you'll have
one version of the sitethat you pick as your
canonical anyway.So that's the one you'd
be working on over time.But in the time frame where
you have both of the versionsup where some of
your links are stillpointing at the old version,
then you will have both of themin your account.And you work on both of them.But in the long run, you'd
have one version of your siteessentially in Search
Console, and that's the onethat you'd be showing.
So you're sayingfor maybe two or
three months, justdo it like that or a full year?
JOHN MUELLER: Sure.I don't know.Maybe half a year
or something likethat if you're doing
a site move type thingor if you're switching
from HTTP to beat HTTPS.Let me run through some of the
questions that were submitted.And I'll try to keep
it short so that wehave more time for all of the
other live questions as well.Or maybe you could just go ahead
and ask your question first.And then I'll run
through the questions.
MALE SPEAKER 1: Regarding
this HTTPS, actually,this was the question.In case somebody has disavow in
HTTP version for his website.But suppose he does
not disavow for HTTPS.Will Google take any action?Or this is really risky
for a future purposeif somebody does not
disallow an HTTPS versionor does not have such
JOHN MUELLER: I
would just make surethat disavow is live on the
current canonical versionof your site.So if your site is
currently on HTTPand you don't have
anything on HTTPS,and everything is
indexed with HTTP,then that's where you
do the disavow file.If your site has moved to
HTTPS or moved from wwwto non-www or to a different
domain, then whereveryour current
canonical is, that'sthe one where you need
to do the disavow file.And you could just
take the disavow filefrom the old version
and just re-upload itto the new version.So you don't need to do anything
special with that file then.
MALE SPEAKER 1: OK, but John,
my question was actuallyis Google really very
concerned that youhave to disavow on HTTPS?Does Google have any
plan to take actionagainst those websites
who do not do this?
JOHN MUELLER: Well, if your
site isn't indexed as HTTPS,then you don't need to do that.It's really just based on
where your site is currentlycanonical.So it's not something
where we tryto throw anything artificial
in the way and say,you need to always do
this on HTTP as well.It's just if that's the
version of your sitethat's canonical at the moment,
then that's the one wherewe need to have these files."When do you guys estimate the
next Penguin update will be?"I don't have any
new news on that.We talked about that
briefly before Barry joined.But apart from that, I don't
have anything new to share.No, just kidding.I don't have any news
on the Penguin up.Usually we take a pause
during the holidaysto make sure that we don't
do anything crazy then.But I imagine we'll hear more
about this in the New Year.
BARUCH LABUNSKI: Soon.Soon.
JOHN MUELLER: We'll see."I want to know about
spam referral traffic.How can we stop it?Our website's rank is going down
due to spam referral traffic."So there are two
kinds of spam trafficthat I've heard about recently.On the one hand, there
is the analytics spamwhere people aren't actually
going to your website.They are just sending
pings to analytics.And that's showing
up in analytics.And that's something that I know
the analytics team is workingon and taking pretty seriously.The other type of spam
traffic is when people justhit your site with a bot
with a fake referrer thatshows up in other types
of analytic software.So for both of
those, that's reallynot something that would
be affecting Google at all.It's kind of a hassle because
you have to filter outin your analytics.And you have to
make sure that thisis traffic isn't
counted if you'relooking at your page views.But apart from that,
it's not somethingthat would cause problems
on Google's side."When testing my site with
Google Page Speed Tools,it says I have to improve the
loading of external GoogleFonts as well as adding
caching for Google Analytics.What do you suggest when
Google complains about Google?"So in general, this
is something that Iwould treat as any other issue
that comes up on your site.We explicitly don't white
list Google's featureswith these tools, because
these features slow downyour website in
a browser anyway.And just because this
feature is created by Google,doesn't mean it doesn't slow
things down for your users.So that's the kind
of thing whereyou have to take the output
of these testing toolsand use your knowledge to figure
out which of these aspectsare really problematic, which of
these are really slowing thingsdown for users, and then
take action to improve thingsin the right way.And that might be
removing some featuresthat Google has implemented
because they do slow downyour site.So that's something
where you haveto take your knowledge
and your experience,apply it to the
output of these tools,and then make
changes in the waysthat you think makes
sense for your users."Will making changes
to a site templateimprove our decline site
traffic or site rankings?"Yes, it will probably change
your traffic and rankings.Obviously, if you
do things reallywell-- you make things
really nice-- thenthat might have
positive effects.If you break things by
making site template changes,then that might have
negative effects.So making changes to
your site templateis essentially making a change
on your website and somethingthat we would pick up on
with crawling and indexing.
BARUCH LABUNSKI: When does the
design core algorithm kick in?I think it's once a year, right?
JOHN MUELLER: Which algorithm?
BARUCH LABUNSKI: Well,
there's the design, whichis 1, I guess, from the 200.Once a year--
JOHN MUELLER: I don't
know if we do that.But as a part of, for
example, Panda where we dotry to recognize
high-quality sites.This is something that
plays into the same thingwhere if you have a website that
looks really, really obsolete,and you're trying
to say, well, thisis like bleeding edge
scientific information here,and you're presenting
it in a way thatlooks like a front-page website
from 1995, then that might notlook as trustworthy as something
that actually looks like youknow what you're talking about.So that's something
where I wouldn'tfocus on specific algorithms.I'd more focus on making sure
that your website matchesthe image that you
want to bring across.And sometimes that
does mean updatinga really absolutely
template to somethingthat looks a lot nicer that
works a lot better where yousee also people go
to your website,and they think, oh, well, I
can really trust this website.It looks professional.It looks like they know what
they're talking about comparedto something where,
oh, I don't knowif this is made by some kid
with some obsolete graphicsprogram that just copy and
pasted some content togetherwhere it's the kind of
thing where you might assumethat the content isn't
really that high quality.
BARUCH LABUNSKI: How often does
the core algorithm role out?
JOHN MUELLER: The
core algorithm--so we make changes to our
algorithms all the time.And these algorithms run
pretty much automatically.So that's not
something where I'dsay recognizing how
the content is--how a website is-- is done once
a year or gated in any way.
BARUCH LABUNSKI: So the minute
that a website changes, thenit can affect the ranking
positively or negatively.
JOHN MUELLER: Sure,
like any other changeyou'd make on a website.And some of these are more
of a technical nature.So for example, if you
changed your internal linkingsignificantly, then
that might be somethingdue to the template change
that you've made, which couldessentially it's affect
the way that your siteis crawled and indexed.So maybe we can find the
content that was missing before.Maybe we can't reach
that content anymorebecause it has such a
fancy navigation that youhave to go through like a search
field or something like that.So these are things
which essentially happenautomatically all the time."Will Google consider using
HTTP2 as a ranking factorto speed things?"I don't see that
happening any time soon.HTTP2 is something that I
suspect a lot of websiteswill just have enabled
by their serverautomatically where you won't
have to do anything special.And it'll just work
in the background.So this isn't
something where youhave to artificially create
something on your serverto make that work.It'll essentially
work transparently.One of the aspects here with
HTTP2 is that a lot of browsersonly support some of
the features in HTTP2if your site is
already on HTTPS.So that's one thing that
might count towards,well, I'll move to
HTTPS as a preparationfor making sure that my
site works even better whenHTTP2 rolls out to my server."As a way of dealing
with scrapers,I'd like to add the x-robots
tag to my XML site mapsand then permit crawling
only to Googlebotand some other
credible crawlers,but not to row
crawlers or end users.Could that be
perceived as cloaking?"No, I don't think that would
be, in any way, problematic.With sitemap files, you can
even cloak them directlyto search engines.That's something that
we would explicitlyallow where if you
test the IP addresses,and you see it's not a Google
or Bing or Yandex or Yahooor whatever IP address, you can
serve a not allowed page if youwant to.So that's the kind
of situation wherethis content is explicitly
only for search engines.So you can choose to
really explicitly onlyshow it to search engines."I've been trying to
clear some manual actionsdue to unnatural links.I've just had two
reconsideration requestsrefused.I had no sample links
quoted in the reply.Is this now the standard?"No, I wouldn't say
this is the standard.But sometimes that happens.So that's the situation
where what I would dois go to the help
forums and get advicefrom some other peers who
have gone through this processas well, which might help to
point you in the directionthat you'd need to work on."Our source code was copied
and iFrame placed over itto create websites
hosted on ad servers.We found over 50 sites
that still show our linksand link back to us.Is Google able to spot
that this is malicious?And other than disavow,
what can we do?"I think disavow
was a great thingto do in a case like that.You might consider using the
spam report tool as well,depending on what exactly
is happening there.But in general, we recognize
this kind of situation.And we should be able to
handle it fairly well.If you see that
we're not handling itwell where some hack site
is shown your contentand ranking above
you, that's somethingwe'd love to hear about.So I'd post in the
help forums fr that."Our site rankings
have been under attacksince beginning of December.Someone copied our site
and inserted some contenton totally unrelated sites.Some of those sites are
ranking higher than us.What can we do?"So, again, in this
case, the web spamreports you mentioned
that they're not helping.That's one thing you can do.Another thing would be
to post in the help forumso that we can get
these escalationsand look at them directly.
BARUCH LABUNSKI: Why
don't he just DMCA it?
JOHN MUELLER: That's obviously
sometimes another thing.But that's something
where you'll probablyhave to look at the legal
situation behind that first.That's not something where I
could make a recommendationsand say, well, you should take
this specific legal actionbecause, obviously,
I'm not your lawyer.I can't give you legal
advice like that."For a website, the robots
tester in Search Console shows
me that a page is blocked by
this rule disallow: *lightbox.Is there a syntax mistake?"Yes, all disallow
and allow linesshould begin with a slash.So in this specific
example, it has
disallow: and then a space, and
then an asterisk and light box.So it would need a slash
before the asterisk there.And then essentially that
should work as expected."In the last session,
you mentionedhidden text is discounted.If relevant hidden
text is placedunder appropriate
headers to makethe page less cluttered
and easier to use,why should that be discounted?"So from our point
of view, the contentthat people see when
they visit your pageshould be what
your page is about.And if there is important
content that people don'tsee when they open the
page in the browser,then it feels like there's a
mismatch there that you thinkit's important content,
but at the same time,you don't actually want to
show it to people by default.So that feels like a
kind of a mismatch there.And if you feel it's
important content,then I'd recommend putting
it on a separate URLso that we can index it there,
so that it's directly visiblethere, and you can still use
within your siteso that pages don't have
to reload and take timeto be refeteched
in the browser whenpeople navigate between
specific parts of your site.But by default, if
you think somethingis really important
for your page,make sure it's
visible by default."Am I correct in understanding
that Google Search is nowrewarding field topic experts
while penalizing thosewith a university degree in the
journalistic style of writingwhich enables
translation of fieldsinto easy-to-understand
copy for the masses?"I don't really
understand this question.But I think it kind of
goes in the directionof why don't you give
real journalists morevisibility in search?They know what
they're writing aboutcompared to these people who
are just writing for the web.And from our point
of view, both sitescould be generating
great contentand could be generating
something that we'dwant to show in search.So it's not that
we're explicitlysaying that one type of content
is better than the other.It's just they're different
versions of content out there.We try to rank
them appropriately."What , according to Google,
is the best way to optimizean entity page like some
celebrity or any topic likethe Nepal earthquake?What kind of content
does Google expectis relevant for
its users and alsohelps us understand that this is
what Google end users wanted?"There is really no
simple rule with regardsto how to create a great
website or a great pageon a specific topic.So that's something where
you as a topical expertwould need to bring in all of
the information and knowledgethat you have.Also, maybe information
about your readers,about the people who are viewing
your web pages so that youmake sure it kind of matches
what they'd expect to find.So there is no simple rule
where we would say, well,you need to have this much HTML
and this clear structure here.Some sites are really
creative in the waythat they present
the information.Other sites present it in
a more straightforward way.And both of those sites have
reason to show up in search.There is no one simple answer
that solves everything.So take your information.Take the knowledge
and experiencethat you have on this
topic and present itin the way that works
best for your users."All of us here must compete
with Google's promotionon its domain, it's page, and
Facebook ads that the websiteowners choose
among Weebly, wiks,and et cetera for the
web building needs.Can Google also add
a line to promotionto use local search for
developers/designers?"I don't quite understand
this question.Maybe if you're
listening in, maybe youcan reformulate it and submit
it again or post it in a chatif you are here in the Hangout.But at the moment, I'm not quite
sure what specifically you'relooking for."Google is not using the keyword
metatag as a ranking signalanymore.But is it if given
use for anything elseas there still might be
useful information in there?Or do you just ignore
it completely?"As far as I know, we just
ignore it completely for search.I believe maybe AdSense
uses it, but I reallydon't know if that's
still the case.
BARUCH LABUNSKI: Google News.
JOHN MUELLER: Sorry?
BARUCH LABUNSKI: Google News
JOHN MUELLER: Google News.OK, Google News in that case.So if your site is
listed in Google News,and if that's
correct, I don't know,then maybe it makes sense
to keep those there.In the past, I've also
use the keywords meta tagjust as a helper
for myself to knowthat this page is on
this topic and keepinformation focused like that.But essentially for web search,
you really don't need it.You can put whatever
you want in there."How does Google
treat SWF content?I run Flash gaming websites.And these games are
distributed by game developers.How does Google then measure
duplication of Flash games?The only way I understand it is
to write a unique descriptionof the game."Yes, I guess in some
cases, we can actuallypull out the content
of Flash filesand include that as a
part of the web page.So that's something that
sometimes we're able to do.Having unique description
for the Flash file definitelyhelps.With regards to site
setter, essentially justa combination of
existing Flash files.That's I guess always a tricky
situation because you reallyneed to provide something
unique and compelling that'smore than just kind of a
spun version of the text.So from my point of view,
what I would look atthere is to make
sure your site shineson the usability
on other aspectsthat are involved there with
regards to really bringingsome unique value to
people who are lookingfor these type of games.So that's something where
I wouldn't necessarilytry to rank for these specific
Flash files, especiallyif you know that they're already
duplicated multiple timesacross the whole web."I revised one of my sites.The site has a few links
to another site of mine.The URLs are not
correct anymore.But there's a 301.Should I just that
correct the URLsand let the 301 do its job?Or does Google give
more value to an H linkthan to a new one?"From my point of view,
whenever you can link directly,I prefer to link to directly.It makes it a little
bit faster for the user.But from a practical
point of view,if you're talking about a
handful of links from one siteto another, that's
not really somethingyou end up seeing any
visible change in the searchresults because of that.So that's something where,
on the one hand, sure,it's a great idea to link
directly wherever you can.With a 301 at least people are
making it to the right content.And it's kind of like you
could do it both ways.I wouldn't necessarily
assume that you'dsee any significant
ranking change justby going one way to other."If a great number of URLs,
more than 1,000 URLs and 1,000domains was removed
from a disavow file,can be seen as a link boom?Could it be understood by
Google as a spam signal?For example, if the URLs were
added by mistake more than halfa year ago."I don't see any
problem with that.That's not something
where I wouldsay this would be seen as a
bad thing or is a good thing.It's essentially just
a technical changethat you're making.You're telling us that
when we recrawl those URLs,we should take those
links into account.And if those are normal, good
links, then by all means,let us take those into account."We've seen a quite
substantial clickthrough ratedrop in Search Console for
apps for all of our accountsfor the 16th and
17th of Decemberwithout search appearance
filter applied.It's a few days after the
search appearance filter beingapplied to Search Console.What can be the
reason for that drop?"I don't know offhand.I'd have to take a
look into My Accountto see what all was
happening back there.The tricky part specifically
around app indexingis that this is
still fairly new.And we do make
changes with the waythat these apps are
shown in search results.And some of those changes
can have a significant effecton the clickthrough rate, on the
impressions, that you're seeingfor these apps specifically.So that's something where
I would kind of expectto see things fluctuate
for a little bituntil we actually
figure out how to handlethese apps in the search
results for the long run.And I would assume
that, I don't know,I'm just making a guess in the
next half year or year or so,things will be
significantly settled downthat you wouldn't see such
strong fluctuations justrandomly happening."Does excessive nofollow cause
de-ranking of a website?"No.If you use nofollow
on your websitewhen linking to other sites,
that's perfectly fine.If you use nofollow within your
own website, then, of course,that could cause problems where
if we can't crawl your websiteand pass Page Rank
internally, then that'ssomething that makes it really
hard for us to crawl and index.But if you're just
meaning that your site hasa lot of nofollow links to other
sites, that's your business.That's the way
you run your site.That's up to you."When having
multiple URLs indexedwith almost duplicate content
caused by a former relaunch,is it better to use 301 redirect
or to choose a canonical tagto fix that?Old pages will be noindexed
as well but still shown."If you're moving URLs
from one URL to another,I'd use a 301 redirect.If you have to keep them up in
parallel for whatever reason,then use a rel canonical.But it's not that there's
like one version betterthan the other.It's essentially what
are you trying to do,and what's the
best tool for that.It depends more on
internal links in casethey're not closed in
robots.text and usefulfor Googlebot.Can this action cause any
penalties from Google?Will page rank
still be transferredthrough these kind of links?Or will they work like
rel nofollow links?"This wouldn't cause any
kind of manual actionfrom our point of view
as long as this isnormal content on your pages.And these links will still
pass page rank normally.So a really common
use case for thisis if you have a
kind of navigationon your site that has
a menu where you clickon the top-level menu item,
various parts of your site.That's something
we've had to dealwith for a fairly long time.We were pretty good
at figuring outhow that's supposed to work.
MALE SPEAKER 2: John,
if I may, I will add.So those are just
functional internal links.Those are site-wide links
like log into my account,drag my order, stuff like this.So I was referring
to those links.
JOHN MUELLER: Yeah,
that's perfectly fine.I don't see any
problem with that.
ROBB YOUNG: I think somebody
else asked something similarin the chat, didn't they?Whether those links would
be considered cloaking?Basically.
JOHN MUELLER: If you're
them when we render the page,then they're essentially
normal links.So it wouldn't be like clothing.
ROBB YOUNG: And it says, in
an effort to combat scrapers,we tested reverse
links with jQuery.Would this be
considered cloaking?Since when the link is hovered
over by a real visitor,the real link is revealed.
JOHN MUELLER: OK,
so if it's on hover,then it turns into a link.What would happen in
a case like that iswe probably wouldn't
pick those link upsbecause Googlebot isn't
going to hover overevery part of the page.It'll pull out the page.It'll render it
once like a browser.It's not going to
interact with the pageto see what's actually
going to happenwhen they do specific things.So if you need those links
to be found by Googlebot,then make sure we can find
them when we load the page.If you just want to make them
available for users, then sure,I think that might be an option.I think, in most cases,
you wouldn't reallywant to do that.And if you are having
problems with scrapers,then I'd try to find something
different to attack thatmore directly than to obfuscate
the links like this, whichcould essentially
end up causing moreproblems for your
website in searchthen the scrapers ever were.
ROBB YOUNG: OK, I was asking
that on behalf of someone.I think they were already
in the chat, actually.
JOHN MUELLER: All right,
I'll have to check outthe chat afterwards."What are the top
three stories you wishI didn't write on my site?"Oh, jeez, I don't know, Barry.Your site is so big.Where do I start?Oh, wait, that you
wish I didn't write.OK.
MALE SPEAKER 3: You can skip it
if you don't want to answer it.There must be one
that comes to mind.I'm just curious.
JOHN MUELLER: I don't know.I can't think of
anything at the moment.
MALE SPEAKER 3:
It's all that good.
JOHN MUELLER: You've
picked up so many thingsthat it's going to be natural
that every now and then,you'll run into something
where I'm just, oh, god, Barry.But yeah, I don't know.I think that's
normal if you end uppicking up so much
content about what'shappening around
search that there'sbound to be one or the other
that one or another personmight not appreciate.Overall, I think
it's really helpfulto have such almost like
an encyclopedia of SEOof what's happening
around the web.
ROBB YOUNG: He'd just
deny everything anyway.What's the difference?
JOHN MUELLER: All
right, site links.Is it possible to get
more than six site linkswhen searching for
brand or company name?I don't know.Maybe.I know this is the type of
UI where people are alwaystweaking on it and seeing
what makes more sense.But I don't know what
specifically thereare limits there or not.So if you've seen
this in search,then obviously it's possible.If you haven't
seen it in search,then maybe it's not
possible at the moment.But maybe it's something that
will happen in the futureor that has happened in the past
where we've tested and said,well, this is maybe too
confusing to users, too many,or maybe we just haven't
tested it out yet."I own a large image website.Is it OK to have the meta
title, image title, and productdescription the same until we
get product descriptions moredefined?"Sure, that's perfectly fine.What will probably
happen is we'll justpick up one of these versions
and keep that to showin the search results.But if they are all the same,
then they are all the same.Then that's not
something where wewould create some kind of
a keyword stuffing filterand remove those pages
from the search results.It's essentially a technical
issue and not an SEO problem."In Webmaster Tools,
or Search Console,I guess, we saw most
of the site URLsget removed from index
status after changingfrom HTTP to HTTPS.Yet our analytics weren't
affected this way.What's the best course of
action to take at this point?"This is something
that's really common.And we've seen a bunch of
people get confused by this.But essentially, Search
Console, if you have your siteas HTTPS or HTTP listed
in Search Console,then that's the
URLs we'll focus on.So if you move your site from
one version to the other,you'll see the number of index
URLs go down in one versionand then go back up
in the other version.So what you're probably
looking at hereis the HTTP version of your
site where it goes downwhen you move to HTTPS.And if you add the HTTPS
version of your site,then you'll see that it's been
going up at the same time.So especially if you're saying
the analytics weren't affected,then what's
essentially happeningis we're shifting
the URLs from oneversion to the other version.And all your traffic is
going with that as well.So that's not something
you need to fix per se,but you can find the
right informationby looking at the right version
of your site in Search Console."Is there some way to
access and download all datafrom Search Analytics
of Search Console,not only in the
last three months?"I think that would
be a great feature.I totally agree with you,
that would be great feature.At the moment,
that's not possible.I know there's some tools that
use API to download this datastep-by-step, where if
you start downloadingthe data, then, of course,
over the course of a coupleof months, you'll
be able to collectthis data during that time.But those are essentially
external toolsthat use the API to
download this information."Some research papers aren't
being properly indexedby Google Scholar and others."This sounds like
something I probablyneed to have more details on.So if you could post that
into the help forums,that would be a great thing.And I'll try to
pick it up there."With the new AMP
blog article markup,should that only go
in the AMP-only pages?Or should it also be added to
the original version as well?I notice the image
objects aren't followingthe same rules as normal.URL appears to be replacing
schema.org content URL."If you're doing something
AMP specific on your pages,then I would put
that on the AMP pagesspecifically, not on the
normal desktop pages."What's the optimal size or
character link for a title?"I don't know.I would look at
the search resultsand think about what would
makes sense for your pages.And this is something
that probablydepends a bit on your
site, probably dependsa bit on your content.I try to keep the title
to something that'snot keyword stuffing that
describes what your page isabout so that people, when
they see it in search results,they can react to
it as appropriate.But that's less
of something whereI'd say it's an SEO issue
and more of something whereyou present your site this
way in the search results.And that should be
fitting for your pages."In an AngularJS site,
the generated head partis considered by Google.So for example, is it
update the head of your page,then as we render
those pages, we'llbe able to pick that up and
use that as appropriate."Can we use a Google
Knowledge Graph APIfor improving our entity pages?Like we have a page
about Taylor Swiftand wanted information
about Taylor Swift.Can we fetch this
data from the APIand show it on our website?"Sure, if that's
what you're usingto provide additional
value for your users,feel free to go ahead.If your website is only content
that's great from other sites,then obviously that's something
that I'd try to discourage.But if you're providing
additional value to usersthrough APIs, through feeds,
or something like that,then by all means, go for that."SVGs can be included
on the web pageeither through an
image tag, object tag,or embedding the SVG directly.Is there any difference as
to how these will be indexed?Will all three approaches
lead to the imagesappearing in image search?"I believe all three
would work nowadays.So if used to be that
you'd have to hostyour SVGs as separate
SVG files and embedthem kind of like images.But I believe all three of
these variations will work now."Any idea when the real-time
Penguin is coming?"No, not yet."I am the webmaster
of clickrent.com.We have an AngularJS system
online for about a year.The sitemap exists in
Search Console-- stillseem to have indexing problems.Our external links are rare
and internal link structureneeds optimization.Do you see any
problems with Angular?By default, I don't see
any problems with Angular.I know there some really
good Angular-based websitesout there that do
fairly well in search.Of course, there
are probably waysthat you can set up an
Angular website thatdon't work that well in
search like you could do thatwith any other
kind of framework.So what I'd recommend
doing there is maybetaking your site, posting
it to the help forums,or posting to somewhere else
where webmasters and SEOs hangout and get other people's
feedback on your sitespecifically.Maybe there are simple things
you can do to improve it.Maybe you are doing
everything right,and it's just a matter of
having even a better websiteto present to users
through search engines."I am new to SEO and
how Google works.I manage my family's
taxi company.We are a local company.We have a business page.And we have our contact
details on a contact page.Also, we now have an SSL
certificate on our site.Will this help rank better?"So SSL certificate, essentially
moving the content to HTTPS,does help a tiny bit.But it's not something
where you'd probablysee a significant ranking
change in the search results.So in a situation like this,
I would consider reallyworking on your pages directly.I'd probably also get help
from people in the help forumsbecause they've seen a lot of
these types of small businesswebsites and can give you
a little bit of informationon what you could
be focusing on,what you shouldn't
be focusing on.In some cases,
maybe they'll eventell you you shouldn't be
focusing on the website at allbecause this niche is way too
competitive for someone newto just jump in and
try things out on.Maybe it makes sense to
be a bit more creativeand drive traffic to
your site in other ways.So this is something
where there'sno simple rule that works for
all websites where it's reallybest to get other people's
advice on your website,your specific business, and
what you're trying to achieve.Let's see-- a few
questions left."What words of
encouragement would yougive webmasters who do the
right thing such as builda great user experience
website-- HTTPS,mobile-friendly, and
work countless hourson their websites only
to be outranked by thosewho use blackout techniques?"That's always a frustrating
situation to be in.That's true.What I would
recommend doing thereis maybe also going
to the help forumsand checking out
with the other peoplethere to see that you're not
missing something totallyobvious and to see what
you can do to improveyour site in general.The thing also here
to keep in mindis that sometimes
these sites thatalso use black hat techniques
do a lot of things really,really well.So the website might be doing
some kind of sneaky stuffon the side, but otherwise has
a really fantastic website.And in general, our
algorithms wouldtry to recognize
that sneaky stuff,ignore that, and rank
the website basedon the other aspects
that are involved there.So just because they're
doing something sneaky,doesn't necessarily mean that
they're profiting from thator that they are
ranking because of that.Maybe they're
ranking essentiallydespite of that sneaky stuff
that you're doing on the side.So if you recognize the sneaky
stuff that people are doing,then that's a good
sign because thatmeans you'll probably avoid
those kind of sneaky thingsas well.But that doesn't
necessarily meanthat your website
will automaticallyrank above these other ones.So continuing to
work on your websiteand really taking the time
to take an honest lookat your website,
at your business,at how things are working,
always makes sense in my view."I'd like to ask a question
about the link badges.So images with keywords
alt tags for many sitesen masse-- for example, members
of a forum, for example, thislooks a lot like a unnatural
link building to me.Is this method not recommended?"I think you'd probably have to
look at the individual casesfor something like that.I haven't really seen any cases
where this would significantlycause any problems or cause
any significant advantagesfor a website.I think if you're the
type of website thathas a really strong
community behind that, that'salways a good thing to have.And that really helps your
website over the long runbecause these are people who
will come back to your websiteregardless of whether
or not your site isshown in search at all.So that's something
where I'd sayit's a good thing to have this
strong community behind that.And just because
they're using badgesto promote their part
of this community,I don't know if that
would necessarily alwaysbe a bad thing or
always be a good thing."They use a block
quote tag as a signal.If I quote another source
with a block quote tagand link back to it,
do you recognize that?"No, we don't, at least
as far as I know,we don't take that
into account at all.So if you quote another
website, then we'llprobably pick up that this
is a quote from that website.But it's not going to be
because of the block quote tag.You can use other type of
markup there if you want."Two months after
a domain change,our rankings are normal.But we don't see the
rich snippets starratings for reviews.Does this mean Google
needs more timeto determine the
credibility of a new websiteor how long can that take?"It's hard to say
because sometimes wedo we re-evaluate the
quality of sites in general.And maybe you've just
coincidentally hit one of thosetimes where we updated our view
of what we would want to showin rich snippets and
what we wouldn't.So it's not necessarily
related to a site movethat you see these rich
snippets star ratingsor that you don't
see them any longer.All right, wow!I ran through those.We still have-- what is it?Almost 30 minutes left.So what's on your mind?
BARUCH LABUNSKI: Regarding
private blog networks,how does Googlebot
JOHN MUELLER: I don't know.
BARUCH LABUNSKI: If
somebody, I mean,like if there's spam going
on with something like that,how do you guys deal with that?
JOHN MUELLER: I
would imagine we dothat the way we handle
most of this kind of linkspam as a mixture of
automatic and manual actionthat we'd have to take there.So I don't think there's
one simple recipe for howGoogle recognizes link spam.
BARUCH LABUNSKI: OK.
JOHN MUELLER: Let me just
double check what was happeningin the comments event.It looks like Chrome
keeps crashingfor some people, sorry."Will it take much
longer to rankfor domains that have
been previously parkedfor many years?"There is probably
a period of timewhere we consider something
still to be parked,even though it has content now.But usually, that should
clean up fairly quickly.So that's not something
that there's a manual timeframe there that things
have to wait out for."There are parts of our
website I'd like to index.But I have no interest
in boosting itfor the results page."I don't quite understand
your question there, Kenny.So if you can post that
in a different way,I'll try to pick that
up for the next round.
JOHN MUELLER: All right, what
have I missed in the chat?Lots of stuff happening.
MALE SPEAKER 4: OK, John,
can I ask a question?
JOHN MUELLER: Sure.
MALE SPEAKER 4: You
just said that-- Imean, if somebody is actually
using black hat techniquesand saying also he's
using the good techniques,don't you think that
Google should not evenlook at the good
techniques he hasput because he's already using
the black hat techniques,because we are struggling
all-- we are putting allthe rules which you tell us.And we don't use any black
hat and stuff like that.
JOHN MUELLER: I don't know.From my personal
point of view, Iwould prefer it if we could
recognize the problematic partand just ignore those
and focus on the restbecause I know there are
lots of people out therethat have followed bad advice
in the past that maybe have hadtheir website set up
in a bad way by maybean older SEO from older
times when maybe this workeda little bit better.And it's not that these
people are explicitlytrying to do something
bad and harm usersand spam the search results.It's just they've
gotten bad advice.They've done something stupid.And they don't even realize
that this is a problem.So in all of those cases, I
think the smarter move wouldbe to just say, well, we
see that this sneaky stuffis happening here.But we can ignore that.And we can be certain that the
quality of our search resultsaren't impacted by people doing
this kind of sneaky stuff,then we'll try to ignore that.On the other hand, if
we say, well any typeof problem on your website
will cause your websiteto be removed from
Google permanently,then I think we would
have pretty empty searchresults pages, because lots of
people do things accidentally.And it's not something where
I'd say just because youaccidentally did
something wrong,we should remove your
website completely.But obviously, finding
the right balancethere is very hard and
especially in situations whereyou're not sure that you're
able to really isolate this oneproblematic thing completely,
then that's a lot harder for usto figure out what the
appropriate reaction would be.
MALE SPEAKER 4: OK, thank you.
BARUCH LABUNSKI: Is
it a problem if youblock your who-is information?
JOHN MUELLER: I don't think so.
BARUCH LABUNSKI: There
was a lot of criticismout there for a long
time that if youdo that, it's not a good thing.You guys don't look
at is a good thing.
JOHN MUELLER: I don't think we
take that into account at all.
BARUCH LABUNSKI: OK.
JOHN MUELLER: If you
want to block that,from my point of
view, that's fine.Sometimes it's really useful
to have that information.So I use that sometimes when we
get an email from some indexingor crawling people internally
who are saying, wow,this website is this doing
something really, really wrong.We need to let them
know about that.And sometimes the
who-is informationleads us to the right people who
can actually fix that problem.So from my point of view, I
like having that information.But it's not something that I'd
say we need to have for search.
BARUCH LABUNSKI: No, just
that if it was registeredby a lawyer, you don't
want your lawyersinfo and all that stuff.So I just wanted to
know if it's a problem.
JOHN MUELLER: Yeah.I don't know.I wouldn't see that
as a problem someone.
Making it transparentis even better, yeah?
JOHN MUELLER: I don't
think we would take thatinto account at all for search.So if you think it helps
for your users, fine.If you prefer to stay
anonymous or stayhidden behind that who-is
domain privacy, that's fine too.
ROBB YOUNG: Is that not part
of-- unless this is also nottrue-- I read that
having a PO box or notreal addresses on your
site can also hurt.Is there any Google
JOHN MUELLER: That's probably
just for local search results.But I don't know how the local
search often handles that.
For Google Places,you need a real address.You can't have a PO box.
JOHN MUELLER: I don't think we
take that into account at allfor search .
ROBB YOUNG: OK.
MALE SPEAKER 2: How about
the duplicated content?Back to my question about
the car-related website,if you remember, John.
JOHN MUELLER: Mhm.
MALE SPEAKER 2: So
I was just wonderingif you had time to take
a closer look to itor possibly did you come up
with some kind of advice for us.
JOHN MUELLER: I'd have to
take a look again, sorry.
MALE SPEAKER 2: Should
I resend it to you?
JOHN MUELLER: Sure.Sure.You could always
ping that again.
MALE SPEAKER 2:
Yeah, because we justcan't sleep because we think
there are some manual actions.That's why we're wondering.
JOHN MUELLER: If there
were manual action, thenthat would only--
MALE SPEAKER 2: Sorry,
I meant automatic.
JOHN MUELLER: Yeah, OK.
MIHAI APERGHIS: Hey,
John, Google AdWordsallows you to track calls to
a phone number on a websiteby using a script
that changes the phonenumber on your website
for visitors thatcome through ads, for example.Do you think that
will be possibleor something that we can maybe
expect to be available for SEOas well?Or people coming from
organic search maybetrack calls via Google
forwarding number as well.I know there would be
some issues regardingchanging the phone number,
cloaking, things like that.That would be interesting
to see exactly howmany phone calls maybe leads a
person gets from SEO efforts--organic results--
JOHN MUELLER: Probably,
I would expectthat you can already do that.I assumed people were
already doing that.So from my point
of view, I thinkthat would be unproblematic.The tricky part, of course,
is if you have thingslike what is it?The local business
markup on your pages.And you're marking up your
phone number like this.And the phone number is
changing from time to time,then that might make it harder
for us to really figure outwhat local business phone
number we should reallybe associating with the site.But if you are doing this
essentially in a scriptthat we never see, then I
don't see a big with that.It's not that
you're significantlychanging your site and cloaking
something to your users.You're essentially just
swapping out the phone numberand using a session ID there.So I would assume that
would be unproblematic.
MIHAI APERGHIS: OK.It would be nice to see an
official script like GoogleAdWords ads that specialists
can implement on the clientwebsite or webmasters.
JOHN MUELLER: I doubt
you'd see somethinglike that coming from our
side because from search, weobviously liked it to see things
as straightforward as possible.But I suspect this is something
that would be trivially donewith jQuery where you just
swap out an element on a pageand show a different number
instead of a existing number.
MIHAI APERGHIS: Well, yeah.It's just that Google AdWords
offers a Google forwardingnumber and allocates
it automatically.And it has all those
advantages, notfigure out what phone
number you need to switchand things like that.
JOHN MUELLER: Yeah, I don't
know how they do that.Maybe it's trying it out and
doing a blog post about it,letting us know how it works.
MIHAI APERGHIS: One more thing
regarding Search Console,Search Analytics specifically--
this is more feedback, I guess.You have that average
position metric,which is aggregated over all the
keywords and their positions.I don't know if it's very
helpful in certain situations.For example, I create
a new landing page.And right after I create
it, I stopped rankingfor certain keywords
regarding that landing page.And I don't trying to
get high right away.So my average position
might actuallybe lower versus last month.But I didn't lose any positions
for the other keywords.So it might be a bit confusing
for certain webmastersthinking, oh, OK, so Google
started to down rank meor I was ranking or
something like that.Maybe something like an
average position changeor changes in position,
something like that.Maybe that would
help a bit more.
JOHN MUELLER: I don't know.It's hard.I think especially with
the ranking information,it's something
where more and moreyou really have to
be able to interpretthat information properly.And like the situation
you mentionedis one aspect where either
part of your website toor part of new content
to your website.And obviously, for
those words, itwon't be ranking that
well in the beginning.That's one aspect.The other aspect is sometimes
a different search resultselements on a page where maybe
they'll be, I don't know,15 entries on a page
that could be on page onewhere you're ranking
15 is still page one.And other times, maybe
your ranking is number 9and you're actually on page 2.So these kind of
additional aspectsthat come into play with
all the modern searchresults pages-- there's
something where you reallyneed to take the
time to interpretthe information that you're
getting out of these toolsand make sure that you're
pulling the right results outof that.
MIHAI APERGHIS: Right.And regarding that same metric,
I know that for certain queriesthat there is a very
low search volume.You don't show them
in the queries report.Does the average position
take that into account.For example, if there's a key
word that hasn't been searchedthat month , its position isn't
taken into account versus lastmonth, for example.There are no impressions
for that keyword.
JOHN MUELLER: Well, if
there are no impressions,then there would be no
ranking information for it.
MIHAI APERGHIS: OK.It's not like you're
tracking the actual keywords.
JOHN MUELLER: No.No, that's I guess sometimes
a part of the confusing thingthere.It's not that we're
tracking the rankings.It's that we're tracking what
we actually show to people.So that could be that sometimes
personalized informationis in there where the
personalized ranking at numberone could be something
very visible for someof these, which might
skew your average ranking.But that's essentially
just because that'swhere we showed it this time.And it's not that we
have any kind of a fixedranking for a page in a query.It really depends
a lot on the user,on personalizations, their
location, all of that.
MIHAI APERGHIS: Right.OK.
Do you guys evertake Google Page Insight to
make it a ranking signal?
JOHN MUELLER: Google
Page Insight--which one do you mean?
BARUCH LABUNSKI: Where it shows
if it [INAUDIBLE] code-wise.[INAUDIBLE]
JOHN MUELLER: Like
the speed information?
BARUCH LABUNSKI: Yeah, exactly.So remember we talked about
passing the 83 as they werementioning at Google
I/O. So just wonderingare you guys ever going
to-- if it's necessary.
JOHN MUELLER: I don't know.Are you ever going to is
a really tricky questionbecause all I can say is maybe.But I think speed is probably
something we'll be takinga look at more with regards
to mobile specificallybecause as we see more and
more sites move to somethingthat's kind of
mobile-friendly by design,but still almost mobile hostile
with regards to usabilityon mobile devices, I think
speed is one of the next aspectsthat's going to be more relevant
there where we can see--
I agree with youbecause I had that issue
where I really wanted to know.I was in Toronto.I wanted to really know what
the weather was this onespecific website locally.And if it wasn't working
because the signal of the hotelwas really bad.So reaching that site
took me over a minute.
JOHN MUELLER: Yeah.
BARUCH LABUNSKI: But
Especially on mobile.That's something where I
think the next big hurdle is.And that's something where
we see a lot of sitesshift their UI to something
that works on a mobile device.But actually, every time you
open a page on that site,it's over 1,000 requests.And it takes a couple
of minutes actuallyto load all of the content
to render it in a way that'sactually visible.So I think speed is
potentially somethingthat we could look at there
where we could say, well, thisis a really fast mobile site.Therefore, we should
make sure they are reallytreated as mobile-friendly.Whereas this site
renderers on a mobile.But it's a really terrible user
experience-- really, reallyslow.It has a lot of
different embeds.It does everything
wrong with regardsto making something
that works quicklyand snappy on a mobile device.So maybe we shouldn't treat
it as mobile-friendly.
BARUCH LABUNSKI: So if
it's over five seconds,then it's a problem.
JOHN MUELLER: I don't know
where we would draw that line.That's really hard
to kind say there.But this is something
that you seewhen you interact
with websites yourselfas well where if
you're on your phone,you're trying to do something.And you can load the page or
you can see that it's loading,but it just takes forever
to actually load, thenthat's not something you're
going to be sticking aroundwith on the website.And if Google points
you at that websiteand you know there's
an alternative that'sreally snappy, then you'd
almost say, well, Google,why do you hate me as a user?Why do you send me
to the site that'sso terrible on my mobile phone.
BARUCH LABUNSKI: Yeah.OK.
MIHAI APERGHIS: John,
any chance I can askyou a site-specific question?
JOHN MUELLER: All right.
MIHAI APERGHIS: It's
a publisher website.It's a pretty big website.And we've been working with
them for three years now.But just for the past
six or seven months,traffic has started trickling
down maybe 5%, 10% per month.And it's really puzzling.We cannot find a specific issue.I know there's stuff to
be worked on, of course.It's just I was
wondering maybe youcan see something you
should first focus on this.Maybe this is a problem.It's fine.There's nothing
we can say either.But I had to try.
JOHN MUELLER: I'd probably have
to take a more in-depth lookthere.I don't see anything
offhand where I would say,well, you obviously forgot
this one stupid thing.Therefore, your website
is dropping in search.Sometimes, it just a way
that the whole environmentis shifting over time
where maybe thereare some really strong
competitors thatare coming up or
maybe people justaren't looking for that
type of content anymore.It's really hard nail
these slow changes down.
MIHAI APERGHIS: Right.Well, one of the
problems or thingswe are trying to implement
is having a static URLor like a reviews, because
we do car reviews and thingslike that.And every time there's a
new model, the URL changes.And we're trying to get a
static URL from year to year.So it gets all of those signals.We're also trying AMP
right now to implement--see how that goes.
JOHN MUELLER: Yeah.I suspect both of those
will help a little bit.But I don't think you'd see a
really strong jump from eitherof those because
those are like tweaksthat you're making on top
of the existing website.And they don't significantly
change the whole modeland the whole environment of how
people interact with your site,how they want to find it
in the search results.That's something that sometimes
just changes over time.And finding out
why it's changingisn't always like to due to a
specific, SEO technical typething on your website.It's maybe just because
people are changing and doingsomething else instead.And maybe that's something
else that you can alsodo on your site.Maybe that's something
else where they say,well, we're going
away from cars.And everyone is just
taking hoverboards now.Therefore, all car
websites are obsolete.These type of things
can always happen.
MIHAI APERGHIS: Sure, well,
we did check our competitorsand try to see where they stand.And we know there's lot
of traffic potentialthere-- a great
lot of potential.So I know this is a subject that
people are still looking for.I'm just trying to figure
out how to, as you said,try to target them to
see what exactly they aremore interested in, I guess.
JOHN MUELLER: Yeah, I guess
what I would do in a caselike this where you're seeing
this kind of steady declineis really do some user studies
to figure out what is actuallyhappening there to see if these
other sites are just rankingsubtly better than your
site-- if it's justa matter of ranking, or if
it's a matter of your sitenot really matching exactly
what they're looking forand really trying to figure out
where does that mismatch comefrom-- where do
these changes happen.Because especially
if it's somethingthat's subtly happening
over several months,it's not going to
be an SEO problem.It's not going to be something
where you can just say,oh I forgot the
rel canonical here.I'll fix that, and it'll
jump back up again.It's probably more of a
general problem than that.
BARUCH LABUNSKI: What
about recording the users?He can record the users
and watch their behaviorand maybe change.
JOHN MUELLER: Sure.
BARUCH LABUNSKI: You know, using
pop ups like OptiMonk whichwill reduce the bounce rate.
JOHN MUELLER: I don't
know if using pop upswould be the right thing.But doing user studies by
recording what they're actuallydoing on your site, that's
definitely a good thing.Maybe doing a lot of A/B testing
to see what kind of contentperforms better-- what kind of
UI performs better for users--where can you get more than just
like a few percent out of it.All these things
I think would bereally useful to
help understand whatis actually happening there.
MIHAI APERGHIS: Thanks.By the way,
regarding AMP, do youthink that once it's launched
it will have a lot of queries,maybe you know a percentage.
JOHN MUELLER: I don't know.
MIHAI APERGHIS: But you
mentioned it will firstfocus on newsy queries.
JOHN MUELLER: Yeah.The demo is specific for I think
in the news box that we havewhere some sites do see a
lot of traffic from that.And that's something where
generally people kind of clickthrough it.They want to read an article
and see different views.So that's something
where if you have,I don't know in your case,
like car specific news.And you provide
them in that way.And we show them
in the news box.And that's probably
something whereyou would see a bit more
traffic just because peopleare actually able
to actually seethis content a little bit
faster and then interactwith that directly
a little bit faster.
MIHAI APERGHIS: OK,
and what happenswith those sites that don't
implement AMP or those queries?I'm guessing they will
see a pretty big decrease.
JOHN MUELLER: I don't know.I mean, on the
one hand, we don'twant to use specific technical
elements as a way of saying,well, you have to
implement this.Otherwise, we won't
show you in search.On the other hand, if we show
things like a carousel with AMPcontent and your site
doesn't have AMP content,then we can't include
you in that AMP carouseljust for technical
reasons, because wedon't have that content
that we can actually there.So that's something
maybe some of these siteswill see a change in
their search traffic.Maybe other sites that
wouldn't traditionallybe shown in there anyway.They wouldn't see any change.
MIHAI APERGHIS: OK, one
more thing regarding AMP--you mentioned, so
again, newsy type stuff.What if I implement it
on an e-commerce websiteand do a static page
with product information?And there's maybe
a button that leadsto the actual website, maybe
Add-to-Cart or somethinglike that.I'm guessing this won't
show up right away,because you're not showing
it up or product-relatedor e-commerece-related queries--
JOHN MUELLER: Probably.
MIHAI APERGHIS: But further
down the line, maybe?
JOHN MUELLER: Yeah, maybe
further down the linethat's always possible.The other thing to
keep in mind with AMPis that it's not
just Google-specific.It's an open-source platform.It's also being used I think by
Twitter, Pinterest, LinkedIn,for example.So if your content is linked
to from any of those other pagecases, and you think that
your AMP version wouldbe a good match for
that, then maybe itdoes make sense
to show that evenfor things that aren't
shown in the Googlenewsy-type environment.So if you have something, for
example, that shared a loton social media that's
used a lot on Twitter,then maybe this
is something whereyou would see a significant
boom out of that as well.And the neat thing I find
about the AMP contentis that you can include
analytics in there as well.So you can see how many people
are previewing your contentthrough the AMP version.And you can make a judgment
call based on that over timewhere you say, well, maybe
I'll just do 10% of my contentlike this to start off with.I'll watch the analytics and
see is it really as somethingworthwhile for me or not.And then make a decision
based on that afterwards.
Yeah, sounds good.
JOHN MUELLER: But as
always, with a lotof these new
technologies, I thinkit's something that
SEOs in particularis really useful for
them to at least havesome experience with that.And another aspect there
is, of course as well,is if you're the first
people within your nicheto implement this
kind of thing, thenyou'll be the ones who will be
able to profit from that first.So obviously, there's
always a little bitof a risk involved there
as well that maybe youspend a lot of influencing
something that nobody actuallynotices.But I think as an
SEO in the beginning,that's something where you build
up a lot of experience as well.And where when clients
come to you and say,should I be doing AMP?You can say, well, yes or no.You can make a
clear judgment callbased on your experience
rather than just on, well,I read this one blog post.That's it.Maybe you should.And this the one site, maybe you
shouldn't, but I don't reallyknow.And if you don't
have any experience,then that's really hard
to give these peoplean educated piece of advice.
MIHAI APERGHIS: Yeah, thanks.
John, every yearyou recommend what
will happen in 2016or what webmasters should do.But where is Google
going in 2016, thisbeing the last Hangout for 2015?
JOHN MUELLER: Where
is Google going?I don't know.I hope they keep
the offices herebecause these are really
kind of comfortable.But past that, I don't know.I see a lot of talk about things
like having an assistant, whichI think is really interesting.I don't how that will work with
regards to websites in general,like if we have to
pick up special markupto make that work really well.But I think that's kind of a
rough very futuristic directionwhere things could be headed
where we see people usingsearch not just to get an
answer to keyword-type queries,but really to answer a specific
question that they have like,should I bring an umbrella
to the office tomorrow?Is it going to rain?What do I have to do next week?I want to go on
vacation in Canada.What should I watch out for?Where should I go?These type of things where
Google has a lot of informationwhere we can theoretically
combine that informationand really help guide
a user to do somethingthat matches what they're trying
to do but where at the moment,search falls short
because we're looking moreat the keywords in the
queries rather than tryingto understand the
actual intent of whatthese people are trying to do.So I kind of suspect
what will happento make that more
possible is that thingslike structured data markup
will become more relevant againwhere if we can pick up the
entities better on a page,understand what actions
could be done on a website,it'll be a lot easier for us
to actually guide people thereand say, well, you are looking
for information about Canada.Here is this really
great websitewith opportunities
in Canada since itlooks like you're not just
trying to go on vacation,but trying to move
there, for example.So kind of having this
structured data markupon these pages I suspect will
become more and more importantover time.
BARUCH LABUNSKI: Do
you have any ideawhy when you move
a site to HTTPS,you can't use the Data
Highlighter right away?
JOHN MUELLER: Because these
pages aren't cached yet.
BARUCH LABUNSKI: OK, so just
wait another couple weeks?
JOHN MUELLER: Yeah, so
the Data Highlighterrelies on the cached pages.And if you move from one
version of a URL to another,then we have to re-indexed
and recapture those pages.
MALE SPEAKER 5: John?
JOHN MUELLER: Yes.
MALE SPEAKER 5: I have a
question in the Q&A for youif you [INAUDIBLE].
JOHN MUELLER: Sure.
MALE SPEAKER 5: About
the multi-language site.
JOHN MUELLER: Let me
switch over again."Why does a multi-language
site take more timethan a single language
site to index in Googleand gain some search results?"That's a hard question.I think from a
technical point of view,one of the difficulties there
is we have to crawl and indexthese multiple URLs first.And we have to understand them
and kind of the connectionbetween these pages first.So if you have one page, then
that's a lot easier for usto understand than maybe
10 different pages thatare on different languages
or different country's focus.So that's something
where it will always,coming from a technical point of
view, take a little bit longer.From a practical
point of view as well,if you dilute your information
too much by putting iton to too many different
versions of pages,then that's a lot harder
for us to get signalsfor each of those
individual pages.Whereas if you have one really
strong page on this one topic,then that's something
we can collect signalsfor fairly quickly.But if you have 10 really
of good pages on onetopic in different languages,
then it's a lot harder for usto understand, well, this page
is really good on this topic.And it's available in
these different languages.So it's sometimes
mostly a matter of time,but also a matter of knowing
when to concentrate thingsone URL compared to splitting
things up onto multiple URLs.
MALE SPEAKER 5: OK, so
assuming that all signalsof both the meta
language are correct,I only need some patience.
JOHN MUELLER: I
would say patience.But maybe also take
a step back and thinkabout does it really
make sense to startwith all of these different
language versions.Or can I start with
maybe two or threedifferent language
versions in the beginningand expand from
there, because it'sa lot easier to have one
or a small set of pagesthat are really strong
compared to havinga lot of different
versions of the same pages.
MALE SPEAKER 5: OK, so
you should just firstto maybe begin with
two or three languagesand then in the
future, add some more.
JOHN MUELLER: Yeah.
MALE SPEAKER 5: Because
I was thinking--I was thinking to choose some
languages where the competitionmaybe is a little lower.
JOHN MUELLER: Sure.
MALE SPEAKER 5: And then call
plan the English language alsoand then in the
future, add some more.OK, thanks a lot.
JOHN MUELLER: That sounds good.
MALE SPEAKER 5: And
have a great New Year.
JOHN MUELLER: Thanks, you too.
ROBB YOUNG: John?
JOHN MUELLER: Yes.
ROBB YOUNG: Can I possibly
ask about our siteagain since it's our two
and a half year anniversary?Is there any, without giving
me any specifics because I knowyou can't-- is there any change
in that underlying issue?It is it something that's being
looked at, discussed, changed.Or is that now an
absolutely permanent changein an algorithm that is
never going to change again?Because when I spoke to you
a year ago or two years ago,you said it might.Or do you-- if you
think of all the thingsthat have changed over
the last two years,there's obviously a
lot of ranking factorsthat have changed.So is it something that might?
JOHN MUELLER: It might.These things might
always change.I mean you moved to your site
too like a different domainnow, right?
ROBB YOUNG: Yeah, we have.But because that
also hasn't reallychanged in ranking in the
last 12 months at all.So we've oscillated between
having some connection to itversus no connection
to it at all.And sometimes when we use
a hreflang to relate oneto the other because
at Christmas, weneed the seasonal-- it bizarrely
gives a seasonal boost.But then it looks like Google
catches us again and says,no, we know what you're doing.Come on.So then we have to
remove it again.But at Christmas,
that's all-importantbecause a conversion rate is
10 times what it is normally.So we almost have to use
it every 9, 10 monthsto say merely a short boost.But we definitely
want to distanceourselves the rest of the time.But it seems like we'll
never get away from it.But what we literally
can't survivewithout that once
a year association,because the natural
position for the other sitedoesn't change no matter how
many quality links we getor how much content
we put on it.It makes no difference.It just stays static,
JOHN MUELLER: The new site
is the one with an E, right?
ROBB YOUNG: Yeah, it is, yeah.
JOHN MUELLER: I really
don't see anythingholding that site back.
ROBB YOUNG: Even
though that's currentlygot the hreflang
from the old one.
JOHN MUELLER: I don't see
anything specific holdingthat back, no.
ROBB YOUNG: But the old one
is still in your special file.
JOHN MUELLER: That's
what I say, yes.But I see it going up and down.I guess that's the hreflang
things that we're doing now.
ROBB YOUNG: Yeah,
and also the Decembergoes up a lot and then
down after the 25thbecause we're a gift company.So we don't have a
quiet period then.
JOHN MUELLER: This is something
where, on the one hand,we try to do what makes sense
for search results there.And we do occasionally
take thingsto different parts of Google
to kind of look at say, hey,this is really obsolete.We need to either get rid of
it or update it or change itor whatever.And I think your site,
at least your old sitewith the way that
it's handled there,falls into that situation.And this is something that
from our point of view,we see extremely rarely, which
makes it really hard for usto push that appropriately
with the teams internallyand say, hey, this is
affecting a lot of sites.We need to change
this really quickly.It's kind of a tough
situation to be in, I guess.
ROBB YOUNG: I know.We've lost [INAUDIBLE]
of our revenue.And it's hard to not
take seriously and notask every time
I'm on these callsand try to contribute
to other stuff.But every so often, I like
to ask because we're notdoing anything--
you know we're notdoing anything inappropriate.
JOHN MUELLER: This is
something we do take upwith the other teams again.So it's good that
you keep askingbecause we do keep pushing
these things at our teams too.
ROBB YOUNG: But is
it really only us?Or is it something that
can affect other peoplebecause I do get other
people ask me, what's wrong?Because they want to
avoid the same issue.But I can't say, well, don't
do this that or this, evenwithin the Webmaster
Forum, because I'veposted in there before.But the problem is that everyone
jumps straight on it in sayingyou've got these four
backlinks , and no, it's not.It's' not that.
JOHN MUELLER: Yeah.
ROBB YOUNG: It's
vicious in there.
JOHN MUELLER: I think in the
last year, when I look back,I've seen maybe
one other site thatran into something like that.So it's something where
for the most part, whenwe look at the evaluations for
that specific change there,they're saying, well,
it's working as expected.It's doing the right thing.And it's really these kind
of cases where we say,well, it's not doing
the right thingwhere we tend to push back
on the teams as well and say,hey, we need to find a solution
here and figure out howto do this in a different way.
ROBB YOUNG: But is that only
one other site reportingthat issue, because,
obviously, Icould have just listened to the
advice in the forum and said,OK, you're right.It's that blog we wrote that
time with all these footerlinks.And then no one would've
raised it further up.So could there be
hundreds of sites?Or is there only one or two,
but you only hear about the one.
JOHN MUELLER: I don't know.It's hard to say, because
the type of algorithms,they're meant to affect
the web in a broad way.It's not.We wouldn't make like an
algorithm that would onlyaffect five handpicked
sites because then itwouldn't be an algorithm.Then it would be
something manual, right?
ROBB YOUNG: No, and I understand
from a user perspective,people are still
finding our competitors.So it's not affecting users.I'm realistic enough to know
that I have to try and maintainmy calm, but because
users find someone else.But there are only two
or three other sites.The frustration is when
you look at the resultsand a lot of those
sites are utter rubbish.But users don't care
because they still find.It's not like they
Google something.And there's 10 empty results.It doesn't happen.
JOHN MUELLER: I don't know.I'll pick it up again.
ROBB YOUNG: If you do, I'll
promise not to mention itagain until about March.
JOHN MUELLER: Why do you
keep mentioning this?No, I mean, these are the type
of things where sometimes wedo bring them up with the team.But if we never hear anything
back from either side,then we assume, well, maybe
it's not a big as a problemas we originally thought.Whereas if we keep hearing back
from the webmaster and saying,well, this still isn't fixed.This is still isn't working
the way I wanted to.Then that's something
where we'll continueto try to push internally.
ROBB YOUNG: All right, if you
could ask, I'd appreciate it.And I'll wait definitely
until the new year anyway.
BARUCH LABUNSKI: Any idea
regarding the search tools?You said you were
going to take it upwith the team-- the search
tools with the location thing.
JOHN MUELLER: Search tools.
BARUCH LABUNSKI: The
search tools, the location,to get a precise look
at how my site is doingin an area which is like--
JOHN MUELLER: Oh, OK, so
the advanced search settingsin the model search results.I haven't heard
anything back on that.I assume this is
just a normal UIchange where we decided
that this makes senseto keep it like that.
OK, so if I wantto find out like
a special hottestchocolates in Switzerland, and I
wanted to go into precise area,I won't be able to find it,
because it's going to give mea need to tap into somebody
else's computer in orderfor me to get that
precise location, even--
JOHN MUELLER: I don't know.If you want to find out the
best chocolates in Switzerland,you probably shouldn't just
rely on the search results.That's seems a bit--
ROBB YOUNG: You should
travel to Switzerlandis what you're saying.
JOHN MUELLER: Well,
or ask Swiss people.
BARUCH LABUNSKI: I'm just
saying the Search Tools--
JOHN MUELLER: People who
know about chocolate.
BARUCH LABUNSKI: It was
a really cool feature.And it's sad that it went away.That's all.There a lot of SEOs
are feeling the same.
JOHN MUELLER: But
on the other hand,we have to make
our search resultsusable by the general public.So this is something that
really only SEOs were using,then I don't know if it make
sense to keep supporting thatand keep updating
the data and makingsure the UI doesn't break.
ROBB YOUNG: If you all
know the best or somethingin any particular area, surely
you Google that instead.You don't Google what's the
best something in England.You Google what's the
best something in Londonand just keep going
down until you--isn't that normal behavior
versus using the Search Tools.
JOHN MUELLER: Yeah.
BARUCH LABUNSKI: Canada,
for instance, in an arealike Toronto, and
there's Mississauga,which is 20 kilometers
away, you'renot going to get the exact
precise-- every result isdifferent now because
of this change,because of removing
the location changebecause you said it's based
on where the user is located.But I said if I'm located
15 kilometers awayfrom the location, I'm going
to get something totallydifferent because of my phone.For instance, every day, if the
IP is different on my phone,it's a different
IP all the time.And the location changes.So I'm not physically in
that area is what I'm saying.
JOHN MUELLER: You could
use a GPS spoofer.You just boot your phone and
install one of spoofing appsand make your phone think
you're somewhere else.I don't know .That doesn't seem like something
that we would officiallyrecommend.But theoretically, you
could do that if youreally, really wanted to see
what the search results arelike there.
BARUCH LABUNSKI: OK.
MIHAI APERGHIS: You can
still use the parameterthat changes your location.
JOHN MUELLER: I think the
parameter still works, yeah.
BARUCH LABUNSKI: But
it's only Canada/US.It's like, come on.
JOHN MUELLER: I don't know.You can specify.You can do it on a city or
a regional level as well.
MIHAI APERGHIS: John, can I
ask regarding Robb's site,how come he cannot do anything
on his site to change thatwould not be-- so the site
wouldn't be affected by thatcertain algorithm, I guess?You mentioned that there's
nothing he can do about it.How come?How cannot he change his website
so that wouldn't be the problemanymore?
JOHN MUELLER: It's tricky.
MIHAI APERGHIS: I've always
been curious about it.
JOHN MUELLER: Yeah.At the moment, I really can't
go into more details there.
MIHAI APERGHIS: OK.
ROBB YOUNG: So
even if it somehowbecame as popular as Twitter,
despite the fact that wehad have a ton of
direct traffic,we still wouldn't rank?
JOHN MUELLER: I
think it would rank.A lot of these things, when
it comes to search results,they are not such that we'd
remove it directly completelyfrom search results.
ROBB YOUNG: It's funny though,
if you Google the exact name,we still come up on Wikipedia
entry and various press.So it's not our brand name
at all because that is found.But that's such a specific
query within a niche area.That's not how we make-- or
not how we used to make money.It was on all of the
generic in different terms.Anyway, I don't want to hog it.I am not Mihai.So I know you
can't say anything.So I'll move on.
JOHN MUELLER: I'll keep pushing.All right, somewhere there's
a question from someonewho can't speak due
to fever and coughing.Let me see if I can find that.Or maybe you can just paste
it into the chat again,and I'll pick it up from there.
ROBB YOUNG: While
that's going on,John, do you want to comment
on the internet of things?
JOHN MUELLER: The
internet of things.
ROBB YOUNG: Yeah,
and how Google willadapt to advertising when you're
trying to talk to your fridge.If everyone is not
actually sittingthere Googling anything,
what are you guysgoing to do to adapt to that.
JOHN MUELLER: When people
are going on their fridge.I don't know what that's--
ROBB YOUNG: Not even Google.They're talking to it instead.Asking I'm out of milk.Can you get me some milk?Would it narrow down your
opportunity to advertiseor would it broaden it.
JOHN MUELLER: I don't know.I could imagine if you're
a store selling milk,then maybe it will
be like, well,would you like to order
here, here, or here.
ROBB YOUNG: Right, that would
be a direct manufacturerrelationship if they do it
right rather than Googlehaving any place there
at all unless youprovide the technology.
JOHN MUELLER: I don't know.I think the whole internet
of things is on the one hand,there's a lot of talk about
it where it's not reallytalk about practical things
just yet, because thereare a lot of crazy
ideas out therethat might make sense
in the long run,but might not even
end up working at all.Where I don't know if
your light bulbs allhave individual IP addresses,
what does that really mean?Does that mean that
your light switchis going to have Wi-Fi built
in and turn things on and off?Or does that mean
that you're goingto have a website where the
light bulbs on your houseare going to be
displayed, and youcan display patterns and make
them blink in fancy ways?Probably not.You're probably just going
to want to turn the lightsor turn them off again.So there are lots of things
that might theoreticallybe possible.But if they'll actually
end up being used that way,it's really hard to say.But I could definitely
imagine thingslike refrigerators having
a bit more smarts built inas being something
that might make sense.But if that's something
that people will actuallyuse in the long run,
I don't really know.
BARUCH LABUNSKI: It's
probably just goingto be all recipe-related
or on the best queries.
JOHN MUELLER: Yeah, I mean
maybe that that makes sense.The tricky part there
is also that people buyrefrigerators for the long run.They buy them to keep
them for 20, 30 years.Whereas if you want
to buy a laptop,you're probably just going
to keep it for a year or two.So it'll become obsolete
after one or two years.And if you have your
refrigerator, that'sobsolete after one
or two years, Idon't know if that would really
make a great refrigeratorfor people, if that's really
something people wouldwant to buy as a refrigerator.
BARUCH LABUNSKI: Are you able
to answer anything regardingthe search to be forgotten?Anything related to that?How fast-- there was no
response at all from the teamafter submitting an application.But when does that
team get back?Or do you do not know?
JOHN MUELLER: Usually they
happens fairly quickly.
BARUCH LABUNSKI: OK,
so just submit it againand wait for an answer.
JOHN MUELLER: Yeah, it's
for people in Europe, right?
BARUCH LABUNSKI: Yeah, exactly.
JOHN MUELLER: I don't
know what the normal timeframes are there.Let me see.With regards to this site that
you just posted, for the movefrom HTTP to HTTPS, I don't
see any problems there.That looks like it's
about working as expected.All right, and since we're
in kind of drifting offinto all kinds of
crazy topics, letme just pop the normal
recording here now.And we can still chat
a bit for a while.And for those of you
watching on YouTube,I guess I'll see
you again next yearor in one of the
future Hangouts.I'll set them up again probably
when we get started againand try to keep them
at the right rhythm,the right frequency
of maybe onceevery two weeks as
we've had it so far.Until then, I wish you
all a great New Year,and I hope everything starts off
really well in the next year.Bye.
MALE SPEAKER 1: Yeah, Happy
New Year to you too, all.