English - Pagination and SEO

Google's Maile Ohye examines the issues webmasters face with paginated content: paginated articles, product category listings, etc. She then explains the options available to those webmasters concerned about SEO, including rel="next" and rel="prev" HTML markup.

12 Languages

Edit Subtitles

Download

Showing Revision 1 created 08/16/2012 by Amara Bot.

Title:
Pagination and SEO
Description:

Google's Maile Ohye examines the issues webmasters face with paginated content: paginated articles, product category listings, etc. She then explains the options available to those webmasters concerned about SEO, including rel="next" and rel="prev" HTML markup.
0 0:00 - 0:03
3020 0:03 - 0:04

MAILE OHYE: Hi.
3830 0:04 - 0:05

I'm Maile Ohye.
5250 0:05 - 0:09

I've been at Google now for over
six years, working with
8580 0:09 - 0:11

Search and with Webmaster
Tools.
11360 0:11 - 0:13

I'd like to welcome
you to my home.
13490 0:13 - 0:17

Let's chat about pagination
and SEO.
16860 0:17 - 0:19

For today's agenda, we'll
first start with some
19280 0:19 - 0:22

paginated content examples.
21870 0:22 - 0:25

Then we'll get into some of the
negative side effects of
24800 0:25 - 0:29

pagination and why you as a
webmaster might want to make
28520 0:29 - 0:32

some effort as to not dilute
your indexing properties and
32270 0:32 - 0:35

to show better results
to users.
35070 0:35 - 0:37

Then we'll cover your
configuration.
37460 0:37 - 0:39

And this comes in two parts--
39310 0:39 - 0:43

for those of you webmasters with
paginated content and a
42540 0:43 - 0:45

view-all page available, and
then for those of you
45100 0:45 - 0:48

webmasters that have paginated
content but
47860 0:48 - 0:50

without a view-all page.
49770 0:50 - 0:53

So there's going to be two types
of configurations there.
52910 0:53 - 0:55

Then we're going to step back
a little bit and talk about
55430 0:55 - 0:59

what Google is doing to help
users with paginated content
58910 0:59 - 1:01

and webmasters as well.
61100 1:01 - 1:04

And then last, given your
configuration, whether you
64450 1:04 - 1:08

have a view-all page available
or you have no view-all page
67680 1:08 - 1:11

available, we'll look at the
options that you have for your
70800 1:11 - 1:13

paginated content.
72770 1:13 - 1:15

So let's go ahead and
start with some
74580 1:15 - 1:18

paginated content examples.
77950 1:18 - 1:21

Paginated content exists
throughout the web, and I'm
81030 1:21 - 1:23

going to cover two of
those common cases.
83410 1:23 - 1:26

One is a paginated article.
86210 1:26 - 1:29

So let's say you go to your
favorite content site, and you
89460 1:29 - 1:32

see the breaking news story.
91780 1:32 - 1:36

"New studies prove that cookies
are superior nutrition
95510 1:36 - 1:39

to vegetables." And that would
be quite the story.
99140 1:39 - 1:43

But your favorite site might not
put this all on one page,
102850 1:43 - 1:47

but instead, paginate it into
several component pages.
106700 1:47 - 1:50

Now this one article has become
three, and this is an
109690 1:50 - 1:53

example of paginated
content articles.
112920 1:53 - 1:57

Another example of pagination
is for things like a product
116630 1:57 - 1:59

category, like what you would
see on your favorite
119270 1:59 - 2:01

e-commerce site.
120780 2:01 - 2:04

So let's say this webmaster
is selling shapes.
123990 2:04 - 2:06

They're selling six
types of shapes.
125870 2:06 - 2:09

But rather than have it all on
one page, they have divided it
129360 2:09 - 2:14

into two component pages, both
of them with shapes, creating
133680 2:14 - 2:15

pagination again.
135270 2:15 - 2:19

So two common ways are with
paginated content articles and
139030 2:19 - 2:22

with paginated product
categories.
141930 2:22 - 2:25

Now, what are some of the
negative side effects of this?
145070 2:25 - 2:27

Well, there's a couple.
146890 2:27 - 2:32

So I'd like to highlight two,
the first being that indexing
151790 2:32 - 2:36

properties, like links and
anchor text, can be diluted
156340 2:36 - 2:40

into the different component
URLs rather than being
159660 2:40 - 2:42

consolidated to the
one article or to
162340 2:42 - 2:45

the one product category.
164520 2:45 - 2:47

So that's one of the negative
side effects.
167200 2:47 - 2:51

The other is that the most
relevant page in the series
170830 2:51 - 2:54

might not be reflected
in search results.
173530 2:54 - 2:57

So if you're the webmaster for
this e-commerce site, you
176550 2:57 - 3:01

might want users to be sent to
page one, say, of your series.
180630 3:01 - 3:04

But because search engines see
this pagination as three
183870 3:04 - 3:07

separate entities, searchers
might be sent to a different
186940 3:07 - 3:10

page that might not be
the most relevant.
190460 3:10 - 3:13

So those are a few of the
negative side effects of
193260 3:13 - 3:14

pagination.
194180 3:14 - 3:17

Now let's talk more about
your situation and the
196930 3:17 - 3:20

configuration you have
on your site.
199560 3:20 - 3:22

We're going to look at this in
terms of two different types
202080 3:22 - 3:23

configurations.
203300 3:23 - 3:27

One is with a view-all page
available, and the other is
206670 3:27 - 3:29

with no view-all
page available.
209190 3:29 - 3:32

Now, if your site has paginated
content with a
211860 3:32 - 3:35

view-all page available, there
are a couple of things you
214810 3:35 - 3:37

want to make sure
you test for.
216730 3:37 - 3:41

One is make sure that you have
still decent latency on your
221430 3:41 - 3:45

site meaning that, if a user
clicks on the view-all
224850 3:45 - 3:48

version, that it doesn't take
them 15 seconds to load
228250 3:48 - 3:50

because it's such a
long article, or
229810 3:50 - 3:51

it's so many products.
230970 3:51 - 3:53

But that they still have
a good experience--
233340 3:53 - 3:56

say, the page only takes
four seconds to load.
236480 3:56 - 3:59

The second thing to check for
if you have a view-all page
238750 3:59 - 4:03

available is to make sure that
the page remains easily
242540 4:03 - 4:06

navigable, meaning that users
can still find the content
246480 4:06 - 4:09

that they want or the particular
product that they
248790 4:09 - 4:13

want by easily scrolling
or viewing headings.
252620 4:13 - 4:16

So that's the configuration of
a view-all page available.
256100 4:16 - 4:19

And then obviously, without a
view-all page available, it's
259310 4:19 - 4:21

fairly straightforward.
261240 4:21 - 4:25

So think about your site in
terms of configuration you
264690 4:25 - 4:29

have. But before we go there,
let's take a step back and
268900 4:29 - 4:32

talk a little bit about
what Google is doing.
272230 4:32 - 4:35

We're, of course, always
working to improve the
274800 4:35 - 4:37

experience for searchers.
276540 4:37 - 4:40

And one thing that we found
through testing is that our
279850 4:40 - 4:43

searchers prefer seeing the
view-all page in their search
283480 4:43 - 4:48

results opposed to an individual
component page.
287760 4:48 - 4:51

And one reason for this might
be because of latency.
291220 4:51 - 4:54

So if you take search results
and you click on a result to a
294450 4:54 - 4:58

view-all page, while that might
take, say, three seconds
297810 4:58 - 5:01

to load that article that new
studies prove that cookies are
301100 5:01 - 5:04

superior nutrition to
vegetables, that might be
303920 5:04 - 5:05

three seconds.
304840 5:05 - 5:08

But on the other hand, searchers
were less happy when
308070 5:08 - 5:12

search results took them to just
page one of the article.
311580 5:12 - 5:14

While that might have just had
two seconds of latency and
314205 5:14 - 5:17

then the page loaded, every time
that user wanted to click
317390 5:17 - 5:20

Next to read more of the
article, it caused some
320150 5:20 - 5:22

additional load time.
321900 5:22 - 5:26

So because of this latency and
other reasons, searches prefer
326000 5:26 - 5:28

the view-all page.
328160 5:28 - 5:32

So given this knowledge, one of
our engineers on indexing,
331730 5:32 - 5:35

Benjia Li, actually came
out with a new feature
334700 5:35 - 5:37

in October of 2011.
337230 5:37 - 5:39

This is--
339370 5:39 - 5:43

"When we detect that a paginated
series also contains
343030 5:43 - 5:47

a view-all version, we're now
making a larger effort to
346630 5:47 - 5:49

return the view-all page
in search results when
349450 5:49 - 5:53

appropriate." So that's
great for searchers.
353010 5:53 - 5:56

And what's even better for
webmasters is that while we
355780 5:56 - 5:59

detect this view-all page, we'll
also still consolidate
359180 5:59 - 6:03

indexing properties, like links,
to the view-all page.
363490 6:03 - 6:06

So again, this is good for
searchers and good for you as
366460 6:06 - 6:10

webmasters for all that indexing
consolidation.
370030 6:10 - 6:13

Now, let's talk about some of
the options that you have as a
373090 6:13 - 6:16

webmaster with paginated
content.
375620 6:16 - 6:19

We're first going to look at the
situation where webmasters
379290 6:19 - 6:23

have paginated the content and
a view-all page available.
382730 6:23 - 6:26

But for those of you that have
no view-all page available,
385840 6:26 - 6:28

it's still good if you pay
attention because some of
387800 6:28 - 6:30

these options will apply
to you as well.
390240 6:30 - 6:34

So you have a site with
paginated content and a
393600 6:34 - 6:37

view-all page, you have
three good options.
397080 6:37 - 6:40

First, you can leave as is.
400100 6:40 - 6:42

There's nothing that you have
to do if you have other
402190 6:42 - 6:44

priorities on your site.
404340 6:44 - 6:47

Paginated content exists
throughout the web, and search
407150 6:47 - 6:49

engines will continue
to do an even better
409320 6:49 - 6:51

job of handling it.
410820 6:51 - 6:54

And as I mentioned earlier, if
you have a view-all page
413760 6:54 - 6:57

available, Google will
automatically try to detect
416870 6:57 - 7:00

that, send searchers there, as
well as consolidate your
420000 7:00 - 7:01

indexing properties.
421430 7:01 - 7:04

So option one is a very
solid option.
424320 7:04 - 7:07

But you also have
a second option.
426860 7:07 - 7:11

The second option is to actually
use rel="canonical"
430600 7:11 - 7:15

to explicitly hint to Google
what is your view-all page.
435070 7:15 - 7:19

So while we try to detect it
algorithmically, you can also
438880 7:19 - 7:22

tell us by writing
rel="canonical" on your
441860 7:22 - 7:25

component pages to your
view-all version.
445150 7:25 - 7:28

And this is kind of a more
explicit hint to us about how
448300 7:28 - 7:30

your site is configured.
450220 7:30 - 7:33

With the rel="canonical," as
many of you already know,
453130 7:33 - 7:36

we'll of course consolidate the
indexing properties from
456180 7:36 - 7:40

the component pages with
the canonical version.
460060 7:40 - 7:43

So things like links will
also be transferred.
462910 7:43 - 7:45

And then, of course,
we'll send users to
465000 7:45 - 7:47

the view-all page.
466850 7:47 - 7:48

So that's option number two.
468460 7:48 - 7:52

The last option is actually done
by two of our engineers,
472430 7:52 - 7:54

Joachim and Benjia.
473950 7:54 - 7:58

And what this is is using the
standard HTML markup of
478480 7:58 - 8:02

rel="next" and rel="prev" on
the component pages in your
482110 8:02 - 8:06

series to signal to Google
that these are individual
485630 8:06 - 8:09

pages, but they all belong
to one series.
489180 8:09 - 8:12

So by adding this rel="next"
and rel="prev" markup, you
492000 8:12 - 8:16

connect these individual
components into one.
495650 8:16 - 8:19

You can do this by adding
rel="next" to page one and
499490 8:19 - 8:24

then rel="prev" and rel="next"
to page two, all the way to
503710 8:24 - 8:27

the last page, which only
includes a rel="prev".
507010 8:27 - 8:28

And then, of course,
on your view-all
508270 8:28 - 8:31

page, nothing is needed.
510830 8:31 - 8:34

rel="next" and rel="prev" is
standard HTML markup, and it's
514380 8:34 - 8:36

been around for years.
515780 8:36 - 8:39

But now, Google is using this
markup for webmasters to let
519210 8:39 - 8:42

us know about their
paginated content.
522309 8:42 - 8:44

So let me explain some
ways that rel="next"
524059 8:44 - 8:46

and rel="prev" work.
526160 8:46 - 8:49

With rel="next" and rel="prev,"
much like you see
529120 8:49 - 8:51

with something like
rel="canonical," we'll
531470 8:51 - 8:55

actually consolidate indexing
properties from the component
534730 8:55 - 8:57

pages of the series.
536720 8:57 - 9:00

And in addition, unlike
rel="canonical" that only
539790 9:00 - 9:02

shows the view-all page in
search results, with
542370 9:02 - 9:05

rel="next" and rel="prev," we're
going to override that
545410 9:05 - 9:09

behavior and send users
to only one of
548560 9:09 - 9:10

the component pages.
550040 9:10 - 9:13

Most likely, this will be page
one, because commonly that's
552910 9:13 - 9:15

the most relevant page.
554670 9:15 - 9:18

So now if you have, say, that
product category, selling
557650 9:18 - 9:21

shapes, if you use the
rel="next" and rel="prev"
561460 9:21 - 9:25

markup, it'll tell us
that these two pages
564640 9:25 - 9:26

belong to one series.
566350 9:26 - 9:30

And then most commonly, we'll
send users to page one.
570320 9:30 - 9:34

Know that rel="next" and
rel="prev" is a strong hint.
574140 9:34 - 9:37

It's not a mandate
by any means.
577220 9:37 - 9:40

The last thing I want to say
about rel="next" and
579590 9:40 - 9:44

rel="prev" is that component
URLs in a series should be
583900 9:44 - 9:46

consistent with their
parameters.
586360 9:46 - 9:51

So let's take the article of new
studies prove that cookies
591280 9:51 - 9:54

are superior nutrition
to vegetables.
593670 9:54 - 9:58

Now, let's say that these pages
contain a session ID.
597860 9:58 - 10:01

All of these values for
rel="prev" and rel="next"
600570 10:01 - 10:03

should also contain
the session ID.
603210 10:03 - 10:06

And this is because our
indexing team looks to
605610 10:06 - 10:09

actually link every page in a
series with what was declared
609220 10:09 - 10:12

previous and what was
declared next.
611550 10:12 - 10:14

And when they do that, they want
to make sure-- say you're
614290 10:14 - 10:16

on page two--
615650 10:16 - 10:19

that the rel="prev" that
states rel="prev" is
618650 10:19 - 10:23

page-1&sid=123, they will
go to that URL.
623380 10:23 - 10:27

But that URL actually
has to list page two
626530 10:27 - 10:28

with the same sid.
628370 10:28 - 10:32

And that's how we can link every
page in the sequence.
631730 10:32 - 10:36

So be sure to keep parameters
throughout your entire series.
636380 10:36 - 10:39

So let's recap those
three options.
639140 10:39 - 10:41

If you have a view-all
page available, you
641280 10:41 - 10:43

can leave as is.
642920 10:43 - 10:46

You could also explicitly state
rel="canonical" to your
646220 10:46 - 10:48

view-all page.
647630 10:48 - 10:51

Or you can override the view-all
page behavior by
651400 10:51 - 10:55

adding rel="next" and
rel="prev." By adding
654530 10:55 - 10:58

rel="next" and rel="prev," you
will help us consolidate
657680 10:58 - 11:00

component pages in a series.
659670 11:00 - 11:03

But instead of sending users to
a view-all page, we'll then
663030 11:03 - 11:06

send searchers to one component
page., most likely
666320 11:06 - 11:09

page one of your series.
668990 11:09 - 11:12

Now, let's talk about the
configuration with no view-all
672170 11:12 - 11:14

page available.
673890 11:14 - 11:17

So for those of you webmasters
that have paginated content
677250 11:17 - 11:21

and no view-all page, you
have two options.
680620 11:21 - 11:23

First, of course, you
can leave as is.
682980 11:23 - 11:25

That's perfectly fine.
684600 11:25 - 11:28

And then your second option is
also to use rel="next" and
687650 11:28 - 11:31

rel="prev." Again, by using
rel="next" and rel="prev," it
691100 11:31 - 11:34

connects the component pages
in the series, and
693750 11:34 - 11:37

consolidates indexing
properties, and helps us to
696890 11:37 - 11:40

send searchers to the most
relevant page, which is likely
699800 11:40 - 11:43

the first page of the series.
702890 11:43 - 11:46

Now I'm going to beat you to the
punch and ask one of the
705890 11:46 - 11:50

most commonly asked questions
about rel="canonical" as well
709920 11:50 - 11:54

as rel="next," "prev." And that
is why rel="next" and
714160 11:54 - 11:57

rel="prev" for a paginated
series rather than
717020 11:57 - 11:59

rel="canonical" to page one?
719382 11:59 - 12:00

Ha!
719720 12:00 - 12:01

I bet you were thinking that.
721440 12:01 - 12:05

The answer is that
rel="canonical" is for
724670 12:05 - 12:06

duplicate content.
726350 12:06 - 12:08

So let's take that article.
728210 12:08 - 12:10

Let's say page two of
the article, cookies
730120 12:10 - 12:12

are superior nutrition.
731940 12:12 - 12:15

If this page actually has a
session ID attached, then it
735440 12:15 - 12:19

can list as the canonical the
same version, the duplicate
738730 12:19 - 12:22

conversion, but without
a session ID Because
741860 12:22 - 12:25

rel="canonical" is for duplicate
content, or it's for
745250 12:25 - 12:27

content which is a superset.
747470 12:27 - 12:32

So here we have page one, page
two, and page three, all
751940 12:32 - 12:34

linking to the canonical
version being
754160 12:34 - 12:36

the view-all version.
755860 12:36 - 12:38

And that's perfectly
fine as well.
758170 12:38 - 12:40

The thing about rel="canonical"
is that it
760190 12:40 - 12:44

only indexes content from
the canonical version.
764180 12:44 - 12:47

So let's go ahead and
take a look at this.
766870 12:47 - 12:51

If we have page two and page
three, page two says "cookies
771310 12:51 - 12:54

are superior nutrition,"
and page three says "to
773960 12:54 - 12:55

vegetables".
775170 12:55 - 12:57

But they both add
rel="canonical"
776910 12:57 - 12:59

just to page one.
779070 12:59 - 13:03

And Google's index will then
cluster page one, page two,
782620 13:03 - 13:05

and page three all together.
785250 13:05 - 13:08

But the only thing that we'll
have indexed is the content
788220 13:08 - 13:11

from page one, the canonical
version.
790710 13:11 - 13:14

So our index will actually
contain "new studies prove
793990 13:14 - 13:15

that."
794982 13:15 - 13:18

And now by using this
rel="canonical" incorrectly,
798100 13:18 - 13:21

this webmaster has totally lost
the content "cookies are
801190 13:21 - 13:25

superior nutrition" and "to
vegetables." So that's why
804590 13:25 - 13:27

rel="canonical" doesn't
work in this case.
806970 13:27 - 13:30

But rel="next," "prev"
works for a series or
810380 13:30 - 13:32

a sequence of content.
812060 13:32 - 13:36

So let's take those two
paginated examples again.
815970 13:36 - 13:39

By using rel="next" and
rel="prev," we'll actually, in
819450 13:39 - 13:43

Google's index, mark
it as a series.
822550 13:43 - 13:47

But we'll have page one, page
two, and page three all
826510 13:47 - 13:48

indexed separately.
828120 13:48 - 13:52

So in our index, we know page
one refers to "new studies
831670 13:52 - 13:54

prove that," page two,
"cookies are superior
834360 13:54 - 13:58

nutrition," and page three, "to
vegetables." And all three
838170 13:58 - 14:02

pages will be indexed and
marked as one series.
841580 14:02 - 14:04

So that's the big difference
between rel="canonical" and
844000 14:04 - 14:05

rel="next" "prev."
845140 14:05 - 14:08

So something to note is that
rel="canonical" can actually
848500 14:08 - 14:13

be used alongside rel="next"
"prev." So let's take a look
852960 14:13 - 14:14

at page two again.
854240 14:14 - 14:17

And this time, it has
a session ID.
856860 14:17 - 14:21

This URL can actually list both
the canonical version
860550 14:21 - 14:25

without a session ID as well as
a rel="prev" and rel="next"
865190 14:25 - 14:28

with, of course, the same
parameters, including that
868370 14:28 - 14:30

session ID.
869690 14:30 - 14:34

So now let's recap your new
pagination toolbox.
873870 14:34 - 14:37

Starting with Google, we have
two new features for you.
877160 14:37 - 14:40

First, we're making a better
effort to detect a view-all
880480 14:40 - 14:43

page, and then send
searchers to that
882540 14:43 - 14:45

preferred view-all version.
885050 14:45 - 14:48

The second feature is if you
want to actually even override
887870 14:48 - 14:49

that behavior.
888920 14:49 - 14:52

So for those of you with a
view-all page available or
891560 14:52 - 14:55

without, if you add markup
with rel="next" and
895110 14:55 - 14:58

rel="prev," it signals to
Google that these are
898020 14:58 - 15:00

component pages in a series.
900120 15:00 - 15:03

We'll then consolidate indexing
properties, and send
903100 15:03 - 15:08

searchers to the most relevant
page, most likely page one.
907790 15:08 - 15:10

Now, let's get into
the types of
909910 15:10 - 15:12

configurations you have available.
912090 15:12 - 15:16

So recapping, if you have a
view-all page available, you
915680 15:16 - 15:17

have three options.
917040 15:17 - 15:19

You can leave as is.
918890 15:19 - 15:23

You can use rel="canonical" on
your component pages, pointing
922710 15:23 - 15:25

to your view-all page.
924590 15:25 - 15:29

Or you can override all the
view-all detection by adding
928690 15:29 - 15:32

rel="next" and rel="prev,"
telling us that these
931830 15:32 - 15:35

component pages belong
to a series.
934530 15:35 - 15:37

And I'd like you, Google, to
send searchers to the most
937260 15:37 - 15:42

relevant individual page,
again, likely page one.
941500 15:42 - 15:45

Now, the other part of the
pagination toolbox is for
945050 15:45 - 15:48

those of you with no view-all
available, and
948220 15:48 - 15:50

you have two options.
949740 15:50 - 15:53

Of course, you can leave
exactly as is.
952540 15:53 - 15:56

Or again, you can use rel="next"
and rel="prev."
955900 15:56 - 15:59

This helps you to consolidate
all the component pages into
959430 15:59 - 16:04

one series and send searchers
to the most relevant page.
963590 16:04 - 16:07

So the great thing about these
pagination features is that
967390 16:07 - 16:10

I've been at Google long enough
to see the infancy from
970310 16:10 - 16:13

when the webmaster community was
talking to us about issues
973130 16:13 - 16:16

with pagination until now when
we have more features
976090 16:16 - 16:18

available to you.
977790 16:18 - 16:21

So thank you so much to all of
you for your helpful feedback
980720 16:21 - 16:23

and for being part of this
webmaster community.
983370 16:23 - 16:26

For more information on
pagination, here are some
985900 16:26 - 16:27

links available.
986990 16:27 - 16:29

And you can, of course, join
us at the Webmaster Central
989390 16:29 - 16:32

Blog or in the Webmaster
Discussion Forum.
992310 16:32 - 16:34

Thanks for your time.
993560 16:34 -

Add a new language!

This video is in English

English subtitles

← Pagination and SEO

English subtitles

← Pagination and SEO

Embed video

Share