Proper non-use of rel="prev" and rel="next" ?

[JS_Harris]

If the pages of a multi-page post are significantly different in that each could be a good landing page(different titles, descriptions and content) wouldn't it be a good idea to avoid using the rel=prev and rel=next tags on such a site? The purpose of the tags is to give Google a hint that they are a series and to consolidate results to the first page in most cases.

The concern is that this results in one page being returned for the keywords it ranks well for when, in theory, all of the pages could rank well for a wider range of keywords if treated separately. Would this be a scenario in which rel=prev and rel=next are actually harmful ?

[lucy24]

and to consolidate results to the first page

Where does that say?

[JS_Harris]

Where does that say?

[support.google.com...]

This markup provides a strong hint to Google that you would like us to treat these pages as a logical sequence, thus consolidating their linking properties and usually sending searchers to the first page.

As I said, the pages could stand alone and still be decent landing pages. if I don't want Google relating the pages or considering incoming links to internal sections as a link to the first section should I avoid rel=prev and rel=next in this situation? The page titles, descriptions and content are different(not wordpress default).

[lucy24]

Well, ###. Sorry for the digression-- but speaking strictly as a human user, if I search for some particular text, I want to be sent to the page that actually contains the text, not to page one of a twelve-page article that will eventually lead me to the text. (Same applies, of course, to text that only becomes visible if you execute some scripted doodahs-- but that was a different thread.)

:: sitting back waiting for someone knowledgeable to weigh in, because I'm now wondering about assorted ebooks that have each chapter as a separate file ::

Edit: At this point I did what I should have done in the first place: an exact-text search to verify that I get pointed to Chapter XV rather than Chapter I. I wonder if it makes a difference whether the linked segments differ in the path itself or only in part of the query string?

[JS_Harris]

Well, ###. Sorry for the digression-- but speaking strictly as a human user, if I search for some particular text, I want to be sent to the page that actually contains the text, not to page one of a twelve-page article that will eventually lead me to the text.

My point exactly, I don't want Google sending my visitors to the first page for this particular site. I took the time to set it up with a unique title for each page, show unique related but stand-alone worthy content. Think 10 different examples of a widget type, all are related sure but I discuss a different aspect of each, complete with different images and text, on different pages. I don't see why only the first should rank, any of the 10 should rank on their own merit. Some people might search for a green widget, others for a rainbow widget, I discuss a different aspect(not as boring as color) on each page and think they should land on the page best suited to their search.

I definitely wouldn't recommend ignoring the tags on a typical wordpress installation that shows the same title and meta description on every page but if set up as above... are the tags a bad idea?

[ergophobe]

sent to the page that actually contains the text,

Sure. In the age of dumb search engines, the text you type in to the search will be the text on the page. Advancements in search engines require that we get away from this, which is why if I type in "nordic downhill pointers" a good search engine might return a page with the text "telemark tips". And that's simple synonym substitution, which is the next dumbest thing to text matching. As you get more abstract in your matching, a good search engine will understand the subject of the article and serve that up regardless of any text or synonyms that match discrete keywords. But now it is I who digress...

Back to the original question - I think you have to ask yourself a simple question: why is this a multipage article?

1. Because it's just too much text to put on a single page and I think users will get click-happy if they don't have something to do other than read.
2. Because I'm selling CPM ads and I want to boost the number of impressions
3. Because it's a complex topic and I want to provide logical divisions between related but separate topics
4. Because it's a super long list, like the Google search results themselves, and if not paginated the load times and server load to return all that would be attrocious.

So, if it's #4 or #1, it makes absolute sense to have a rel/prev and that's really it's point and I believe it is these type of pages Google is aiming at consolidating. If it's #1, it begs the question of whether it really makes sense to paginate at all.

If it's #1, please direct me to your neck so I can ring it. These slideshow "pages" are the scourge of the internet right now. Better for all of us if you don't rank for any of the pages.

If it's #3, which I think it is, I don't think you'll have a problem. You can create a linked series and with or without the prev/next, I think each page will stand fine on its own and users will land on the page that most closely corresponds to their query, provided the search engine understand it and your page clearly. Anything you can do on-page to help that (titles, schema markup, etc), will help getting the right users to the right pages. Sadly gone are the days when you could see which queries were sending them to which pages and adjust and hone accordingly.

[lucy24]

In the age of dumb search engines, the text you type in to the search will be the text on the page. Advancements in search engines require that we get away from this

I don't think that changes my point, though. If I click on a search result, I expect to be taken to the page that actually contains the promised Telemark Tips-- not page one of a seven-page article, in which the Telemark info is buried on page five.

[ergophobe]

No, it does change the point.

- if you type in text that is missing from all pages in the article but is deemed to be on-topic for page seven you should and normally will be brought to page seven in most cases where the article author doesn't do some crap slideshow pagination that Google can't follow

- if you type in text that is missing from all pages in the article but is deemed to be on-topic of the general subject on which the overall article is written you should and in most cases will be brought to page one.

- if you type in text that is found on one and only one page as exact match, you will be brought to that page.

My point was, however, that ultimately what you want is not a typically a match for your text, but an answer to your question. I do tons of linguistic searches (medieval and early-modern usages) and in my case the intent of my question is to see the word in context, which means that a good search engine would take me to the location of that text for those queries. But in many case, the intent of the question can indicate that the answer is not best provided by the page with an exact match keyword.

Think of real conversation. Somebody asks you a deep question that requires background and setup. You don't typically start your answer with the part of you answer that has an exact match for some of the words in the question. That isn't typically what people want when they ask a question. Search engines still function this way because of their limitations.

[lucy24]

it does change the point.

No, it doesn't. (Does! Doesn't! Does! Doesn't!) How can it, when your expanded answer says exactly the same thing I said in the first place? The search engine should take you to the specific page that answers your question, not to page one of an article that will answer the question on some subsequent page.

I do tons of linguistic searches (medieval and early-modern usages) and in my case the intent of my question is to see the word in context, which means that a good search engine would take me to the location of that text for those queries.

How weird. To me that sounds like an absolutely textbook case of a situation where you would never, ever want the search engine to guess or use synonyms.* ("When I said 'ax' I most emphatically did not mean 'ask', since that was the whole point of my inquiry.")


* "When you said 'umiarjuap', did you mean 'marijuana'? When you said 'umiarjuup', did you mean 'markup'?" I did not make those up. In these specific case the search engine didn't have the arrogance to assume I did mean B when I said A, but this seems to be entirely at their whim.

[ergophobe]

Sorry, but no, there is all the difference in the world and that is my point.

What you assert you said in the first place:

The search engine should take you to the specific page that answers your question

What you actually said in the first place

if I search for some particular text, I want to be sent to the page that actually contains the text

And as for the linguistic searches, you are still not seeing my point. And your "how weird" comment takes my meaning to be the exact opposite of the plain sense of the words you quote. My point in the bit that you quote is that search intent matters. As I said, in the case of a linguistic search, I *DO* want the exact text at the exact location, no synonyms, nothing clever. In the case of other searches, I want an answer to my question. Currently search engines are not smart enough to consistently see the difference, but they are getting better.

Anyway, all of that relates to JS_Harris question how? In my experience, you will typically get taken to the page that most closely corresponds to an answer to your query. In some cases that will be the page seven you want and in some cases it will be the root page of the article.

So back to the original point - if you are talking about an article series, I would not ever expect Google or Bing to take me to page one for a query that relates to content on the third article in the series, regardless of prev/next markup. If, however, you're talking about a long article that can be split across various pages or not (in say a "printer friendly" view), I would expect to be taken to the head of the article. In my experience, though, the overwhelming default is to take you to the middle of the article even if it isn't the place a human teacher would start you in order to answer your query.

[JS_Harris]

So back to the original point - if you are talking about an article series, I would not ever expect Google or Bing to take me to page one for a query that relates to content on the third article in the series, regardless of prev/next markup.

OK so using the rel=prev and rel=next tags on the type of multi-page articles I have is a bad idea. I sincerely don't want Google to aggregate anything back to the first page at all, which is why I worked out how to title each section differently. Thanks.

[ergophobe]

I don't know if it's a bad idea, but in those cases, I have not seen it result in devaluing the lead article. I've seen a middle article rank the highest, because the first article is introductory fluff and doesn't get a lot of link love.

Please note, though, I'm speaking from my experience as a web user and my reading of what Google and Bing say in their documents, not from before/after experience with adding prev/next markup to a page and seeing what happens.

In any case, the examples Google and Bing give for paginated content where prev/next are appropriate are things like
- forum threads split across multiple pages
- search results
- reverse chronological list of blog posts or news stories

Exceptions are
- pages with arbitrary breaks that have a "View full page" option available. In that case you typically want the full page version to be your canonical (according to Google, though obviously Buzzfeed wants you to view it as a slideshow)

- [support.google.com...]
- [blogs.bing.com...]
- [googlewebmastercentral.blogspot.com...]

In any case, in the examples I mentioned, I think prev/next makes sense and would probably help your user (do you really want to send them to the third page of the listing of blog posts? It seems like they are going to want either one of the posts referenced on the that third page or the first page of results).

My point in the long diversion, though, was that if you have general search terms that indicate that the article/series as a whole is relevant, but based on the general topic, that's when you would want to take the user to the top page of the article and that's where I would expect (note "expect" - this is a guess, not knowledge) the prev/next markup to start to come into play.

Really, though, I think the key is good in-article navigation and generally, Bing and Google (again based on the articles linked above) try to discern how pages fit into a series (based I would *guess* on things like position in recurring navigation blocks, use of Part 1, Part 2 in titles and nav, etc).