8

While this is a question, it is as well a serious warning to publishers that separate their posts using <!--nextpage-->.

Bare with me. I run a highly trafficked multiple author platform that rank very high on search engines. While making SEO related improvements, I noticed that only the first page of an article that is separated into several pages is being indexed by Google. How did I notice this?

Because someone grabbed the content of the unindexed pages of an article and posted it in their own site. This resulted in them ranking higher than me on that particular keyword. Actually, let me rephrase that, my site's article is not even indexed beyond the first page.

My question is, how can I make the search engines index the whole article if it is separated into several pages? Ideally, a script that displays the full article if the visitor is a search engine (if that would not be harmful for SEO). Other suggestions are welcome.

On a side-note, if I was into Black Hat techniques, there would be a goldmine of content out there that I could exploit. Think about it.

And on a personal note, this issue should be addressed in a future update of Wordpress.

Update: A similar question that detailed how Wordpress creates the same canonical URL for all pages in a paginated sequence was asked here, however the answer that was posted and marked as correct does not answer the question. That solution works for paged comments only, not paginated posts.

Update 2: According to this blog post by Google, we can use rel=”next” and rel=”prev” to indicate the relation between pages. So for instance, this is what we should include in <head> on page 2:

<link rel="canonical" href="http://domain.com/article/2" />
<link rel="prev" href="http://domain.com/article/1" />
<link rel="next" href="http://domain.com/article/3" />

And there should not be a rel="prev" or "next" if there isn't a page after/before it.

Improve this question
3
  • Seems like you've found the answer. Care to format it as such and accepting it? I think a lot of people will benefit from this.
    – Tomas Buteler
    Commented Feb 21, 2013 at 14:45
  • @tbuteler While I found the solution which is to add a unique Canonical URL on each paginated page and to include the link rel="prev"/"next" whenever necessary, I do not have a script for it written. However, toscho mentioned in the chat that he may have something. Watch this space for updates...
    – Christine Cooper ♦
    Commented Feb 21, 2013 at 15:26
  • I've updated the code in my answer on the question you referred to. I have tested it, and it solves your problem. Good work on finding about this flaw!
    – shea
    Commented Feb 24, 2013 at 4:37

2 Answers 2

Reset to default
2

The basic problem for a script solution is: rel_canonical does not offer any useful filter. So we have to replace that function:

remove_action( 'wp_head', 'rel_canonical' );
add_action( 'wp_head', 't5_canonical_subpages' );

The next problem: $GLOBALS['numpages'] is empty before setup_postdata(). We could call that function already here, but it might have side-effects.

Here is a solution that adds correct prev/next links too, and it prevents conflicting relations caused by adjacent_posts_rel_link_wp_head. We have too hook into wp_head not later than priority 9 to deactivate the latter hook.

remove_action( 'wp_head', 'rel_canonical' );
add_action(    'wp_head', 't5_canonical_subpages', 9 );

/**
 * Extend version of the native function rel_canonical()
 *
 * @wp-hook wp_head
 * @return  void
 */
function t5_canonical_subpages()
{
    if ( ! is_singular() )
        return;

    if ( ! $id = $GLOBALS['wp_the_query']->get_queried_object_id() )
        return;

    $post = get_post( $id );
    setup_postdata( $post );

    # let WordPress do all the work
    if ( empty ( $GLOBALS['page'] ) )
        return rel_canonical();

    $permalink = get_permalink( $id );
    $canonical = t5_page_permalink( $permalink, $GLOBALS['page'] );
    echo "<link rel='canonical' href='$canonical' />";

    # next and prev links
    if ( 1 < $GLOBALS['page'] )
    {
        $prev = t5_page_permalink( $permalink, $GLOBALS['page'] - 1 );
        print "<link rel='prev' href='$prev' />";
    }

    if ( isset ( $GLOBALS['numpages'] ) && $GLOBALS['page'] < $GLOBALS['numpages'] )
    {
        $next = t5_page_permalink( $permalink, $GLOBALS['page'] + 1 );
        print "<link rel='next' href='$next' />";
    }

    # avoid conflicting pev/next links
    remove_action( 'wp_head', 'adjacent_posts_rel_link_wp_head' );
}

/**
 * Helper to get correct permalinks for sub-pages.
 *
 * @param  string $permalink
 * @param  int    $page
 * @return string
 */
function t5_page_permalink( $permalink, $page )
{
    if ( 1 == $page )
        return $permalink;

    # no pretty permalinks
    if ( '' === get_option( 'permalink_structure' ) )
        return add_query_arg( 'page', $page, $permalink );

    return $permalink . user_trailingslashit( $page, 'single_paged' );
}
Improve this answer
14
  • Thank you very much for the script. Do you have the time to add the link rel="prev"/rel="next" feature as noted in Update 2 of my question.
    – Christine Cooper ♦
    Commented Feb 24, 2013 at 22:31
  • That would require a setup_postdata() – something that should happen much later during page rendering.
    – fuxia ♦
    Commented Feb 24, 2013 at 22:32
  • Should I open a separate question for it even though the 'link rel' is required for this indexing issue to be resolved?
    – Christine Cooper ♦
    Commented Feb 24, 2013 at 23:45
  • @ChristineCooper I have added that now. Not sure what side effects it might cause.
    – fuxia ♦
    Commented Feb 25, 2013 at 0:08
  • 1
    I left this a while to test and see how Google indexes the posts now and I am happy to confirm that this script works perfectly. Well done!
    – Christine Cooper ♦
    Commented Mar 12, 2013 at 17:01
0

Just FYI: Yoast's WordPress SEO plugin takes care of adding rel=prev and rel=next meta data to the <head> of pages. That guy (Yoast) thinks of everything!

Improve this answer
1
  • 1
    -1 I am aware of this but I am not willing to install a massive SEO plugin just for one minor part of its features, neither should you unless you will be using everything it comes with.
    – Christine Cooper ♦
    Commented Mar 1, 2013 at 1:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.