Forum Moderators: Robert Charlton & goodroi
Also, when an article is indexed properly it sometimes causes other articles to become indexed improperly. This is happening because the anchor text of links pointing to other articles is ranking more highly than the other articles and so a search for blue widgets will result in serps with red widgets ranked highest and the words blue widgets in the description.
Is there anything more I can do to sort this type of improper indexing out besides waiting it out ?
1. Links to the page I want to see ranking from more than one place on the site - and even better, direct backlinks from somewhere else.
2. Making sure the page includes good semantic variation around the key terms: co-occuring vocabulary rather than a too-close focus on just "the" keywords.
There are some easy things you can do there (with all the usual disclaimers that I defer to Ted and others on any SEO questions; I'm just mentioning implementation):
1. Custom excerpts, which is a ridiculous nomenclature because in fact a custom excerpt is not an excerpt at all, but custom text that is not taken from the page itself. This means it is not overlapping content, so at least they aren't competing on the exact same phrases in all cases (obviously there will be some overlap, especially for single-word phrases). To do this, you'll need to use the_excerpt() in your theme.
2. Custom "read more" text. Instead of just "read more" you can have "Read more about blue widgets". I've heard that Google may not give much consideration to the anchor text in second and third links from a given page to another, the first-link priority issue (Ted?), but assuming that there is some consideration of anchor text in second and third links on the page, it gives you a chance to get another variation in your Read More link. The Headspace2 plugin will handle this.
3. Custom meta title to allow semantic variation between H1 and meta title. Matt Cutts has said this is a good idea (in a Wordcamp presentation). The meta title will be unique to the article page and won't appear on the index page at all. You can do this with Headspace2 or the All in One SEO Pack.
I've heard that Google may not give much consideration to the anchor text in second and third links from a given page to another
There was one well-publicized test that seemed to show this - but that's one test. Later testing by others seems to show that it's not true. I wouldn't say the idea is either proven or disproven at this time, but I lean toward saying it's disproven.
As far as I'm concerned, Google would be foolish to ignore this kind of signal - but even if they do, it's a bit of an aid for your readers.
>>it's a bit of an aid for your readers
That's the main reason I do it. You can have a more compelling call to action than "Read more" and you can give another clue as to what the page is about.
This is the same site that I performed some serious micro managing of pagerank on a while back and those changes are still being noticed as Google re-crawls pages.
It uses custom excerpts and displays just the custom excerpt on the index (most recent articles) and in categories. The value of category pages has been greatly reduced on this site, articles rank extremely well instantly EXCEPT when the index trumps the article page, in that case the index is ranking 3-4 pages back in the index for the same term.
An update, Google WMT was begining to display a backlink for many new articles that also has 7-8 other links attached, those links are from the similar articles section of each article. When these articles are pushed off the index page with new articles the article doesn't get indexed, instead the backlinks disapear as do all of those articles. I would have thought the article would replace the index but that's not the case (else it is happening very slowly, 4 weeks later+).
Summary: articles still rank extremely well and have improved in positioning since making the pagerank flow changes. The index page is most valuable, article pages are roughly 3 times as valuable as they were before inside the internal link structure... category value has been diminished by 66%. (I use some software to run pagerank flow tests).
My problem is with new articles, it's turning into a crapshoot on wether or not the index page will pick off the actual article. If it does the article is as good as lost right now. On June 15th i'll remove the most recent articles from the index page if it continues to see what happens then but i'd rather not.
Telling Google very specifically which pages are most important isn't a walk in the park, it's not helping that Google is making assumptions either, I wish there was a switch in WMT to tell Google, here... let me tell you which to place more value on.
Edit: the only metric I really need to watch is traffic and that's increasing nicely. I do see a major problem though if more articles simply fail to rank at all after being trumped by the index page. 4 weeks and counting is a long time for an article to stay awol, well, in Google anyway.
[edited by: JS_Harris at 5:16 am (utc) on June 1, 2009]
When you talk about "pagerank flow changes" do you mean using nofollow on links to category pages or are you doing more complex stuff than that?
I didn't quite follow this:
EXCEPT when the index trumps the article page, in that case the index is ranking 3-4 pages back in the index for the same term.
Anyway, good luck!
re: pagerank flow changes - as of tonight I've removed all internal nofollow links thanks to mixed messages from Matt Cutts. He had previously stated that you could control the flow of pagerank via nofollow but he recently changed his wording and pagerank sculpting as most understoof it may not work at all, his words. I made the call to halt the process before Google updates every page.
[edited by: JS_Harris at 8:11 am (utc) on June 4, 2009]
