|New pages not ranking, even with good external and internal linking|
| 7:56 pm on Mar 8, 2009 (gmt 0)|
I have an ongoing issue with one of my websites at the moment. I'm quite an experienced SEO myself, so not new to the game.
This particular site is PR6, has hundreds of thousands of pages indexed, 1.3 million members and ranks #1 - #3 in Google for hundreds of thousands of related terms.
Think of the site like Youtube, in that we have galleries and downloads, though we support multiple formats. It's a very similar setup to all of our main competitors.
Here is an example of this strange phenomenon that's going on at the moment and has me baffled. I created 2 new galleries for new "formats" that have recently been released. I then posted a blog and forum thread to inform members of the new galleries, the posts contained links to said galleries. I then proceeded to build unique and relevant 1 way links to the pages though not aggressively. One of the links was from the number 1 SERP for this particular model.
These pages receive a lot of internal link juice, they're only 1 level deep, assuming we're not classing the top level as a level.
Anyway rather than these new pages ranking as I'd expect, random posts from either the blog or the forum are ranking for the targeted terms, rather than the gallery page. These posts are virtually orphaned. In an attempt to convince Google it was the gallery pages that were important for this term, I added a link from the post to the gallery page to absolutely no effect. While discussing this with SEOmoz, they also suggested I do this to fix the issue but as I explained, I'd already been doing it.
The gallery pages aren't filled with tons of unique textual content, but then neither is youtube, nor none of my competitors. It's the nature of our industry. HOWEVER we have made an effort to provide some decent textual content on each gallery page. Whereas our competitors have made zero effort to do so.
We have a higher PR than all of our competitors and less pages, therefore more link juice to go around. I can't see this being a duplicate content issue as our competitors would surely be having the same issue or worse, as they have more gallery pages, lower PR and less unique content on each gallery page.
As I said earlier in this particular niche, we rank #1 - #3 for hundreds of thousands of related terms and for the most part, these ranks appear unaffected. But new pages for whatever reason are struggling.
Right now I can't even clutch at straws as I nor nobody I've spoken to can see a problem.
I'm hoping one of you have had a similar experience and figured it out.
Many thanks for reading all of that, it was hardly a short post!
| 10:48 pm on Mar 8, 2009 (gmt 0)|
The post covers a good many of the questions I'd normally ask about... site structure and PR distribution being an important one with regard to new pages.
The unanswered questions that do come to mind are mainly about linking....
- How old are the external inbounds?
- How "natural" and varied is the anchor text?
- Is the inbound anchor text perhaps too varied? ;)
You say the one-way inbounds weren't built too aggressively, so I assume that traffic to the new pages was then significantly greater than the links to them.
I have seen natural inbound links that took six months to get fully credited. I should add that in this case I'm assuming that these links were the factor that caused the page to move up. I've seen other inbound links cause a significant boost in a week or two or three and then trail off in their boosting effect.
I do feel that text on the page is very helpful to make inbound links effective, so you might look to adding some to your pages, which I gather from the post are thin.
| 11:31 pm on Mar 8, 2009 (gmt 0)|
Hi Robert, thanks for taking the time to reply.
- How old are the external inbounds? They were added about a week ago, not long after the new pages were created. It may seem early to start jumping to conclusions but so far this is a replica of a situation I had a few months ago with a new batch of pages I released. These pages still fail to rank and are stuck on page 3 while related blog posts rank on page 1.
- How "natural" and varied is the anchor text? Even naturally, the pages tend to receive the same anchor text which is "Brand Model Content Type". But to Google that could be perceived as unnatural even though it's not. I guess I could actively try to link to it with more random text.
- Is the inbound anchor text perhaps too varied? No, probably the opposite.
There are only a handful of good links pointing to the page, it's not as though it has tons. In the past I've been able to rank number 1 for similar terms, from internal link juice flowing to the page alone. As time passed I've become better and better networked and so my new important pages can gain some decent external juice quite quickly. Perhaps that's part of the problem, it all happened too quickly.
Maybe it is a case that the internal and external links are just too targeted and thus tripping a filter. Maybe it's worth altering some of the anchor text and see how it goes.
I don't try to be clever or smart when it comes to my associations with other sites, I'm completely transparent. The websites that linked to these pages, had linked to the site before, as many of them do regular features on our content. Everything I do is natural and because it aids my users experience.
Every single one of my competitors cheats. All of them. They buy links, thousands a week. Try to be smart when it comes to linking their ridiculously big networks of duplicate sites.
Times like this frustrate me when I see my competitors cheating without penalty and enjoying 500,000 visits a day. But what can I do, continue to do what I believe in or join them on the dark side.
| 11:53 pm on Mar 8, 2009 (gmt 0)|
The situation you described does remind me of a head-scratcher situation I had where the intended page just "would not" rank but some other page did instead. I eventually resolved that one by expanding the on-page text content to include a few 3 term co-occurring phrases. When I read the page objectively, I saw that it was both thin and way too focused on the specific keyword.
| 12:34 am on Mar 9, 2009 (gmt 0)|
|I eventually resolved that one by expanding the on-page text content to include a few 3 term co-occurring phrases. When I read the page objectively, I saw that it was both thin and way too focused on the specific keyword. |
Tedster, thanks very much for sharing that with me. It could well be the issue with lack of textual content and the textual content that is there, is quite focused.
I'm going to run some tests with the anchors and text content.
I'll make sure I report back with any changes I notice. Perhaps others will find this experience useful in the future.
| 2:04 am on Mar 9, 2009 (gmt 0)|
I've had a bit of a play. I've just changed a couple of pages for now.
I'm a bit bunged up with flu and tired, my creativity juices are all dried up. I struggled coming up with alternate phrases..... urgh I seriously need some sleep :)
[edited by: tedster at 3:24 am (utc) on Mar. 9, 2009]
| 3:23 am on Mar 9, 2009 (gmt 0)|
I'm not talking about alternate terms or synonyms. Rather I'm talking about "co-occurring" terms, the vocabulary that naturally occurs within content about the topic. For example, in an article about "apples" you would usually find some (but not all) of the terms "cider" "juice" "orchard" "Gala" "heirloom" and so on.
| 4:17 am on Mar 9, 2009 (gmt 0)|
|..."co-occurring" terms... |
This is one of the reasons for not being stingy with text on a page. You not only want to keep the template itself from dominating, but you also want enough text that there's opportunity for richness and natural variety in the vocabulary you use.
If there's not much text beyond your template (ie, beyond the global content that's common on all your pages), then your pages may get filtered as duplicates. If there's not much text beyond just your target terms, your pages won't be seen as documents that are enough about your content to be potential resources.
But if you take the idea of co-occurring vocabulary too far, say by trying to enhance search or Adsense attractiveness by including an unnatural of synonyms for your target phrases, you're liable to get into an area that Google regards as spam.
So, there's some tightrope walking between not enough co-occurring vocabulary or too much. Google is getting better, I think, at avoiding collateral damage, but I feel I'm still often avoiding what would be natural repetition to not trip the filters.
[edited by: Robert_Charlton at 4:20 am (utc) on Mar. 9, 2009]
| 12:17 pm on Mar 9, 2009 (gmt 0)|
Thanks, I understand now.
At present each gallery pages have a small, related article of about 150 words on it. But below the gallery itself. Maybe worth moving it up. But then that compromises functionality. *shudder*
This whole issue troubles me somewhat. As I've always believed that sites like youtube don't need much textual content, because their content isn't documents but video's. Their contents worth is proved via masses of deep links. My own experience has always backed this up. Quality content is quality content.
Working on some more in-depth documents now for the struggling pages.
| 7:34 pm on Mar 9, 2009 (gmt 0)|
|...related article of about 150 words on it. But below the gallery itself. Maybe worth moving it up. |
Maybe not. If the gallery is essentially images, there's no reason it can't be above the text.
Also, while 150 words is shorter than I'd pick for some types of sites and articles, it's a lot longer than I thought you might have. I'd concentrate more on defocusing the vocabulary.
| 8:37 pm on Mar 9, 2009 (gmt 0)|
|I'd concentrate more on defocusing the vocabulary. |
I'm also leaning heavily toward an issue with overly focused anchor text.
In the past, this has never been an issue for us. But Google is updating all the time. Perhaps Google is now more focused on percentages rather than sheer abundance. So even if only 10 external links point to the URL, if they all have the exact same anchor text, that's 100% and thats too focused / manipulative = Penalty.
I was just discussing the duplicate content issue with a well respected SEO from #*$!. He said that if it were a duplicate content issue, the pages wouldn't be indexed at all. And as the pages are indexed, would this suggest it's not a duplicate issue?
Thanks so much for both your help with this.
| 8:20 pm on Mar 10, 2009 (gmt 0)|
Just an update. I added a 50 word intro to the page and expanded on the review, avoiding keywords for the most part.
I also went back and altered several of the internal and external links. Will update as I see changes.
FYI, after making the changes last night, the page that had been at position 8 for over a month for targeted KW, dropped to #12 by this morning. I'd not altered anything on this page or any links to it.
| 12:34 am on Mar 11, 2009 (gmt 0)|
Just curios do the gallery pages link to the blog or the forum pages?
| 2:06 am on Mar 11, 2009 (gmt 0)|
|Just curios do the gallery pages link to the blog or the forum pages? |
The only link to the forums and blog is from the main navigation. There are other links but all nofollowed.
| 3:09 am on Mar 11, 2009 (gmt 0)|
Update: Each gallery pages has a series of categories. These categories are only ever linked to, from it's gallery parent.
Google is now ranking an almost completely textless category of the new gallery, rather than the main gallery page which as I say has links, this one in particular mainly internal.
This is really started to drive me insane. I feel like I'm chasing my own tail.
| 12:19 pm on Mar 12, 2009 (gmt 0)|
This is insane. Google has totally lost the plot.
Yesterday /green/brand-model-widgets.html was ranking and today it's /gizmo/brand-model-widgets.html. These pages are very sparse in their textual content and only have links from the main /brand-model-widgets.html page.
The main brand-model-widgets.html page now has much more textual content, more varied anchor text.
I'm at a loss. Thanks for all the assistance.
[edited by: Robert_Charlton at 6:38 pm (utc) on Mar. 12, 2009]
[edit reason] examplified specific [/edit]