Interesting ranking results of a post-Penguin test

[Shepherd]

As I've mentioned before we had some penguin damage on a few sites. One of the directions we went post penguin was to test new sites (in the same niche). We were unsure whether junk links or interlinking between owned sites could have been causing problems.

With one of the new sites we employed the same junk link strategy with no interlinking between owned sites.

We targeted a few main keywords and used an EMD matching the main keyword we targeted. The site is a very thin site, minimal content.

domain: widgetexample.com; keywords: widget example, discount widget example

Links were all pointed to widgetexample.com (home page) with some varing anchor text (like "website" and "widgetexample.com") but the majority of the links were anchored with the target keywords.

After a few months of link building the site started to rank, currently #9 for "widget example" and #7 for "discount widget example".

The interesting thing is: google refuses to rank the home page for these terms, they rank deeper pages like widgetexample.com/red

No link building was done for any pages other than the home page.

Rankings have been steady, within 2 spots for over 3 months now.

[tedster]

The interesting thing is: google refuses to rank the home page for these terms, they rank deeper pages like widgetexample.com/red

That is interesting - especially because we have several members here, struggling with the opposite situation where the home page is ranking and they would prefer an internal page. Sounds like an area worth exploring some more.

[Dymero]

Despite the thin content, maybe the internal pages are showing higher relevancy for the keywords than the homepage with things like title tag, etc?

[somecat]

I noticed yesterday that if I search for red widgets related to my page, I see a lot of amazon and ebay ahead of me, but when I added amazon to the query, I'm number one lol

I'm not an affiliate of amazon, nor do I have anything to do with it.

[scooterdude]

If you are saying that interlinking between owned sites appears to be your culprit, well, that exactly my experience.

interesting.

[Shepherd]

If you are saying that interlinking between owned sites appears to be your culprit

Not really, no way to know that at this point since it is commonly believed that penguin has not been re-run since October 2012(?).

I was talking more about google ranking pages other than the pages being targeted by junk links for the keywords we are targeting. It feels like it's a "loophole" in the algo. Almost like they know that you are creating junk links so they're not going to rank the page for the keyword but then something bigger kicks in and say this site must be about "keyword" so lets find a page to display.

[outland88]

Is that index page more commerce related (even pointing to a mother site) or is this an info site?

[Shepherd]

widgetexample.com home page thin content (300+- words) w/ 50 links to widgetexample.com/red, widgetexample.com/blue, widgetexample.com/green, etc.

widgetexample.com/red: thinner content (50+- words) w/ link to widgetexample.com/red/red-widget-example.pdf.

red-widget-example.pdf: thin content with link to ecom site.

[Panthro]

OP describes behavior I've noticed post-Penguin and post-EMD update.

[outland88]

thin content with link to ecom site.

I thought so. That’s why I asked because of the index page problem you mention.

I've been observing people doing what you mention for about a year. More than a few brands use the methodology. From what I have seen it can be condensed more than you think. Some sites seem to vanish with Penguin updates but the pages reside about in the same positions you mention with the link building.

[Shepherd]

and post-EMD update.

oops, didn't even think about the EMD update. Need to reboot test for a generic domain now.

Some sites seem to vanish with Penguin updates but the pages reside about in the same positions you mention with the link building.

Having trouble following you there outland88.

[outland88]

To clarify

From what I have observed the link building may lead to rankings as high as positions 8-20 in very competitive areas. Seldom rises above that. Once there is an update the rankings often vanish. I speculate that relates to the very distinct link profile. This doesn't seem to matter to the companies involved because they just turn out newer sites to achieve the same goal. In some areas what you're doing may only require 10-20 links but in the areas I track it may take hundreds of links.

It typifies the sad state of affairs Google is in.

[Shepherd]

OK, I see outland88. Is this something you've seen post penguin or prior?

We've just started another round of link building which, if history is an indicator, should push the site into the top 5 for the targeted keywords.

[Shepherd]

Seeing a significant drop in rankings for keywords we targeted with junk links for this site.

[Awarn]

I have that issue with the other pages ranking better than the home page. Odd part it is the real junk pages that rank. I have some pages that prior to Penguin were strong and ranked very well on their own and they had links to those pages. None of those pages rank well now. However pages for items that rarely sell (junk) rank better than ever and at times rank better than the homepage. It is strange too because they jump around. I might see the homepage for a while (a week or so) then it is replaced by one of the junk pages, then switches to a different junk page then after a week or so the homepage comes back. Simple fact is Google is screwed up.

[atlrus]

Like the OP, I have something similar happening on a couple of websites.

I think it has something to do with the way Google is selectively applying penalties to single pages, rather than entire websites.

The only way I can see this work is if the home page is triggering a non link related flag with the algo, good enough to ban the page, yet not the entire website. So the algo is left with all those "good" links pointing to the website, yet nothing to show, thus it simply pulls out the "next best" page.

OP - does your home page rank for anything at all?

[Shepherd]

OP - does your home page rank for anything at all?

only for a search for the domain name: "widgetexample.com", also (i just checked) for quoted long text unique to the page.

[atlrus]

That's in line with my observations.

Since this is just a test site, why don't you try an redo the home page completely and see if it will recover?

[Shepherd]

why don't you try an redo the home page completely and see if it will recover?

first, this test site was started after the most recent penguin *known* refresh/update (october 2012) so it technically should not be affected by penguin. That said, it does display the same ranking characteristics of sites that we have that were *presumably* affected by penguin.

second, it's my belief that whatever causes the target page not to rank (penguin or something else) has to do with inbound links and not on page factors. Something like: if (page) has x% of inbound links using (target keyword) as the anchor text then (page) will not rank for (target keyword).

[atlrus]

second, it's my belief that whatever causes the target page not to rank (penguin or something else) has to do with inbound links and not on page factors. Something like: if (page) has x% of inbound links using (target keyword) as the anchor text then (page) will not rank for (target keyword).

I doubt that, since in your case: if (page) has x% of inbound links using (target keyword) as the anchor text then (page) will not rank for (anything)
I would understand if your home page did not rank for "widget example", but ranked for "discount widget example", but you say that's not the case.

In your case page rank flows:

link>home page>inner pages

so if the home page is taken out of the equation because of the links, there would be nothing to pass juice to the inner pages and they would certainly not rank 1 page on 50 words alone for "widget example", assuming it's not something that returns only 10 sites :)

[Shepherd]

Link building was targeted to the specific keyword, no other links were built. There would really be no reason for the page to rank for any other keywords, on page factors or inbound links. The page does rank for quote text unique to the page and for the domain name.

So what I am taking from this is that the page is not penalized in general, it is penalized specifically for the keywords being targeted.

A good test for this theory would be to build links to the page that is currently ranking for the keyword that are keyword anchor text rich and see if that page stops ranking for the keyword. I'm betting that it would.

As for passing juice, i'm not thinking that the links are devalued, just the page for the target keyword. If that's the case then there would still be link juice and it would still be flowing throughout the site.

[Shepherd]

assuming it's not something that returns only 10 sites :)

currently 496,000 results for "widget example" and 186,000 for "discount widget example"

[atlrus]

In your OP you said:

We targeted a few main keywords and used an EMD matching the main keyword we targeted.

and now you say:

ink building was targeted to the specific keyword, no other links were built.

Did you use only one anchor for the links? If so, there is a chance for over-optimization, although I still don't see how the links will pass so much juice to the inner pages, if they are discounted by google, that those pages rank in the 1 SERP. If, on the other side, your home page is banned for something not link related, let's say keyword stuffing, then those links could pass full juice and google simply uses your next best page.

Of course, if I am right, adding more links could simply lead to your home page getting enough authority to overcome a page-related issue, I don't know.

Now that you got me thinking, I might pull one of those websites of mine and do the home page redo experiment myself :)

[Shepherd]

the specific keyword(s), i forgot the s.

checking keyword density...

each of the keywords are in the title tag once, not in the description tag, and only occur once on the page.

on the page that is being displayed, there is only one occurrence of the keywords, they are in the on page text, none in the title or description.

no on page links contain the target keywords in anchor text, however the domain is an emd so the href does include the keyword.

[Shepherd]

also, interesting, the home page tbpr is 0, interior (ranking) page is 1. in fact, all 50 interior pages are tbpr 1.