No longer showing for 'money' phrases but long tail combinations rank high?

[lchoice]

I hope someone can shed some light on this as I've been pulling my hair out so much there's hardly any left!

Background: 12 year old website that for about 10 years had Top 3 rankings for 100's of phrases but rankings first dropped off August 2011. Panda seemed to be the cause but finding the exact issue is hard. We are an online travel agent and every hotel page has duplicate content copied from other websites. This has not been changed although lots of sections in the site still rank well, so do the hotel pages themselves. Lots of internal duplicate issues have been resolved but with no effect.

Our old style link, link, link all day long with our 2-word main key phrase as anchor text has given us an unnatural backlink profile but no message has been left by G about this in WMT (yet).

Internal link structure is poor with all pages linking back to the homepage with our 'money' 2-word phrase in 3 places.

Penguin wiped two thirds of all our backlinks back in May 2012.

Why then, do we still rank for our 'money' phrase on the homepage when it has some extra words included and becomes long tail?

e.g.

CityName Apartments (money phrase) - Now ranks page 2-3

CityName Apartments to rent for the night - Ranks #2 on Google in all countries

To make things more confusing other pages rank really well for similar money phrase

e.g.

CityName Apartments Offers - Ranks 2nd on 185,000,000 results (not homepage)

It seems only the homepage is effected (where 95% of inbound links point) but if the site wide duplicates or unnatural link profile was flagged it would effect more than one page of the site. Wouldn't it?

[Marcus_Miller]

Hi Simon,

Can you drop a link? A quick look at the backlink profile would really help here as it certainly looks like a penguin problem for the 'money keyword' from the description but without checking, it's hard to tell.

My thinking would be that the longer search keywords are not competitive so you still rank for these terms. That said, if you have internal / external duplicate pages that are not ranking, if you can't or won't change the content, then maybe they could be worth noindexing as a test at least.

Re the links, I have been working with various sites with a bad link profile and subsequently done a bit of research about what constitutes a safe and natural (or super natural) link profile and the basics seem to be:

• 50% of links should be from topically relevant sites
• keep keyword anchors below 50% and ideally around 30%
• aim to have at least two of the top 5 anchors as branded URLs
• aim for around 70% of the link profile as branded url links
• link to a wide variety of documents with a wide variety of terms
• keyword anchors for product pages should be below 45% of total anchors
• keyword anchors to category pages should be below 25% of total anchors

More details here:
http://www.bowlerhat.co.uk/blog/seo/anchor-text-ratios-and-link-building/

This is possibly a little conservative but is the approach we are using to repair link profiles and for future link building efforts.

Hope this helps
Marcus

[lchoice]

Hi Marcus

bit.ly/aZeroB

I know this sounds like Penguin but this issue kicked in nearly a year before Penguin was announced which is why I'm thinking Panda.

Thanks for taking a look Simon

[Marcus_Miller]

Hey Simon

It's fairly easy to diagnose a panda issue and start by checking the dates you lost traffic. Lets assume you have a panda issue and that has pulled down lots of the rankings for various pages and sections of the site.

Now, that does not mean, that subsequently, you have not picked up a penguin problem as well and looking at your link profile, it would certainly not be surprising:

• No URLs in the top 5 anchors
• Links from low quality sources (http://zurichhotelsweb.com/links/links13.html)
• Links from topically irrelevant sources
• High percentage of keyword anchors
• low percentage of URL anchors

So, if you can check the date you first lost organic traffic in analytics you can pretty much determine if it is a panda problem.

You can read up on the types of thin content here:
http://www.seomoz.org/blog/fat-pandas-and-thin-content

Then, you need to identify all the thin content pages and either update them or remove them. If you can't remove them or they have a purpose then you can noindex them and ideally nofollow the links to them from the rest of the site.

Once you have done that, and you need to be thorough, then you should see a bounce back from those issues but I am fairly sure you will have accrued some penguin issues as well so I would also then check the penguin dates for further drops.

I have one client that got hit by Panda on the 19th of April and lost around 25% of organic search volume from Google and then by Penguin on the 24th and lost around another 50% of the remaining volume. We have solved a lot of the penguin problems but are now looking at the panda problems which can be larger in scale to diagnose and resolve.

Fact of the matter is, you have to be fastidious in your approach and remove anything that is duplicated or thin and then go after the Penguin problems.

Good luck and keep us posted!
Marcus

[lchoice]

Hi Marcus

Thanks for confirming my suspicions.

It's time to bite the bullet and start re-writing all the content!
Simon

[JustDucky]

For whatever it's worth, have a site which was also mauled by Panda & Penquin in April, 2012.  Probably lost around 30% to Panda and Penguin knocked it down to around 25% of the pre-mauling traffic.   The Penguin mauling is algo based and is probably caused by anchor text in manually written comments on relevant blogs or pages.  Without doing anything (as in taking an extended vacation from blogging, linking, etc.) traffic is now at 50% of pre-mauling (probably half of the Penguin effect is gone).  Seems to benefit from each Penguin update by 10% - 20% per month.  I am curious how this compares to "link removal" efforts.