Effects of Panda on Thin Affiliate Sites

by rishil on July 11, 2011

This is not going to be a long post, nor is it going to teach much. I am putting it out as a way of an early warning system for thin affiliate sites that survived the panda effect.

Basics  Of Thin Affiliates

I run a few data collecting (and money making) thin affiliate sites. I dont hide the fact and I dont overly try and push them either. I build them and let them run. Most of these survived the Panda 1.0 update. Panda 2.0 was a different matter.

For those who dont know what a thin affiliate site is, Its pretty much a thematic site put together with a series of affiliate feeds on related products that automatically create product pages. The game here is to try and rank for the long tail – often Google would give priority to thematic sites built on Exact Match Domains, over the original retailer. The other advantage to date has been that some ecommerce sites are so huge that they only concentrate on the larger volume, money making SERPs, and tend to ignore smaller volume products.

This is where a number of thin affiliates who make money play at. The SERP competition is low, and it is often easy enough to outstrip the original retailer in the SERPs for their own products. Its nothing new and many have made a living out of it comfortably. (Sadly not I – I get bored too easily).

In essence, though not exactly “scraping”, these sites often tend to be repetition of content that exists on other sites, passed on through the blessings of the original content producer.

So what happened?

Initially Panda 2.0 halved the traffic to the sites. Literally. and I thought to myself, well 50% aint so bad. And I thought that Panda updates are going to be periodic, bulk algo changes. However, if what I am seeing is the norm, then the roll out is far from finished. Here is the aftermath as of yesterday:

Light Affiliate Site Destroyed by Panda Sloww Roll out

Light Affiliate Site Destroyed by Panda Sloww Roll out (click to get full image)

Learnings

Panda 2.0 isnt a one of release, and can potentially affect sites weeks after its roll out – just because you survived, or partially survived the aftermath, doesnt mean your sites are safe if the content isnt great.

I lost the HOMPAGE. Thats right – the home page does not rank. In this example and 5 other hit sites. In fact, an inner page ranks for some of the KWs that the home page used to rank for. Is this a signal? I dont know. but I know that the home page of a blog based site tends to be the strongest part of a site – and often short term content that appears on it helps sites rank for KWs even before the inner post pages are indexed.  Whats the easiest way to stop sites ranking quickly for constant low quality content? Kill the home page.

Blood Thirsty Panda Image Source http://forums.voogru.com/off-topic/13741-red-white-pandas-4.html

Blood Thirsty Panda

Damage Control?

I dont think that the site is worth recovering, despite the volume of traffic it used to drive. Why? Because I know that it was a crappy affiliate site, I knew that there wasn’t much content on it that was original, and I know it deserved to die. So I will let it R.I.P .

So what will I do? If you are new to my writings, you should head over and read about SERP Sniffing. But in essence, this is a strategy that allows you to capture long tail keywords that drive volume, AND are easy to rank for. Thus armed with this data I intend to build this site on a new domain, without all the crappy element, write custom content for it, make it useful, and hopefully make it pay off.

What should you do? Be wary and keep an eye out – it isnt over yet.

Image Source http://forums.voogru.com/off-topic/13741-red-white-pandas-4.html

Share and Enjoy:
  • Twitter
  • del.icio.us
  • Facebook
  • Google Bookmarks
  • FriendFeed
  • Sphinn
  • LinkedIn
  • PDF
  • StumbleUpon
  • Suggest to Techmeme via Twitter
  • Yahoo! Buzz

Rishi Lakhani is an independent Online Marketing Consultant specialising in SEO, PPC, Affiliate Marketing and Social Media. Explicitly.Me is his Blog. Google Profile

{ 15 comments… read them below or add one }

Mark July 11, 2011 at 9:42 am

Two questions:

1) You lost the homepage – was it an exact match term/domain?

2) Going to a longtail strategy – but you’re planning on writing content for thousands of keywords? Hum?

Reply

rishil July 11, 2011 at 9:51 am

1. Home Page – it was an EMD, but used to rank within minutes of any post going live for that content. This no longer happens. So I lost home page “authority”.

2. 1000 Keywords can be targetted by under 100 pieces of good content. Remember that this is a “thematic” site. Cost of that content would be about $10 per page for the initial outsourced material, then hours tweaking each. The aim is to make each page as relevant and as useful as possible. And not to rll out at one go – a decent strategy would be to:

- target the high volume KW sets first
- Bundle similar KWs and content that fits thematically
- optimise for that, and then source related KW sets
- bring in other long tail variations that match the content into those specific target pages.

Its not going to be easy, nor is it going to be cheap, but then the SERP Sniffing taught me which were more valuable, and which to discard, as well as what the value of each KW set is to me financially.

Reply

Ariel Ozick July 11, 2011 at 10:00 am

Why do you think this is panda and not a penalty or something else?

Reply

rishil July 11, 2011 at 10:05 am

Panda, by its intent, was to clear out scraping and duplicate content of low value. This defined the site. Second, the drop offs coincide with the 2.0 roll out.

AND it wasnt just the one site – 4 others got hit exactly the same way. Penalties work slightly differently IMHO.

Reply

Ariel Ozick July 11, 2011 at 12:04 pm

When you talk about EMDs not ranking for their name, that’s less Panda and more something else algorithmically, IMHO..

Reply

Fredrik Eriksson July 24, 2011 at 4:26 pm

Yea sure they happend to be EMDs, but the content got penalties from Panda. EMDs has really nothing to do with it?

EMDs have some advantages over a non-EMD domains, and even if it was an EMD, it still got problems when the Panda-udate rolled out.

Reply

Mark July 11, 2011 at 10:42 am

Good luck with that strategy. I burnt about a grand on content for one of my sites and got about hundred or so decent articles, they tried to fork off stuff for longtail. In short, it really didn’t work, ended up with about 8k visitors a month, which was disappointing to say the least.

I don’t think you’ll get content of any quality for $10 a pop and when you’re talking longtail, you need to start thinking in the 10,000s – not 1,000 IMO.

As a contrast (and because I was pissed off) – here is how 500k of instantly generated pages have handled Panda: http://i.imgur.com/ZGTVm.png

Panda is a complete clusterfuck and just is aimed at squeezing out potential aff competition, nothing to do with quality.

Reply

rishil July 11, 2011 at 10:55 am

lol. Thats a good success rate on the auto gen. The reason why I hope it will work is because most those SERPs are low competition. Without trying, I will wonder for the rest of my life whether it would have worked or not. The 1K a day visitor figure translated into decent affiliate revenue, so all I am losing is money that the site brought in – it was doing about £250-£450 a month, plus double that during Christmas and Valentines. So worth replicating with quality – because low bounce rate=more visitors into mechant sites – its original bounce was about 60%…

Re Auto Gen – am looking at a 800K spin bundle to flow on a custom platform, so will kep you updated on how that works…

Reply

Mark July 11, 2011 at 11:05 am

Btw, my “Good luck with that strategy. ” was a geniune good luck, in retrospect it doesn’t look that way. Just wanted to clear that up.

Reply

netmeg July 11, 2011 at 4:18 pm

There was one person on WMW speculating that the various iterations of Panda were in turn taking on various types of long tail, starting with two word phrases, then three word phrases, then refining, etc. I haven’t seen any Pandalized sites up close and personal (i.e. with analytics and log files and whatnot) myself so I can’t confirm that, but it does make sense to me in a weird sort of way.

Reply

Mark July 11, 2011 at 4:59 pm

Great post, thank you for the insight. Has anyone noticed panda hit subdomains?

Reply

Michael Martinez July 11, 2011 at 10:52 pm

Would you mind contacting me via email? I have some questions I’d like to ask if you’re willing to share a little information. I don’t have an opinion on whether you’re seeing Panda or a new algorithmic penalty but I’d like to form one.

Reply

rishil July 12, 2011 at 7:57 am

Michael Martinez seems to think that this may be part of a new algo penalty rather than Panda. So am going to dig ito deeper.

Reply

Robert July 13, 2011 at 10:43 pm

I find it hard to believe it’s over just because G seems to have missed out a number of sites which arguably *should* have been hit, according to the perceived criteria…

Reply

Nick November 3, 2011 at 1:37 pm

Just a quick question as i am interested in doing something similar. To create pages that rank well for these keywords, they have to have good title tags, h1 tags etc. So your page could for example “Cheap iPads for sale” to make this page panda proof people have to find the info on this page, well… decent enough for them to stay a few minutes.

Is this going to happen?

People looking for “cheap iPads for sale” are not looking for info, they are looking for “cheap iPads for sale”, hence why thin affiliate sites are failing after Panda. A small % will click on ads, affiliate links but a larger % will just hit the back button.

Obviously i don’t know what niche you are going to target, probably not “cheap ipads” But just wondered how you would get around this? How could you create decent content for a page like “cheap iPads for sale”?

Reply

Leave a Comment

Previous post: What We Learnt From a Pills Link Hacker

Next post: 9 Ways in Which PR Teams Fail SEO