tedster - 1:39 am on Sep 21, 2012 (gmt 0)
The Panda history for eHow was quite involved. They escaped Panda 1.0, but were hit strong by the next Panda update. Then they made some very interesting adjustments by migrating their best content to subdomains dedicated to their best authors. As these URLs were indexed and Panda computed opver the next few months, eHow began to climb back up. One year later, eHow was getting past the Panda algorithm pretty much all around.
I do give minimal credit to Demand Media for figuring out how to get around Panda - and then being willing to do the work. What a shame that their work did not include more accuracy, real depth and editorial oeversight, rather doing "just enough" to get past Panda no matter how much their content pollutes the web.
At least the "How to Pour Milk" style articles are pretty much gone. But the scraped-content-spun-with-inaccuracy pages are still happening. That's a joke. but it's also not so easy to catch with a pure machine learning algorithm and no manual spam actions. How can an algorithm today identify inaccurate information? Maybe as the Knowledge Graph matures, there's a chance of that happening, but we're not there right now.
When it comes to describing affiliate sites as a business model that is under attack, I'd say that's not exactly true. It's reproducing feeds with no (or minimal) change that's under attack. I know too many affiliate sites that are doing well to say the entire business model is under attack.
The barriers to affiliate entry are certainly higher today, but it's still easier than opening a new offline business.
And as an end user I am happy with that. I remember well seeing page #1 with ten different affiliates all selling the same item with the same language. That was a really bad SERP experience for the user, no matter how much the sites enjoyed their income. I'd rather see a mix of one e-commerce site plus 9 informational sites than that kind of duplication.
In fact, other kinds of thin e-commerce sites are also struggling on Google these days. There's trouble in business models like lead generation, or fulfillment programs such as drop ship, hold your own inventory, even build your own products. If the brand is not actively growing happy and repeat customers and generating decent buzz, then it simply has trouble holding onto its previous level of organic traffic.
I do think there's something here that is overly severe - with too high a bar being set. This has often been the case in the past, too, when a new algo factor is rolled out. Remember the "minus 950" penalties? Over time those mellowed and faded into the background.
I still expect to see the Panda and Penguin ranking adjustments mellow over time, too. But since the algorithms are more complex, it may well take a longer time to mellow. It's like that cheese on TV ads that isn't mature yet.