While Google’s previous method of rating websites in accordance to their visualizations helped a lot of internet users to promote themselves and their sites, the major problem that occurred was that this basic algorithm was allowing dishonest methods to flourish.
A generalized hunt for keyword relevance translated into the very popular techniques of word stuffing and site scraping. By means of these methods, search engines were being fooled into positioning high on their ranks those websites where a keyword occurred most often. The old method also encouraged copying of materials from original sites and their posting on fantom sites, in the same hope of achieving popularity to the detriment of honest content generators. Because these techniques worked rather well, Google decided to put a stop to them in order to allow genuine content to be created and to eliminate exploitation of quality materials by mere online racketeers.
Why so often?
As Tom Gregan emphasizes in .Net Magazine, Panda has made it possible for Google to kick out of its services entire domains responsible even for a few unfit pages. This, of course, puts a lot more pressure on content generators to deliver sharp, original, well-written and properly organized materials. Gregan also points out that, on the other hand, Penguin “has brought into focus over-optimization,” which is reflected precisely in the techniques mentioned above. So the advise for web developers is that they produce “not only a significant amount of text, but also a variety of content.”
This way, the website is sure to be following the regulations imposed by Google’s algorithm updates, which are brought about so frequently in order to gradually eliminate from the market those SEOs who don’t submit to fair play but who for some strike of luck have escaped the previous updates. Google’s tactics also work towards raising awareness among users that shortcuts and cheats can’t go too far, and to implement a kind of mentality that is SEO-friendly from the get-go.
Another reason: Preparation for Web 3.0
But there may be another reason why Google is pushing out with so much insistence its algorithm updates. As pointed out by Siva Vadhyanathan Google is preparing for the fast approaching Web 3.0, in which search is going to be done in accordance with queries much closer to human speech and human interests.
At the moment, web queries are responded to by returning a list of sites whose relevance to the query is calculated in accordance to the frequency of the relevant keywords. That means, when we navigate the net we don’t quite get the answers we are looking for, but rather charts of popularity. With Web 3.0 it will be possible for the user to get exactly the reply they want, without any reference to what others may have thought or posted about the topic. But in order to do so, search engines are in need for huge masses of human-generated language. This is why Google has launched its Google Books service, which promises to provide copies of fundamental texts produced throughout the history of writing.
By amassing this enormous amount of texts, Google is hoping to generate a database large enough to get as close as possible to responses aligned to human language. Under such circumstances, spun texts, over-optimized content, websites full of stuffed words, or scraped sites are entirely unwanted because they do nothing but clutter the net, without adding anything of value to it.
So the hope of eliminating the unproductive websites may very well be the primary reason why Google is updating its Panda and Penguin algorithms so often. This also explains the emphasis placed by Google on content poorly written. Although possibly original, this kind of content is certainly not reflective of proper human communication in a mediated environment such as the net, where texts are still judged in accordance with the grammatical and stylistic rules that govern the languages used in navigation. Google is interested in acquiring prominence in this field which is already tested by smaller experiments, and with good results.