If you are not into the technical side of website design and search engine optimisation (SEO) the title of this article probably means very little to you. Never fear! I will explain as we go. Regular readers of this blog will know that I am currently learning how to make money on the internet from writing. It is because of this activity that SEO has become an important topic in my life. I wasn’t born an internet nerd, it is an acquired disease!
What is Keyword Stuffing?
Keyword Stuffing used to be an activity that web designers indulged in to make sure that their web pages ranked well for a variety of search terms. So, if you searched for “digital camera” you would encounter a page of text such as:
“My best digital camera was the digital camera I bought on holiday at a digital camera store in a large digital camera market…” etc. etc. ad nauseum.
Clearly this kind of document reads badly to a human, but old style search engines loved it.
You see, until fairly recently the internet was indexed like an academic document. If you said (in the html page title) that the topic of a web page was “free health care” and stuffed the page with keywords about “free health care” then the search engine would accept that “free health care” was indeed what the page was about, despite the fact that the page had endless pornographic images and links.
The Google Effect
Google is currently the world’s largest search engine by traffic volume. The guy’s at Google became increasingly fed up with this kind of web page ranking highly (along with the rest of us) and so they started to change the methods used to index a page in the Google directory.
The exact method’s used are by Google are a hotly debated subject and employee’s have none disclosure agreements and so we will probably never know exactly how Google indexes its pages, but we can say with certainty that a simple tactic such as I just described no longer works.
And so keywords died, or did they?
You see Google excels at answering the questions people ask. When you enter search text in a search engine you are effectively asking “Where can I find this resource?” If Google’s reply is relevant then you are happy and will return. If the page is full of useless spam links then of course next time you search you might just try another search engine. So, if you ask Google about “BMW cars” then this is what is needed as a reply, there is no way around it. This is what the industry calls a keyword phrase and is a valid means of looking for information.
Well, these day’s keywords appear to work a little differently than in the past. Keywords in an article can be measured by how many times they appear e.g. 4 appearances in a 100 word article is a keyword density of 4% and so on. It seems that modern indexing methods penalise a keyword density over about 2-3% (opinions vary).
The search engine algorithm also appears to alter this penalty based upon content type. If you are talking about “Tiger Woods” then Google will expect to see a few repetitions of “golf” and maybe a few other topical keywords besides!
The point is that it appears that the algorithm is adaptive to content and context.
While this isn’t science fiction standard artificial intelligence, it is certainly heading down that road. When a sci-fi hero speaks to his voice activated computer and asks for information the results are generally exactly what he wants, even if the context in which the question was asked is ambiguous.
Perhaps the first artificial intelligence really will be a search engine and not an insurance actuarial computer or something similar. If it is a search engine I just hope it doesn’t get caught watching porn or downloading warez while it should be working!
Anyway, I said that keywords are dead. What brought them back to life?
Well again it seems that topicality of a search is determined by: Uri, directory (if used) and then html page title. So www.spamsite.com (hopefully not a real website) will always have topicality if someone is searches for “spam site” with big G. So now we see a rash of “niche” sites abusing trademarks of well known companies in order to gain traffic.
Optimising a Webpage for Relevance (Long Live Keywords!)
The best option for ranking purposes seems to be to include your keyword in the url of the site, in a local directory if necessary and then in the html page title. Additionally this keyword should be in the H1 heading and possibly also used in h2 and h3’s if they exist also.
For example: http://spamsite.com/spamsite/spamsite.html
The “spamsite.html” document would be of the form
<h1>spamsite main title</h1>
<h2>spamsite secondary title 1</h2>
<h3>Spamsite additional information 1.1</h3>
<h3>Spamsite additional information 1.2</h3>
<h3>Spamsite additional information 1.3</h3>
<h2>spamsite secondary title 2</h2>
<h3>Spamsite additional information 2.1</h3>
What amuses me about learning all this over recent weeks is that this “technique” for SEO, so successfully employed by websites with nothing useful to say, is that it is also the defacto standard for referencing academic documents via html and I am pretty sure it was part of a web standards document released by http://w3.org some time ago.
Of course there are other factors involved in determining where a particular page comes in the results of a search such as authority, age of the site, age of links and so on. I don’t pretend to know it all about these subjects, but the sheer irony of guys with almost nothing to say earning a living by bucking the system and mimicking academic style was just too good to pass over.
If we do eventually develop a science fiction style super-intelligence; when we ask it a question will it just repeat keywords to us in answer?