Search Engine Spam: Alt Attributes
November 13, 2005 – 9:45 am
I've decided to write up some sections about Search Engine Spam, how it's used and why you shouldn't use it besides the obvious, it's wrong to spam Search Engines and these techniques will undoubtedly get a black mark against your web site if not dropped altogether. Perhaps spam techniques may work very briefly but the technicians over at places like Google are working harder than ever to combat the spam sitting in their indexed.
So what can you do with the Image alt attribute? There is an art known to SEO enthusiasts as "keyword stuffing" whereby instead of using the alt attribute for what it was designed for, to explain what the image represents, blackhat SEOrs put as many keyphrases and words in there as possible. A large percent of your visitors will see the image and not the text that represents it. However, the text is read by search engine spiders and can be used within the indexing information for the web site. There are several aspects to address here:
Proof of Indexing
My own business web site, 3.E.Media, contains various images of fruit in the top right hand corner (or elsewhere depending on the style sheet chosen). These all contain their relevant alt attribute. Now, despite myself being an SEO enthusiast I don't even try to get my own site high in the rankings for phrases like "web design", "web development" etc. Why? Because those subjects are so saturated on the web I would be fighting a losing battle! My web site is there for one purpose, to show new and prospective clients our work, and because I should have a web site! However, looking in my statistics I can see that a lot of people come to my web site searching on the words "blackberries", "kiwis", "strawberries". Well I don't mention these words anywhere on my website besides those alt attributes. So there is the proof, the spiders will read and index the alt attribute text.
Implications of this Spam method
So what are the implications of using this method? Afterall, everyone browses with images switched on these days right? No one cares what the alt attribute says right? Wrong. The most important group of visitors this will affect are your partially sighted users who require a screen reader or braille reader to let them understand what is on the web site. If you have a logo for a web design that says
Our Company Logo
However the alt attribute contains the text.
web design web development photoshop design graphic design brochure design designs web design agency
Then that isn't very fair to those visitors that rely on that attribute to tell them what the image displays is it? Why should they lose out to the complete browsing experience of your page? Afterall they could be a potential customer or potentially your biggest client, so alienating them is not a good idea. Then, putting the disability discrimination act aside, some people do still browse with images turned off. Not everyone has superfast broadband (and believe me sometimes the broadband isn't that fast either), this month I've also received a visitor to one site using the Lynx browser (text only browser) which of course doesn't display images either and gives the visitor the alternative text.
What should you have?
You should explain what the image is quite simply. If it's a text graphic, put the text that is used. If it's a photo of a beach ball, put "Beach Ball Photo". If the image has no bearing and is just there for a display purpose either state this or else leave an empty alt attribute i.e.
<img src="images/blackbox.jpg" alt="" />
There are further options available to you where images are concerned. There is the option to give the image a title using the title attribute. This can give a little more information on what the photo is for, for example:
<img src="images/beachball.jpg" alt="Beach Ball Photo" longdesc="Girl playing with Beach Ball" title="A Photo of a Girl playing with a Ball on a Beach in Cornwall" />
All of this will be picked up by the spiders and will have slight weightings on your indexing, however please note that whilst the text will be indexed it will not be considered as important as page text, lists, link text, headers etc.
A Possible Solution?
I've had a think about this. Most alt attributes only require a maximum of X amount of characters so surely if the techies over at the search engines could develop an algorithm that picks up if the number of characters in the alt attribute is between X and Y then flag it for possible spam. If it's over Y characters then it's 99.9% spam and prevent the attribute from being used as a bearing on the web site. Okay it's a rough idea but I don't see why it's not been looked at.
The new Jagger update has started to iterate throughout the Google servers now and I actually checked a known web site that use these tactics within their site. For a top key phrase they are still within the top 3 results. I only know about this site as it's the competitor of a new client of mine. The trouble is how can I sit and explain to my client that good SEO means no techniques like this when his main competitor is top of the listings for keyword stuffing??! Admittedly I've not started work on his site yet, just written a report (ripping his site to pieces in the process) explaining the downfalls and what we can do to improve it. The finished site will be live in January/February so I'm hoping that it will make the difference, however that remains to be seen.
Sorry, comments for this entry are closed at this time.