My recent editorial about metatags generated some helpful reader responses. My basic point still holds. Metadata as currently structured, in the context of a non-ideal world, is open to spam and deception, and most search engines give meta tags very little weight. The overworked site owner would probably find it a relief to simply stop using meta keywords and possibly descriptions entirely.
There were some cloudy parts to the argument that I’d like to clear up, though. The first is to reaffirm the utter usefulness of emerging metadata standards in the contexts where they’re likely to thrive. Information science professionals and company knowledge management mavens reminded me of how vital it is to classify information, at least in contexts where the classifications may be expected to be relatively free of spoofing. Corporate intranets and libraries (though some may think to the contrary) are more likely contexts for what Jurgen Hagermas calls communicative action (in which interlocutors are motivated by ‘an orientation towards mutual understanding’). Much more so than the public Internet, which, as we know, is a marketplace open to all forms of strategic action, in which participants are often motivated by a desire to get their own way.
Another reason to continue to use tags even in the context of the free-for-all public Internet is for site search. Well-tagged pages might not help you rank well in search engines, but once visitors *do* find your site, they may then search its pages, and the more structured data you can provide to the site search tool, the more likely it is that your visitor will find the exact pages they’re looking for. Summary descriptions are always an aid to navigation; they’re just time consuming to write. I tend to think that pages that contain hard-working content merit good description tags. Unfortunately, there is also a lot of bad content or non-content out there masquerading as content. If a site can barely produce a coherent sentence on its visible pages, then it would probably take longer to write each page description than it took to create the page.
Corporate portals might also want to tag their public pages in a way similar to the way they tag their private pages – not necessarily so “the public” can find them on Google, but so employees and partners can find *all* relevant documents, including those which are openly available to the public.
But the general state of consumer Internet research will probably not be improved a whole lot through the use of these tags. Search engines discovered in the early days that if you fail to develop schemes and technologies for taming the chaos of the net, it’ll look a lot more like a gigantic landfill that overwhelms a bushel or two of good apples. In a context where the creation, distribution and dissemination of pages can be done at nearly zero cost, the outlaws can make themselves seem orders of magnitude “bigger” than the sincere participants, even if they may be in the minority.
Without good countermeasures (like discriminating, smart, customizable research tools and search indexes), the Internet would be like “two billion channels and nothing on.”
In such a context, search engines *cannot* take meta tags too seriously. They must come up with clever methods of ranking content, and weeding out the bad stuff. They will never finally succeed, at least in this particular realm. In spite of its stellar reputation, we can see that Google’s index has its share of link farm spam, keyword spam, and so on. If somebody makes $50 in affiliate fees every time a visitor to their web site signs up for an American Express card, there is no downside to deceiving Google to get the site’s pages ranked high for the relevant keywords in order to generate free traffic. So that’s what many web site operators continue to do. As consumers, we have to cheer for the search engines and their ability to weed this crap out.
Google doesn’t completely ignore meta tags as I claimed. They may give a very low weight to keyword tags (we’ll never know how much). More importantly, the index will grab a description meta tag as a “fall-through” if a site’s home page doesn’t contain much if any usable text, nor alt tags on images. Another thing Google does in some cases is to use the description written by a human editor at the Open Directory, if the site is listed there.
Human oversight combined with a smarter way of ranking which web pages matter most to the searcher give us pretty good search results most of the time, even if the humans at the major Internet directories aren’t perfect, and even if every ranking method can be “cracked” to some extent by strategic articipants in the ongoing quest for free search engine referrals.
Given the wide open nature of the medium, it’s unlikely that there can be any sort of fixed taxonomy that classifies things “just right.” Nor can there ever be any guarantee of sincerity on the part of those who
For all of the overworked web site owners who just want to get on with the business of producing great content or selling great products, I hope you’ll someday feel less compelled to bother with those pesky keyword and description meta tags.
The very short summary to my point remains: if you’re thinking corporate portal or library, think taxonomy; if you’re thinking about creating a large Internet search index, think noncooperative game theory. No, that doesn’t mean run out and spam the engines! It simply means that the engines can’t put too much emphasis on ranking methods that are too wide open to spam. If the engines don’t pay much attention to meta tags, they might be a waste of your time.