• Register
• Log in

SEO Theory - SEO Theory and Analysis Blog

Algorithm analysis, Web community relationship analysis, SEO practices and techniques, industry news, etc.

• Home
• About SEO Theory
• SEO Glossary

The Wikipedia Principle - How We Devalue Web Content

Posted by Michael Martinez on October 31, 2007 in SEO Theory

The Theorem of Search Engine Optimization tells us that generating optimal performance from search engine results diminishes the naturality of search results.

The Theorem of Query Space Optimization says a query space exists only as long as there is interesting content populating the query space.

As the World Wide Web becomes a more significant communication medium we will see an increasing reliance upon new media channels of information. The concept of the television network won’t translate very well to the Web but search networks will become front-place resources. A search network is not simply a search engine. The search network is part of the medium, part of the content. For example, Xenite.Org’s site search is now part of Microsoft Live’s search network. The distinction between my search function and my content is blurry because services and tools are content.

Search engine optimization is not concerned with what makes content interesting, but it depends on interesting content to acquire value and establish relevance. Search technology has to identify relationships between interests and content. We as optimizers seek to help the search technology accomplish that task but as Web marketers we’re more concerned with winning the competitive race to place our content in front of our target audiences.

There is a direct conflict of interest between our cooperative goal of improving the quality of search results and our competitive goal of ensuring that our content captures valued search visibility. You cannot objectively help improve the search results if you’re subjectively promoting your own content through those results. All the moral persuasion in the world won’t change the fact that you’re influencing search technology to provide favorable listings for your own content.

The economy of search visibility is built on three factors: interest, content, and indexing/ranking technology. The interest factor is the consumer demand for content. The content is the product of the supply channel. The indexing/ranking technology is the channel by which the supply satisfies the demand. Well, that’s the macro-economic view. The micro-economist would say that you have a high-value query space as long as you have less content than there is demand for.

In other words, competition levels the playing field because it oversupplies the market and deflates the value of the goods and services being provided. And that is why you have industry shakeouts every now and then. When the cost of manufacturing and marketing products (or providing and marketing services) exceeds the revenues they generate you no longer have an incentive to stay in the market. Some cut-throat competitors use this fact to undermine their weaker competitors, sustaining significant losses in order to bleed the weaker competitors dry.

In search engine optimization, commoditization destroys the value of the services we provide and the techniques we rely upon. The easier it is to do a basic SEO task, the less expensive it becomes for everyone to do it, the more likely that the revenue generated by performing the task will decline. If you can sustain enough volume in your exchanges of value you incur economic growth.

Volume is the spammer’s tool. For anyone else involved in search, however, volume is a problem. The more algorithmically relevant content you find in the search results the less valuable any particular content really provides. This is really why search results positioning is so important. When presented with 1,000 results, most people will just click on the first result and work their way down until they find something that satisfies their interest. If the task of looking at all the available content is too daunting, people let search engines make the choices for them.

Search technology is presently incapable of distinguishing between high value content and low value content if only because we all value content according to different criteria. And that is why benevolent search engine optimization is in everyone’s best interest. The spammers lose if the search engine loses market share. The non-spammers lose if their content is drowned out by too much spam. The searchers lose because they don’t find what they are looking for. Benevolent search engine optimization at least helps searchers find something both relevant and satisfying even if it’s not the most relevant or satisfying content.

And that is the SEO Dilemma: searchers will settle for any content that is minimally relevant and satisfying. Call it the Wikipedia Principle, where a search engine intentionally promotes low quality content that is minimally acceptable to searchers because it costs less to do that than to promote better content. At the SMX 2007 conference in Seattle earlier this year, Matt Cutts made the point that most of the people in the room were technically more savvy than the average search engine user — hence, the average search engine user didn’t know that Wikipedia content is not necessarily the best content.

He did not say, “We’re cheap and trying to make a buck by shlocking off the first available content”. He said that most people don’t know whether the Wikipedia content Google serves is accurate or fair. Hence, if it’s good enough for the ignorant masses, it’s good enough for Google because the ignorant masses click on the links and are satisfied. From the search engine’s perspective, search is all about satisfying the user. From the user’s perspective, search is all about finding the right information (or something that will do). From the Web marketer’s perspective, search is all about bringing more traffic to favored Web sites through search.

The complexity of search visibility is that none of the three groups can predict what the other two groups will do. Searchers have no way of knowing if the content they need is there or if the search engines will show it to them (a search engine may choose NOT to show relevant, satisfying content simply because that content violates some guideilne). Web content providers have no way of knowing what searchers are looking for or whether the search engines will choose to show their content to the searchers.

And search engines have no way of knowing whether searchers are looking for the content the search engines show or whether the content providers are providing content the searchers want.

This three-way symbiosis means that any change in one partner’s situation affects the other partners. If you provide good content but get banned you’re not hurting just yourself, you’re hurting the search engine and the searchers. If you ban good content for any reason, you’re not just hurting the quality of your search results you’re hurting the searchers and the content providers.

The searchers really bear little to no moral responsibility in the relationship. They do their thing and are at the mercy of the search engines and the content providers. So one would think that competition would be good for the searchers because they, at least, can take their dissatisfaction to another search engine. And searchers do that, but not enough to affect the badly compiled search market share statistics.

Which means that the decisve role in the three-way relationship comes down to either the search engines or the content providers. But search technology is simply not capable of producing the correct results from randomly collected data sources. The search engines are black boxes only in theory. In application they are complex mechanisms but they have defined paramaters and those definitions limit the search engines’ capabilities.

Google is no more capable of telling me where to find the most reliable and authoritative information on the works of J.R.R. Tolkien on the Web than it is of predicting whether we’ll have snow in Minneapolis on March 7, 2009. But though I cannot predict the weather I do have some basic familiarity with the works of J.R.R. Tolkien.

I can, actually, tell people through my Web sites where I think I’ve found good quality content on J.R.R. Tolkien. I can even tell people that I feel I provide some of that good quality content.

Google, in its limited fashion, can use my recommendations to help direct other people toward that quality content but because Google is NOT an intelligent machine it has to function in a very limited fashion — Google cannot distinguish between my recommendations and my warnings, for example. Now, I can be careful to advise Google in some crude way that I don’t recommend certain content, but the same vulnerability that subjects Google to my linking whims subjects Google to other people’s linking whims.

That is, Google has no way of knowing if someone else is as well-informed, better-informed, or less informed on the works of J.R.R. Tolkien than I am. So what is to prevent Google from giving equal weight or greater weight to someone else’s recommendations? What is to prevent Google from misinterpreting someone else’s warnings and giving greater weight to those misinterpreted warnings than to my positive recommendations?

Nothing. Absolutely nothing.

Which takes us back to the Wikipedia Principle, where “the only qualified expert on a specific topic is not allowed to be the authoritative voice on the topic” — because it’s too expensive to be right in an environment where being satisfying is sufficient.

Satisfaction is more important than accuracy or correctness on the Web. And that’s a pretty weird concept. It’s like saying we should legalize intravenous drug use for people of all ages because having free access to heroin will make them feel good. You could say that’s an unfair comparison. So let’s say that we give everyone (including children) the freedom to eat only donuts (also-spelled doughnuts). Donuts are fried carbohydrates. At some point the minimal nutritional value of the average donut is exceeded by the health risks associated with high fat and carbohydrate content.

Eating one donut is fun. Eating nothing but donuts is unhealthy, though it may be satisfying. And some people, given the choice, will eat only donuts regardless of the health risks. If you shorten the timeframe for consideration even more people will choose to eat donuts (because they can always eat something more healthy later). But the principle remains the same: satisfaction is preferred to correctness or accuracy. A satisfying meal is preferred to a healthy meal.

In order to get more people to prefer healthy meals, we need to make those healthy meals more satisfying. We have to lower the cost of preparing healthy meals AND make the meals taste better for more people (without sacrificing any of the nutritional value). That would be a significant improvement in the quality of healthy meals for the donut-loving people in the world.

Search engine content needs to be as satisfying and easy-to-make as donuts. Accurate, correct information needs to be as easy to promote as inaccurate, incorrect information. Which means that every site needs to be as well-designed and as self-promotional as Wikipedia.

You see, the Wikipedia Principle works both ways: the structure of a Web site promotes its content without regard for the quality of the content. People need to feel it’s as easy to link to your Web site as to link to Wikipedia. That means the information you provide on the topic should be at least as extensive and exhausting as Wikipedia’s content. That means a random stranger landing on your Web site needs to be able to find the information in your pages with as little or less effort than they use on Wikipedia.

This has nothing to do with brand value. It’s all about the economics of search. Every user attempt to find information needs to be satisfied. If you can satisfy those attempts with good information as easily as or more easily than with Wikipedia information, you can beat Wikipedia. That doesn’t mean just create a big long page with a lot of copy. It does mean that if a search engine has as much reason to consider your page as to consider a Wkipedia page, both pages should appear close together in search results.

And that is where the SEO Method comes into play: you experiment, you evaluate, you adjust.

When you see your page listed under Wikipedia’s page, you have less work to do than if your page is ranked 20 or 30 positions behind Wikipedia.

Your job as search optimizer is to do what the search engine cannot: recuce the cost of promoting the most informative, accurate, correct, and complete information to the point where the search engine cannot logically refuse to promote that content.

But in turn the search engine cannot impede your progress. For example, if a search engine imposes an arbitrary restriction on Web content indexing and ranking the search engine is not acting benevolently. The users are not being presented with the most relevant, complete, correct information available. They are simply being presented with information that meets a minimal level of political acceptability.

Google’s Supplemental Results Index, for example, imposes a limit of political acceptability on Web sites. They are using the Supplemental Results Index to suppress spam but in the process they are suppressing many high value, legitimate unique content Web sites. The Supplemental Results Index increases the cost of capturing search visibility, so instead of leveling the playing field Google has favored those sites with the most resources at their disposal. And that means competition for search visibility will decrease because fewer people can afford to compete at the level Google requires.

To bring this all together, we know that search optimization destroys the natural randomness of search results. As long as benevolent search optimization dominates the process that should be a good thing, and the benevolent search optimization can be provided by either search engines or Web content providers. We also know that a query space (traffic and content) exists only as long as there is interesting content that sustains the queries. And we know that search results tend to favor minimally acceptable content because of the prohibitive cost of identifying highly acceptable content.

We can therefore conclude that the life expectancy of a query space is tied to the value or quality of the content created for the query space combined with the genuine fairness of the search results. A loss of quality in content or fairness in search results significantly impacts the life expectancy of a query space. The more fair the search results tend to be, the longer the query space should survive. The more acceptable the content tends to be, the longer the query space should survive.

More briefly, the level of competition for search visibility directly impacts query space longevity because as competition lowers the cost of acquiring search visibility the quality of content and the fairness of search results declines, thus shortening the life expectancy of the query space. In other words, the more search engines and optimizers dink with a query space, the less attracting the query space becomes to searchers.