• 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

Building trust amid SEOs’ paranoid dementia

Posted by Michael Martinez on October 25, 2007 in SEO Theory

The latest brouhaha over Google’s so-called PageRank Edit underscores the dichotomy of SEO Toolbar mentality: Toolbar addicts can’t live with Toolbar PageRank and they can’t live without it. The truth is that many SEOs cannot live without some sort of artificially validating metric.

People who look at the Web through a toolbar place themselves at a competitive disadvantage in several ways. First, they think only in terms of an undocumented third-party evaluation. How many times have you seen someone say in a forum or blog, “I only want links from PR X sites”? How many forum posts have you seen start out with, “I have a PR X site in a competitive vertical …”?

Second, they disassociate themselves from the content they are optimizing. Toolbar PR is not about content, topicality, verticals, or relevance. It’s just an automated assessment of a Web page’s possible importance compared to the rest of the Web. Google even admits the assessment is not based entirely on linkage (although that has been an on/off admission).

Third, the Toolbar-toting crowd worry about what the toolbar may be doing to their search visibility, their rankings, and their content’s quality reputation. We can speak of quality reputation as the perceived quality or value of a Web document as determined by external criteria not directly associated with the document. Hence, no one wants to associate with a Web site that doesn’t have the right Toolbar PR, AlexaRank, or whatever.

The quality of a page can be determined by any one or any collection of criteria. You as the judge of quality are free to choose and define all the quality factors you want. But the larger community of Web surfers tends to look at page quality in a relatively small number of ways.

For example, surfers want to know if a page is safe. Will this page harm your browser, cause your browser to download malware, or draw you into some sort of Federal sting operation you never would otherwise fall prey to?

Surfers also want to know if a page is relevant. Does this page include any information regarding the topic you’re currently interested in? People click on links to learn more about something. It doesn’t matter if the links are provided by a search engine or a typical Web site. If someone clicks on the link, they are looking for information.

And surfers want to know if the page is informative. Does it answer their questions? If a page is authoritative in tone, it stands a better chance of changing minds or clarifying opinions than if the page just rambles on randomly. But most users cannot discern between low-quality (unreliable) and high-quality (reliable) authoritativeness.

So another important measure of quality for most people is page credibility. One measure of page credibility is links from other Web sites. Bet you thought I would never say that. But it’s true. People do very often look at links as if they are indicators of reliability, which is why the United States government does not endorse Web sites through links. On the other hand, links are not reliable indicators of opinion.

You can often find opinion expressed through the text surrounding a link. So a semantic analysis of link-associated text could be boiled down to three values: recommends site, does not recommend site, and no opinion expressed. But you still have a problem of credibility. After all, what is to present me from creating 100 domains where I link to SEO Theory and write 100 variations on “SEO Theory is the geratest SEO blog of all time. You absolutely MUST read it!”? And if I do that, then what is a search engine to do with the distilled recommends site value it could semantically extract?

Credibility cannot be measured algorithmically. One liar can recommend another liar. One fool can endorse another fool. One scientist can point to the work of another scientist, establishing a baseline that is adhered for a generation until some future scientist shows the first two were full of flawed logic. If we as people do not know how credible endorsements and negative comments truly are, how can we possibly create software that makes such determinations for us?

A lot of work has gone into studying how we determine which Web sites are credible. But even these studies cannot distinguish between faux expertise and real expertise. Pick any Wikipedia article at random and tell me whether it was writte by experts or faux experts. You may be able to make a guess but how good is your guess?

For example, the current Wikipedia page on link farms starts off with a very erroneous statement. How many of you know what that erroneous statement is? If you’ve read about link farms on the Web, you’ve probably read many different definitions. This particular Wikipedia article is one to which I contributed. You can see my comments in the discussion page. I clearly explained how link farms were developed and why. Nonetheless, the very technically correct language that I inserted into the article months ago has been replaced by very technically incorrect language.

Wikipedia’s policies on credibility and sourcing are very confusing. In fact, its system is designed to favor consensus over correctness (for example, “If you seem to be the only person who feels that the article should be the way that you have made it, perhaps it is better the way everyone else thinks it should be”). When the only qualified expert on a specific topic is not allowed to be the authoritative voice on the topic, how reliable and credible is the article? Not credible or reliable at all.

Imperfect systems and mechanisms therefore deter credibility on the Web. That is, you cannot trust what you fin on the Internet because you have no way of knowing just how many blind fools are following other blind fools. How many times have you seen the “Good Times” virus warning? How many times have you seen the “Nieman Marcus cookie recipe”? How many times have you been told that someone’s brother-in-law is a lawyer and he said that it was perfectly legal to forward an email asking you to send money? How many times have you received an email asking for a lnk exchange because it builds PR and is good for both your site and the other site?

Credibility is an expression of trust. We think a page is credible and therefore we trust it. In search engine optimization, however, trust is one commodity that is in rare supply. Almost as soon as someone puts out a great new idea for capturing links, someone else is warning you that it’s not ethical, that it no longer works, that it’s cheap, shlocky, and unprofessional, etc. Almost as soon as someone claims they have cracked Google’s algorithm someone else is claiming the first person is an idiot.

There is a fairly small cadre of SEO bloggers and consultants who all support each other. Reading their blogs is like looking for physical variation in homogenized milk. You’ll only start picking out discrepancies when the milk turns sour — when the SEO collective thought-leaders have a falling out amongst themselves. They usually divide and argue over new ideas, over the meaning of significant events, or over the ways they express their otherwise homogenous opinions. After a brief sputtering period of angst and hostility, you usually see the cadre close ranks and go back to telling everyone that SEO is all about links.

How credible is that cadre of SEO thought leaders? Extremely credible, if most of you folks haev been telling me how you honestly regard those people through the years. But an equally important question is, how right are they on any given topic? Their collective random chance of being right is about 50% — which is to say that it’s no better than your random chance of being or my random chance of being right.

And the reason why the collective SEO wisdom is no better than guesswork is that no one actually tests what other people say in the community. Rarely do you see a radical new idea challenged, questioned, probed, tested, or challenged in this field. In fact, most of the radical new ideas are spread through lower-tier SEO networks that push the ideas up to the thought leaders through grass roots induction. That is, the SEO community thought leaders usually identify trends and popular ideas and help shape the messages that advocate those trends and ideas.

But no one actually bothers to conduct a scientifically or statistically valid test to determine whether the ideas are good or bad. So about half of all SEO ideas are probably good ideas, and about half of them are most likely bad ideas — at the time they are proposed.

On the surface that may not sound so bad, but it diminishes the value of the trust we place in each other. So, as a community that is obsessed with trust and picking friends and allies to help us advocate our causes, we do a pretty bad job of establishing the value of our criteria for trust. Hence, typical SEO advice on how to choose good linking partners, good methods for constructing content, good methods for search engine optimization in genral is really no more authoritative or reliable than pulling ideas out of a hat.

Which is not to say that the SEO community is a group of idiots. We’re people. We form opinions. We like some folks and we dislike others. But when we’re algorithmically challenged to determine how search engines trust content and link, we’re at a severe disadvantage. Let me put it this way: when we want to influence a search engine’s measure of trust and credibility, we’re pretty limited in our ability.

If we were all equally fair-minded and altruistic about the Web, that would be a very bad thing because we would all be working to help promote the best content for every topic and the search engines would not be listening to us very well.

If we were all equally self-centered and selfish, that would be a very good thing because we would all be promoting ourselves for our own benefit and the search engines do listen to us to some extent.

In this kind of game, self-promotion wins out over altruism every time because self-promotion is always grounded in solidarity of self-interest and collective altruism is always grounded in consensus of opinion. It only takes one self-promoter to destroy the equilibrium of altruism. It takes nearly everyone’s participation to make altruism work but altruism by itself does not ensure quality or legitimate expertise.

So let me boil this down to a few points:

1. We have imperfect criteria for determining which pages are credible, authoritative, and expert
2. We have imperfect criteria for determining which pages are self-promotional and altruistic
3. We have imperfect criteria for determining which links are recommendations for, recommendations against, or abstentions in opinion regarding other pages
4. We have imperfect criteria for determining which pages are completely informed, partially informed, or uninformed

Through these imperfect criteria we must build relationships, promote Web sites, and compete with each other for search visibility and dominance. The competitive advantage lies with those people who have the more perfect criteria (although no one will ever have perfect criteria).

So, if we were playing a game, each player would have to choose from a set of broken tools. Which broken tools would you trust more? The broken tools that hinder you the least or the broken tools that help you the most? How you value those tools is tied to whatever strategy you’ll apply toward winning the game. One of the tools every player has available is the freedom to make his or her tools available to everyone else.

In doing so, we open up a new set of questions for each player. That is, each player chooses tools that hinder least or help most but then is offered tools that hinder least or help most. If you went with least hindering tools to begin with, how much would you trust the tools that help most when offered by other players? Their recommendations are going against your better judgement, so whose judgement is better?

In search engine optimization, you have two players: search engine and optimizer. Each player would benefit if they both worked together. Neither player has any incentive to work with the other (because the benefit of taking advantage of the other player outweighs the benefit of cooperation). Each player gets to choose from tools that hinder least or help most. Each player is free to offer his tools to the other.

If both players select tools that hinder least, their cooperative strategy will be less effective than their competitive strategy because in a competitive strategy each player can take advantage of the other. The least hindering tools make it easier for them to take advantage of each other.

If both players select tools that help most, their competitive strategy is less effective than their cooperative strategy because in a cooperative strategy each player is better equipped to help the other.

If one player selects tools that help most and the other player selects tools that hinder least, the player with the least hindering tools must choose a competitive strategy and the player with the most helpful tools must choose a cooperative strategy. If either player chooses the wrong strategy for his tools, the other player wins.

In a situation like this, non-disclosure becomes a competitive advantage. That is, if neither player knows which tools or strategy the other player has chosen, both players are forced to play conservatively in case their guesswork is wrong. The first player to figure out what the other player is using and doing can devise a counter-strategy and move ahead.

If one player knows which the other player is using, or if one player knows which strategy the other player is using, then the knowledgable player can either change tools or change strategies. But even if the knowledgable player changes to most helpful tools and a cooperative strategy, he stands to benefit more than the other player because he is more likely to take risks.

Competitive knowledge therefore increases risk-taking. Hence, when you look at a Web page, determining how credible, authoritative, expert, reliable, informed, altruistic, and popular it is helps you shape your strategy for dealing with that page. You can either be cooperative (link to and help that page) or competitive (try to outrank it in the search results).

That is, you play the “search engine versus optimizer” game against other optimizers in an attempt to second-guess how they are playing against the search engines. You play the “search engine versus optimizer” game against the search engines in an attempt to second-guess how they are playing against your competitors.

So we can call this the Two Competitors With Two Tools and Two Strategies game, because in the end you have to decide if you want to cooperate with or compete with your SEO competitor. But you also have to decide if you want to cooperate with or compete with a search engine.

How do you compete with a search engine in SEO? Take Wikipedia. How many SEOs hate Wikipedia? Most feel at least some competitive frustration toward Wikipedia, in my opinion, because Google tends to favor Wikipedia above more useful and reliable content in search results. Hence, most SEOs probably find themselves competing with Google to promote Web sites above Wikipedia.

The search engine may or may not know what tools and strategy the optimizer is using. The optimizer may or may not know what tools and strategy the search engine is using.

The better informed each player is about the other player’s choices the better the informed player’s chances of wnining become. You CAN win by helping other people if THEY are helping other people. You CAN win by competing with other people if THEY are competing with other people.

I’ll come back to this concept in a future post.