• 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 Theorem of Query Space Optimization

Posted by Michael Martinez on October 30, 2007 in SEO Theory

The Theorem of Search Engine Optimization says that achieving optimal performance from search engine results diminishes the naturality of search results.

That is a very empowering statement. For a search engine, it means helping users find the content they need despite the content’s poor organization and presentation. For a search marketer, it means you have the ability to recognize when someone has stepped in and organized the search results.

The search algorithms need to embrace optimization without letting abuses run free. Search optimizers need to recognize when their efforts are overpowering the search algorithms. Even black hat SEOs have to stop and think about the consequences of their actions. Or, rather, it makes economic sense for black hats to limit their tactics.

Search engine spammers are Once-lers making thneed pages. They really serve no purpose other than to generate money for the Once-lers of the Web, but the search spamming process can so pollute search results that all the fish have to get up and walk away. That is, we need to clean house every now and then because the spammers cut down all the Truffula Trees.

It’s in a spammer’s best interest to create just enough content that produces money without tripping filters. It’s in the searchers’ best interest to find as little spam as possible. What might work better for everyone is the adoption of what we could call green SEO tactics that emphasize value and longevity. You can still produce revenue-building content in volume but you don’t have to constantly replace that content the way you have to replace spam.

Of course, not all spam is revenue-producing content. Some spam is just faux link-building intended to make faux content look algorithmically normally. Since spammers outperform most people in link building, it doesn’t make much sense for search engines to utilize links in their ranking algorithms. But that’s a thought for another day.

In addition to the Theorem of Search Engine Optimization we have to look at a principle that Mike Grehan made popular, the so-called “Filthy Linking Rich Effect”. Grehan pointed out that as a Web document accumulates links, its ability to accumulate more links improves.

There are two reasons why a document’s ability to capture links improves. First, it becomes more visible through increased linkage. Second, a document’s perceived credibility is based upon the number of endorsements it receives. Those endorsements do not have to be peer endorsements, although peer endorsements are generally assumed to be a more reliable indicator of quality than disassociated endorsements.

The problem with gauging quality through the eyes of endorsements, however, is that endorsements are not reliable. For example, suppose every member of the Republican Party stood before the United States Congress on July 1, 1969 and swore an oath that Richard M. Nixon had great integrity? The credibility associated with such an endorsement would be huge right up until the 1972 presidential election, when a small group of men working indirectly for the Nixon campaign were caught breaking in to the Watergate hotel.

On the other hand, contrary endorsements are equally unreliable. Suppose the vast majority of Americans said in 1980 that Jimmy Carter was a terrible President of the United States and that he should leave office? In real life Carter was voted out of office in a landslide victory for Ronald Reagan, but today President Carter is widely praised for his humantarian activities. His reputation has improved considerably since he left office.

Endorsements are therefore only useful in a limited time frame, after which they need to be discarded. Although links are not endorsements, links are treated as endorsements by search engines and links also experience periods of limited relevance.

A link that passes value today may stop passing value at any time in the future for any of the following reasons (among other possible reasons):

1. The linking page is taken offline
2. The destination page is taken offline (this includes changing, moving page URLs)
3. The linking page is penalized or filtered
4. The destination page is penalized or filtered
5. The linking page changes the format of its link

Even when treated as endorsements, links tend to have finite contexts.

A Web document’s link profile therefore can be used as a metric to gauge shifting trends over time. If a document’s link profile grows, we know that the document participates in young but maturing queries. If a document’s link profile remains stable, we know that the document’s queries has achieved full maturity. If a document’s link profile shrinks we know that the document’s queries are either dimming or have lost all activity.

The correlation between queries and links is very strong. The more people who search for a specific term, the more people who will provide links to content for that term. The fewer people who search for a specific term, the fewer new links that appear for that term.

A Web document’s productive lifespan may be very short or very long. The longer the lifespan of a Web document, the less active its relevant queries tend to be. We can group queries together in classes or sets of related search expressions. Each Query set defines a Query Space that is populated by Web documents which are all relevant to each other because they are all relevant to the queries (the documents may or may not have overlapping topics).

A long-lived query space is either driven by a lot of news or else it’s a low-activity query space driven by a fundamental need. You’ll see a lot of traffic for a celebrity like Britney Spears and you’ll see a fair amount of traffic for house movers. One day Britney Spears will be no more popular than Donna Douglas is today, but people will still be interested in moving houses.

The buildup and falloff in queries, content, and links defines a wave-like pattern if you graph it. The queries start first. The content starts second. The links start third. Content and links tend to peak after queries peak. Links tend to peak before content peaks (I think because late-comers create content that doesn’t generate any buzz or interest among other content creators).

Content peaks after a query space has become saturated with links and content. There is really no more need for truly useful links because all the content that the average searcher needs is already there. It becomes increasingly more difficult to promote new content to the top of search results in a mature query space until the queries drop off. Once the queries drop off content and links begin expiring faster than they are replaced. By the time it becomes easier to optimize for a query space that enjoyed a long run most of the traffic has moved on to other interests.

We can look at the performance of query spaces. A query space’s scope is determined by the number of related queries that produce a significantly similar number of documents. A query space’s longevity is measured by the number of time frames in which the query space produces at least average activity. That is, a query space becomes disfunctional or breaks down (loses coherence) after its traffic drops below the average level of traffic for all query spaces.

You don’t have to know what the average amount of traffic is for all query spaces. You only need to know what a query space’s average traffic was for the past two years. If you see a decline in traffic then you know the query space is degenerating. If you see peaks and valleys where the peaks seem to about the same you’re looking at a mature query space that is still productive. If you see an increase in traffic over the past two years then you’re looking at a young query space.

A query space is coherent as long as it produces relevant results. A query space’s coherence degenerates as traffic declines and/or as relevant content declines. That is, spam can kill a query space by making the query space less useful or productive for searchers.

Hence, the Theorem of Query Space Optimization tells us that queries remain productive only as long as there is search interest in them, and only as long as relevant content is promoted for those query spaces. In other words, a query space exists only as long as there is interesting content populating the query space, which means that a query space can actually become self-sustaining if and only if the query space produces continually new and interesting content.

In short, creativity is king.