• 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

Finding the Perfect SEO Formula

Posted by Michael Martinez on August 15, 2007 in SEO Theory

People who rely solely on links for rankings put themselves firmly in the last possible competitive rank on the race track. They rarely take the lead and seldom keep it for long. Why is that? After all, when SEO pundits try to analyze successful Web sites, they always find thousands and thousands of backlinks. Surely those links help with rankings, right?

Yes. And no.

If a Web site has 10,000 links the odds are pretty good that there is at least some variation in the anchor text. And there are usually plenty of deep links as well. In a natural link profile there is a LOT of variation in the anchor text, so you don’t get the benefit of 10,000 links pointing at the same page with the same anchor text.

SEOs who try to analyze why sites like Adobe, CNN, and Google dominate certain queries inevitably make the mistake of treating all links as though they all pass PageRank and anchor text, as though they all point toward the root URL, and as though they all pass the same anchor text. Spammers might do that but natural link profiles look very different from that formula.

There is no doubt that Adobe ranks first for “click here” because of links pointing to Adobe’s most famous download page. But that always happens with a link bomb and the Adobe effect for click here is as much a link bomb as the George Bush effect for failure in the white house.

Many a so-called expert (including more than one who is tightly intertwined with the SEO conference circuit) has told me that “click here” proves that you have to rank through links. That’s why I so often talk about SEO idiots. But enough about the people who don’t know how to do basic search results analysis. They still make good money through the SEO conference conduit.

I’ve said it before and I’ll say it again: Just because you CAN rank through links doesn’t mean you HAVE to. In other words, there is more to a search engine’s ranking algorithm than looking at links. Yeah, you can overwhelm the algorithm with link anchor text. Some people are very well equipped to do that. But most family cars are not equipped with nitrous oxide so they aren’t likely to go shooting down the highway at more than 100 miles per hour.

Nor are most SEOs and business Web site operators equipped to generate hundreds or thousands of spammy links that all point home with the same anchor text.

So the general rule of thumb is that you don’t need many links with link anchor text in order to influence search results, but links alone don’t guarantee a first place position unless you have a LOT of links (and I mean tens of thousands or more) agreeing with each other.

Even Google will allow a page to rank on content if the page includes the keywords several hundred times.

SEO contests are a waste of time because they are not really search engine optimization contests. They don’t test your skill or knowledge of the art of optimizing Web pages. SEO contests are really just link building contest because everyone who wants to seriously win just tosses all their link placing robots and inventories of link boosting pages into the mix. Some peopel will compete on the social media circuit and ask friends to link for them.

That’s just not search engine optimization. In fact, there is nothing optimal about pointing 500 links at a spammy page, as the links won’t pass value forever.

To do it right, to keep content ranking for the long term, you’ve got to come into the mix with both content and links. You can lead with links if you’re lazy but sooner or later you have to work with content (the only alternative being an endless battle to get more links into place than the search engines can delist).

Since link chasing is not search engine optimization, you have to really understand what the search engines are looking for in order to work with real search engine optimization. I would guess that more than half of today’s SEOs — including 95% of all the bloggers and “me too” commenters in the top ten or twenty SEO blogs and forums — couldn’t optimize a page if their lives depended on it.

They depend on links. They live, eat, and breathe links. They link bait. They thrive on building meaningless traffic out of curious people who have nothing better to do with their time than visit “link worthy content”. They spam. They cut corners, cheat, lie, and otherwise pretend to be doing search engine optimization.

Some of these people are pretty good Web marketers, but they’re very bad, inexperienced SEOs. And most of them have been involved in the community for years.

Now, the reason why they are bad SEOs is that they look for formulas. They like techniques. They don’t think through the consequences of their choices. They share every tip and tidbit that they can find without regard for the damage they are inflicting on the Web, the search indexes, and the SEO community.

Formulaic SEO is a formula for disaster. Sooner or later you lose because the formula gets out, is replicated by hundreds or thousands of other mindless, unthinking SEOs, and you collectively build a huge footprint that creates a statistical black hole in the search results.

And the search engines don’t need much help to identify a statistical black hole. In fact, I would say that even Google is capable of identifying a statistical black hole (and I only pick on Google — bless their hearts — because Google remains the easiest search engine to game). We’ve seen several massive updates since the beginning of 2005 where Google just wiped the slate clean for thousands of spammy Web sites.

That’s the price you pay for playing the formulaic SEO game.

No formula lasts forever. I’ve tried them all, one way or another, and I’ve watched them all fail. Today there are cute little spammers who buy worthless spambot software that drops links into forums for them. I delete these links on site. A lot of these idiot wannabe spammers cannot even configure their messages properly and they break all their links. The non-English speaking idiots compose messages that are so garbled a marijuana junkie could realize there is something wrong with the robot-seeded crud.

Spambot freaks probably swear by their software. Most of them swear at it in their frustration because they cannot keep the links coming fast enough to please the search engines. What’s funny is that with just a little thought these same idiots could generate a huge torrent of original content that would pass muster for the search engines and allow them to generate all the ad revenues they want. But these guys don’t think, they just read testimonials and don’t ask for proof that the testimonials are real and they run their scripts and plague forums with useless links.

White hat SEOs can be just as bad as the spammers, however. They fall into two camps: those who understand how machine-measured relevance works and those who don’t. Those who don’t understand machine-measured relevance just say you should write “good copy” and not worry about things like title tags, page URls, bold, and italics. Not coincidentally, these folks tend to do well with long tail optimization.

White hat SEOs who understand machine-measured relevance walk a fine line. They are algorithm chasers just like the link spammers but they are actually on to something worthwhile. When all else is said and done, the search engines are more interested in content than in links. And content spammers know you don’t need to use bold and italics, you just need to repeat your keywords a lot.

SEOs who talk about “keyword density”, “co-efficients”, “theming”, “relevant content”, and “relevant links” are just blowing smoke out their ears. They don’t know what they are talking about any more than the link-dropping script kiddies.

The first rule of measuring search engine relevance is that no one knows how the search engines measure relevance. So you can whip out your 101 probable ranking factors check list and vote for your favorite nonsense like “site wide PageRank” but the fact you use such a list only means that you don’t know how search engines determine relevance.

There is no shame in admitting that you’re just as ignorant as I am. I feel like you’ll be keeping good company with such admissions, but that’s just me.

You can easily document the on-page factors a search engine has to work with. Google tells you what their index is concerned with (even if the list begins in a 1998 paper that nearly everyone agrees no longer describes the way Google works). The IR scoring mechanisms are pretty limited in what they can look at. That’s why Google and other search engines have been playing with other ways to weight IR scores.

You can jiggle the scoring a little bit and say this week BOLD is worth more than ITALICS, but eventually you get frustrated with the whole process and wish you could prequalify the content before you score it. That really is what the search engines try to do and they are not relying solely on links for prequalification. Anyone who points to links first isn’t thinking the process through.

We all feel that older domains achieve success more quickly than newer domains. Why is that? Is it because the search engines trust older domains more than newer domains, is it because the search engines crawl the older domains more often, is it because older domains are more likely to earn links that use anchor text which agrees with content, or is it for some other reason(s)?

We don’t know. But what we do know is that anyone with an older domain that adds new content acts pretty cocky around the SEO blogs and forums. We also know that older domains do get more links faster than newer domains, when all other factors are equal. And we know that older domains with a lot of content get crawled at a prodigious rate.

But how much does the age of a domain matter? In fact, it hardly mattters at all because — as with links — domain age is not the only gateway to success. You can (and many people do) take a new domain and get it to rank fairly quickly for targeted keyword expressions.

So just because you can rank easily with an older domain doesn’t mean you must use an older domain to rank easily. Older domains are just older domains.

A lot of people like sub-domains, too. The search engines do tend to treat sub-domains as if they are unique domains. I’ve long complained about sub-domain spam but to be fair to millions of innocent Web site operators you really cannot justify taking away the sub-domain factor. On the other hand, I know of quite a few high profile free Web hosting services that offer sub-domains whose inventories are almost completely in Google’s Supplemental Results Index.

If a page is Supplemental, it’s not visible (unless Google changes its policy and decides to play fair and let all honest, legitimate, unique content have a fair chance to compete for queries). Yes, Google now seems to be indexing more content from Supplemental Results Pages — perhaps even all the content — but the Supplemental Penalty is very real. If your pages are Supplemental, your content is penalized in the search results.

Furthermore, sub-domains can appear and vanish quickly if you don’t validate them with substantial, meaningful, unique content. Google does seem to look very closely at sub-domains. So some sub-domaining spammers are probably doing well, but that’s not really a good formula for SEO.

In fact, there is no good formula for SEO. Even the so-called “best practices” that white hat SEOs advocate can become suspect over time. There is often only one distinction between white hat and black hat SEO: excess.

Google doesn’t like orchestrated reciprocal linking, even though such links predate Google and often have nothing to do with gaming search engines. Yes, we invented link farms to manipulate Inktomi’s search engine — but the manipulation was not to achieve better rankings. Rather, we were just trying to ensure that pages were recrawled frequently enough to stay in Inktomi’s primary index every month. Inktomi maintained two indexes but only selected search results from the primary index, which was rebuilt from scratch every month.

The Inktomi results were limited to 110 million pages, so when the Web was seething with 500 million pages there were a lot of unhappy Webmasters. Business site operators banded together to help each other get crawled often enough that their pages didn’t lose visibility.

That’s mostly what the link farms were all about. We couldn’t help it if Larry Page and Sergey Brin were a couple of morons who didn’t understand the Web (but boy did they sure get rich trying to make a stupid idea work). Indexing content on the basis of linkage doesn’t work because linkage has never been about relevance or value. Web linkage was always used to create connectivity, especially between unrelated documents. That was the whole point of the Web, in fact: to bring everything together without regard for relevance.

If you’re going to manipulate search results on the basis of relevance you need to become very good at triggering “signals” because people who search for content don’t care how many links point toward a page. All they care about is finding the right page, and links are an unreliable indicator of relevance. That is why so many link-dependent SEOs struggle.

You cannot build enough links if you’re pointing at the wrong content. You cannot maintain customer satisfaction if your search engine insists on serving up the wrong content. People eventually get tired of seeing the same crap over and over again.

The best SEOs create their own standards for relevance and build up from those standards. Yes, you test ideas and evaluate their effects in the search results. But if you want to be really good at this, if you want to earn the rank of Master SEO (rather than remain stuck at Journeyman SEO), you have to let go of your pet ideas and theories.

SEO Masters don’t know how search engines determine relevance. SEO Journeymen think they have all the answers.

The perfect SEO formula makes no assumptions.

The perfect SEO formula uses only the right variables and factors, discarding those elements that don’t help.

The perfect SEO formula is cold, calculating, and emotionally detached.

The perfect SEO formula is the SEO Method: Experiment. Evaluate. Adjust.

But you knew that already, didn’t you?