Best SEO Blog|Launch your site.

There is growing demand for SEO analysts. Unfortunately, the majority of information you’ll find on the Web about search engine optimization and SEO analysis is outdated, incomplete, or just plain wrong. Anyone wanting to be an SEO analyst should understand there are no standards in the industry and you’ll find a plethora of competing opinions on what is right.

Here are a few tips on how to evaluate the SEO information you find. You can use these tips to improve your SEO analysis skills.

• Ignore any advice that focuses solely on PageRank, Links, or Authority.
• Question the value of any case study which does not provide historical data.
• Assume that any explanations about recent changes in search results are wrong

Ignore link-explicit advice
Search results are determined on the basis of many factors. If you’re only concerned with Google, you need to develop a balanced approach to three basic aspects of search optimization:

1. PageRank
2. Trust
3. Hypertext Matching

PageRank is the ticket that gets you into the top ten search results in competitive queries. Google will not allow the most relevant documents to reach the top of search results if those documents lack PageRank.

Trust is the ticket that gets you into the top ten search results for certain hyperoptimized queries. Queries become hyperoptimized when people rely extensively, perhaps even solely, on links to achieve high rankings. Trust is almost certainly tied to links but may also be tied to historic behavior, including outbound linking patterns, inbound linking patterns, absence or presence of malware, use or non-use of sneaky techniques, and other practices. Good behavior may earn Trust; bad behavior may lose Trust.

Hypertext Analysis covers many topics but it can usually be broken down into on-page repetition and emphasis. It can also include on-page repetition and emphasis for linking documents. That is, you may improve your search results by obtaining links from documents similar to your own in content, structure, and subject.

Question the value of case studies
SEOs use case studies to make convincing arguments about all sorts of points. The problem with case studies is they typically don’t include verifiable raw data so that people can check the claims being made. A good SEO case study should provide historical data and explain how a trend developed as a result of optimization. You cannot have a trend with less than three data points, so any case study that only shows you beginning and end results is pretty much useless.

Assume algorithmic explanations are wrong
In more than 10 years of studying and interacting with the Search Engine Optimization community, I have never yet seen anyone correctly explain a major search algorithm change in less than six months. John Scott came the closest to achieving that dream in 2004 when he proposed that Google’s so-called Sandbox Effect might be tied to links.

A lot of people through the years have tried to explain how Google works. You can safely ignore all PageRank tutorials, all ranking factors documents, and all single-factor theses. The PageRank tutorials are filled with technical errors and bad math. The ranking factors documents do a good job of collecting opinions and showing consensus but they are not informed. And single-factor theses are almost all obsessed with some aspect of linking.

Every time a major algorithmic change occurs people crank out the crank theories (usually just rehashes of failed crank theories from previous algorithmic changes). The most respected minds in the SEO community generally make fools of themselves trying to dissect these changes. You’ll do better to stick to the basics than to try to adjust what you’re doing on the basis of speculative explanations that have statistically low success rates.

In short, the best SEO analysts take everything with a grain of salt, look for hard data, and invest a lot of time in testing hypotheses and looking for patterns. If you’re good at pattern analysis you’ll have an advantage over the average SEO. Most of your conclusions will probably be wrong (that’s a universal rule, nothing more — there seem to be no exceptions). You have to learn to become detached from your ideas. If you become emotionally tied to an idea you will be obsolete within two years.

You’ll do better to analyze and question the anecdotes people share than to rely upon their opinions and explanations.

You’ll do better to try to replicate successes than to assume that anything will be successful just because someone said it worked for him.

You’ll do better to just stick to the basic stuff for your live sites and leave the experimentation for special sites you don’t mind compromising.