SEO Is Mostly Quack Science
teddziuba.com
teddziuba.com
The research itself was presented in, what I (and many, many others) believe to be precisely the right fashion. We pulled page 1 results from 11,000+ search results, then looked at Spearman's correlation co-efficient for each element.
This blogger pulled out a small section of the post and disagreed with the recommendations we made, which I think is perfectly fine and reasonable. However, to discount/discourage future research or suggest an entire field is illegitimate seems inappropriate.
It would seem, for example, highly subjective to include the bit he did without also including this portion:
Methodology
- We collected 11,351 search results from both Google & Bing via Google AdWords suggest data for the various categories (you can see these keywords yourself via Google's AdWords tool)
- We looked only at the first page of results (which typically included 10 results, but sometimes contained a higher or lower number). We ignored all no-standard results (meaning universal or vertical results such as video, images, local or "instant answers")
- The correlations relate to higher/lower positional ranking on page 1 of the search results
- We controlled for search results where all (or none) of the results matched the metric. Thus, for example, if we were looking for correlation with .gov domains and no results in the set included a .gov domain, we didn't use that SERP for that dataset.
- We've used Spearman's correlation coefficient, as it is the standard (and in our opinion, best choice) for ranked datasets. You can read more about this selection via Ben's comments here and here.
and this:
Understanding Correlation Significance
The correlation numbers we show range between -0.2 and 0.35, where a perfect correlation would be 1.0 and no correlation would be 0.0.
The standard error for each result set is also included, but tends to be so low in most cases that displaying it on the bar graph would make it nearly invisible. This is thanks to the large number of results collected - we've got very high confidence in the statistical significance of these.
Correlation ≠ Causation
It's long been held in statistical analysis that even very high correlations do not necessarily mean one data set is the cause of the other. People holding umbrellas don't cause rain. Ice cream sales don't cause hot weather.
For example, the more I wear suits, the more I speak on panels about SEO. Does it therefore follow that wearing suits gets me onto panels about SEO?
It's critical to know that the data below, like data from other types of SEO tests, requires careful consideration and analysis. Parsing a bigger correlation as a direct sign that one should do X or Y more would be a fallacy.
I guess I'm just a little frustrated because this topic has been discussed over and over again, yet old stereotypes and ad hominem attacks still pervade.
See previous pieces:
http://news.ycombinator.com/item?id=1402544
http://searchengineland.com/an-open-letter-to-derek-powazek-...
http://searchengineland.com/from-my-inbox-more-defense-of-se...
I also think the author chose the wrong article to pick on. There are so many poorly written and poorly researched pieces of advice on SEO out there - the article they chose isn't one of them.
Fundamentally speaking, SEO is about three things:
1. Making an information architecture that makes it easy for spiders and humans to find what they want (301 redirects, good top-level navigation, sitemaps, clean URLs, etc...);
2. Sending strong contextual signals to spiders about what a page / piece of content is about through anchor text, alt attributes, good title tags, and basic good web design;
3. Strategically building links to key pages.
That's it - SEO in its totality isn't junk science, it's just that the hair-splitting crap that goes on at places like SEOMoz is just that: hair-splitting over minutia that really makes no practical difference at all.
You largely echo this sentiment of the article. However you're missing on 2 very important points IMO.
One is that there are hidden algorithms used to determine where a page ranks in the SERPs. By testing and iterating one can determine sufficient detail to game the algorithm and produce better rankings. This is science. Hypothesise, test, analyse, improve, iterate.
Two is that a minute difference in page "score" can be a lot in terms of rank and hence expected return assuming visits can be turned to sales.
Certain SEOs can and do acquire and maintain top rankings. They're not doing it by doing the same as everyone else, the algo's shift and these SEOs maintain their positions.
Those minutiae you're hair-splitting are like deviations to the fine-structure constant, the results cascade to greater final effects.
The point of the article is that the people who conduct this type of research get the science wrong more often than not.
The real difficulty is in conversions. For example, I'm looking for a full-time job in my recently-adopted homeland of Poland. I know that I am a kick-ass programmer, but the fact that I struggle with the Polish language means that I'm un-optimized.
I'm now working on the language side of things, and yeah, you'd better expect that 12 months from now I'm gonna kick a metric shedload of ass!
So, take the same view of your website - what's wrong, what doesn't convert, and why - and you'll be sorted.
All it takes hard work, but it is by no means insurmountable.
I have often thought that one of Google's great coups in marketing is convincing many programmers that there algorithm, unique among all others known to man, has moral authority. Nobody ever says that about travelling salesman heuristic approaches... Maybe this is due to the pervasive misconception that search engine algorithms actually measure quality of content, or that they could.
A very optimistic person could say that the search results are due not just to Google's algorithms but the cooperation of Google's algorithms and webmasters optimizing for them, and that without both of these elements, the Internet would be a much poorer place than it is currently.
One example is meta description tags. That is useful for site owners (good snippets increase CTR), useful for users (good snippets make the search results magically better by telling you what you're going to get before the click), and useful for Google. Good meta description tags only exist because of SEOs and do not follow under the frequently argued "Meh, that is just a common web usability guideline" shibboleth. To the best of my knowledge no common user agent surfaces them anywhere users can see them. They exist because they're used for search engine result pages, and that's all.
Research intended to "show the validity" of a hypothesis, rather than to disprove it.
Of course, that statement only applies to the items looked at in the research. We don't, for example, know how the two engines may value different links differently.