SEO workers tend to believe that they are able to tell why page B is # 11 at Google for a quite
competitive search query and page A is # 10, and not vice-versa. The sad truth is that this is practically  impossible. Indeed, if you compare page C at # 200 with page A at # 10 you can often point to a major difference in the number of inbound links, the quality of the metatags, the age of the site and other observable factors that are known to be measured by the ranking algorithm while determining the relevance score of the page and, hence, its ranking. 

However, while comparing two adjacent pages at or close to the top of competitive arenas the picture is very often reversed - the better ranked page has less inbound links, the metatags seem to be written by a careless designer rather than a pedantic SEO expert, the text is too short and so on. How can that be?

The answer is one SEO workers like to ignore: the search engines are very successful in keeping secret most of the data we need. We do not know the real relevance scores of each page and cannot measure the results of our actions and experiments; we do not know the weight of each 
parameter in the ranking algorithm, we do not know the real value of each inbound link - some might have a negative value and some might be worth each hundreds or thousands of ordinary links.

Nevertheless, such "explanations" also serve to "calculate" the price of SEO projects, another fallacy of the trade.