Actually the magic rests mostly in the content. Yes
description meta-tags, and so forth are important to help weigh keywords, but their effect is far less than people think. The reality is that they are clues and not triggers.
In 2008, people paid little attention when Google released scholar, but in reality, this was an important event that effects us today. This is because we all deal with scholar today without really knowing. Here is what I mean.
Scholar, when released, was a citation analysis and index engine. For some, citations are links. Boy are they wrong! Citations are found in everything and that is what Google realized with scholar. Scholar as an engine, once perfected, was rolled into the general set of algorithms that Google uses. In this case, the
description meta-tag became clues as the engine analyzed the meaning of the content, the topic, any similarities to other works, author recognition, and so on. In fact, the Google
author tag and
Google+ are extensions of the citation analysis to help remove some ambiguity. These elements came out of the Google TrustRank exercise where known quality sites were manually analyzed and used for comparison for other sites. This was about when citation analysis was rolled into the general algorithm as Google realized that just a seed was needed for analysis and not ongoing manual processes. For citations, the seed already existed in scholarly works and any manual adaption could be made by members themselves.
Let me give you an example. I write a scholarly piece that gets very little attention, but is indexed in Google. Let's say you find this work and quote it without linking, mentioning the author, and so on. What will happen is that Google will recognize the quotations as being from my work and a citation is born. A relationship exists where the content of each piece can support each other and similar works can be found. To further the example, say another person writes a piece with no quotes, links, author, etc. An analysis of the work can determine that the topic is related to other works using keywords, phrases, and even structure and language analysis. Google can recognize the work of an author by looking at other know works. So even if an exact match cannot be found, sometimes a citation can be made using other analysis to identify an author, a period, level of expertise and so on.
In a manner of speaking, that is what happens with
description meat-tags, and so forth. If the keywords and phrases are a match, then they support the weight of the topic keywords, and phrases for indexing. And if they don't? Nothing. Because these things are following citation analysis algorithms, the
h1 tag, and
description tag should support each other as works to be related but not match entirely. It is in this way that keyword usage works in SEO. They are mini-works. They are clues. They are not triggers. They all have to be supported by the page, the site, and most importantly organic links which by the way are also analyzed as to be natural or not. Keywords and phrases in organic links are weighted very high simply because they are unscripted and can support a topic in a narrow and focused way.
Now let me give you another example. A scholarly work was published some time ago and my name was on it with others. No other identifying information. My name is somewhat common. I do not use the author tag, or use Google+. I do not participate in any social media at all except for this site, and yet, my site got a sudden boost of about 10,000 extra page views a month and tapered off exactly as the accesses and popularity of the scholarly work tapered off. Google knew who I was by name and previous works and knew which site was mine of course by Google Webmaster Tools which I think was part of your question. Spooky huh?
How clear was that? Did I do a good job of explaining it? At least was I entertaining?