Online, there is always someone searching for something; learning how to position your content, site architecture and landing pages in search engines is the premise of SEO.
Search Engine Optimization is built on the foundation of information architecture and information retrieval. As an SEO practitioner, you can gain immense insight for optimization from speaking their language, by lifting the hood to find out what makes search engines tick under the surface. This implies mastering the metrics responsible for creating and fine-tuning relevance; since those metrics are algorithmically responsible for delivering all relevant natural search results.
When a query is executed against the stored data, a facet of inverse document frequency uses an [is/is not] shingle level analysis (segmenting data like a game of concentration to find exact and broad match portions of keyword/query data).
In an overly simplified summary, IR (information retrieval) works by breaking data/information into segments and then matching the query with the fragment of data or document most suited as a direct-hit or correlation to the query as a contingent of relevance and relevance score. There are multiple weighting schemas utilized, some of the more popular are the vector space model and term discrimination to parse, assign and sort documents for relevance.
Search engines often cluster probabilistic thresholds for keywords when a suitable amount of topical content is found within a global profile on a website. Typical aspects are known as nearest neighbor queries (NN) or probabilistic nearest neighbor (PNN) level algorithms .
In other words, if enough content/nodes exists, and overlap with multiple related objects (your content), when one node/keyword/mention acquires authority (i.e. 75-100 occurrences, pages, internal links, inbound links or citations), relevance from one node to another will spill over into other topical areas and rankings/from retrieval from the index can emerge that cross-reference the gap between the two or more keywords, documents, etc. i.e. the process of stemming.
In other words, as you add more topical information of a subject, the threshold of relevant data assigns a profile to your pages and your website for containing a specific degree of data on the subject or topic.
Herein is the basis of adding topical content to reinforce continuity on a subject as well as create a broader array of destinations for query space inquiries. For example, if you search for “purple snafenhalgers” and everything that is not purple or contains any of the key shingles comprising the cluster re-cog-nized as “snafenhalgers”, is instantly eliminated; particularly, if no reference or citations exist to solidify the existence of the sought data to match against the index.
Search engines are programmed to retrieve the most relevant document and based upon the on page and off page SEO factors your pages exhibit. Search engine optimization uses components such as coherent title tags (that summarize the contents of that page), h1/heading tags, multiple occurrences of singular and/or plural synonyms (containing a shared subject matter) and the internal and external link graph both to and from that page; referenced within a site or from other websites determines the degree of authority a page or website has for a said occurrence/term.
It is an aspect of global term weights, the relevance model /weighting scheme and page level thresholds, but inverse document frequency otherwise known as IDF  parses both page level and site level occurrences of keyword occurrences when looking for an ideal match.
For example, if you have multiple pages on “search engine optimization” and multiple pages on “marketing or marketing tips”, depending on the correlations used and the query used (such as “search engine marketing tips”), your page could be returned as a direct hit.
Internal links and virtual theming are an effective method to assign a specific pecking order for specific keyword queries. Aside from that, link and domain popularity, click through data and the volume of content or inbound links to each page can shift search engine ranking factors in your favor.
For instance 300 pages on a topic, all internally linked to a landing page, as well as deep links to each of those pages are enough to produce tangible shifts in rankings. On the contrary 20 pages could be enough (depending on the query, industry size, how many quality links are required to cross the tipping point).
Considering that only the top 1000 results matter, there are always websites with a clear advantage; based on when they started optimizing their content, and how much time they have had to cultivate their collection of documents and popularity on the subject.
Google organic search results rewarded topical relevance by adding a plus sign accompanied by a double ranking or third ranking in the top 10 results stating “show more results from” with the search result (thus indicating a high concentration of the said terms).
At the end of the day, in order to dominate the SERPs (search engine result pages), you need relevance and popularity. Despite the obvious, built more content and links takeaway, there are other more significant undertones to this article.
To summarize, there is no need to recreate the wheel, study what works and find out why (such as reading a few papers on information retrieval) and then when something works, you know why and how to reproduce it.
SEO is an ever-changing endeavor, by studying the principles that support it, you can gain significant competitive advantages and apply them to a content management system, your approach towards content development or your internal or external linking practices.