SEO Glossary: Keyword Strategy Terms

Eleventh and final entry in the running SEO glossary, closing out the three-post split of what was originally a single “24 essential terms” post. This one groups the keyword-research and semantic-search vocabulary.

Keyword

A keyword is the word or phrase a person enters into a search engine to find information, products, or services, functioning as the basic unit through which search intent gets expressed and matched to content. A keyword implies more than just its literal text: it carries signals about the searcher’s intent, the level of demand for that topic, how competitive it is to rank for, its relevance to a particular business, and its likely conversion potential. Modern optimization increasingly targets entities and topic clusters rather than individual keywords in isolation, reflecting how search engines now interpret meaning and relationships between concepts rather than matching strings literally.

Keywords are commonly grouped by length and specificity: head terms are short, broad, and high-volume (“shoes”), while long-tail keywords are longer, more specific phrases with lower individual search volume but often clearer intent and less competition (“waterproof hiking boots for wide feet”). A site targeting only head terms tends to face steep competition and vague intent matching, while a strategy built around a wide base of long-tail terms, each individually smaller but collectively substantial in traffic, often converts better precisely because the searcher’s need is more precisely defined.

Keyword Cannibalization

Keyword cannibalization happens when multiple pages on the same site target the same keyword or search intent, effectively competing against each other rather than against outside sites, which can suppress how well any single page ranks and confuse the signals a search engine uses to decide which of your own pages to show. It’s typically identified through rank tracking that shows inconsistent ranking behavior for a given keyword, with different URLs from the same site swapping in and out of the ranking position over time. Fixing it usually means either consolidating the competing pages into one stronger page, or clearly differentiating them so each targets a genuinely distinct intent; either path should start with mapping existing traffic and rankings per URL before merging or restructuring anything, so you don’t accidentally consolidate away a page that was quietly performing well on its own.

Cannibalization isn’t always accidental; some sites deliberately publish multiple pages targeting closely related variations of the same core term to capture different intent angles, and that’s a legitimate strategy as long as each page genuinely serves a distinct search intent rather than existing as a thin, redundant near-copy of the others. The distinction that matters is whether a searcher landing on either page would get a meaningfully different, useful answer, not simply whether the target keywords look similar on paper.

Keyword Density

Keyword density is an older SEO metric measuring how frequently a target keyword appears on a page, calculated as the number of keyword instances divided by the total word count and expressed as a percentage; practitioners in earlier eras of SEO commonly aimed for a specific range, often cited as roughly 2-3%. The metric is now considered outdated. Modern ranking systems rely on natural language processing and semantic understanding rather than counting literal keyword repetitions, meaning a page can be well-optimized for a topic without hitting any specific density figure, and forcing an artificial repetition rate tends to produce worse, more mechanical writing without a corresponding ranking benefit. This matches current guidance from Google’s own John Mueller, who has repeatedly downplayed keyword density as a meaningful signal.

The metric’s persistence is mostly a legacy habit rather than a reflection of current practice; many SEO tools still calculate and display a density percentage simply because it’s an easy number to compute, not because it reflects anything modern ranking systems weigh directly. A page that reads naturally and covers its topic thoroughly will land at whatever density it lands at as a byproduct of good writing, and reverse-engineering that number by inserting the target phrase repeatedly tends to produce noticeably worse, more repetitive prose without any offsetting ranking benefit.

Keyword Difficulty

Keyword difficulty is a metric, typically scored on a 0-100 scale, that estimates how challenging it would be to rank in the top organic results for a given keyword, based on factors like the authority of currently ranking pages, their backlink profiles, the depth and quality of their content, search volume, and the presence of SERP features that reduce available organic real estate. It’s important to treat any difficulty score as an estimate rather than a guarantee: the number reflects general competitive conditions for that keyword, not a prediction specific to your own site’s actual authority and content quality, so the same keyword can be realistically achievable for one site and genuinely out of reach for another regardless of what the tool’s score says.

Different SEO tools also calculate difficulty using different underlying methodologies and data sets, so the same keyword can show meaningfully different difficulty scores across tools; treating any one tool’s number as an absolute truth, rather than a relative comparison against other keywords measured by that same tool, is a common way this metric gets misused.

Keyword Research

Keyword research is the systematic process of discovering, analyzing, and prioritizing the search terms people actually use when looking for information related to a business or topic. It typically involves identifying seed keywords to start from, expanding that initial list through related and adjacent terms, analyzing search volume, assessing competition, classifying the underlying intent behind each term, analyzing what’s already ranking in the SERP for it, and prioritizing which terms are worth targeting first. The most important governing principle here is that user intent matters more than raw search volume; a lower-volume keyword that closely matches what a business actually offers, and what a searcher genuinely wants, generally outperforms a higher-volume keyword with a weak or mismatched intent fit.

A practical way to check intent fit before committing to a keyword is simply examining what’s currently ranking for it. If the existing top results are overwhelmingly product pages, that’s a strong signal the query carries transactional intent; if they’re overwhelmingly long-form guides or comparison articles, informational or commercial-investigation intent is more likely. Trying to rank a product page for a keyword the SERP has already shown Google interprets as informational is a common, avoidable mismatch.

Keyword Stuffing

Keyword stuffing is a black hat tactic involving excessive, unnatural repetition of a target keyword in an attempt to manipulate rankings, and it explicitly violates Google’s spam policies. It shows up in several forms: simply repeating a keyword far more often than natural writing would call for, dumping unstructured lists of keyword variations without surrounding context, hiding keyword-heavy text from users while still exposing it to crawlers, and inserting keywords in places where they don’t add genuine meaning. The tactic largely stopped working because modern algorithms use natural language processing specifically capable of detecting unnatural repetition patterns, and because user engagement signals tend to reveal when a page reads poorly to an actual human, even if it briefly ranked before that engagement data caught up with it.

Knowledge Graph

Google’s Knowledge Graph is a semantic database containing a very large number of entities, such as people, places, organizations, and concepts, along with the relationships between them, letting Google interpret a query’s meaning and return direct factual answers beyond simple keyword matching. It draws on structured data pulled from multiple sources across the web rather than any single input. Structured data markup on a site can help contribute alignment signals that support entity recognition, but implementing schema markup does not itself guarantee inclusion in the Knowledge Graph; inclusion depends on broader recognition and corroboration of an entity across many credible sources, not on any single site’s markup alone.

Knowledge Panel

A knowledge panel is the information box that appears prominently in Google search results, typically to the right of organic results on desktop, summarizing quick facts about a recognized entity: its name and description, a representative image, key attributes, links to official social media profiles and websites, related entities, and “people also search for” suggestions. Businesses, public figures, and organizations can sometimes claim and help manage the panel associated with their own entity, though the claiming process has changed over time and isn’t available or guaranteed for every panel that appears. A knowledge panel is the Knowledge Graph made visible: the panel’s contents are pulled directly from the same entity data described above, which is why panel accuracy problems are usually graph-recognition problems, not a separate, panel-specific bug to fix independently.

Landing Page

A landing page is a webpage specifically designed to receive traffic from an external source, such as a search result, an ad, or an email campaign, typically built around a single, focused conversion goal rather than serving as general site navigation. Effective landing pages tend to share a few characteristics: a clear value proposition stated early, one primary conversion action rather than several competing calls to action, content that closely matches the intent of the traffic source driving visitors there, minimal distracting navigation, visible trust signals, and a conversion element that’s easy to find and complete. Because landing pages are frequently the first (and sometimes only) page a visitor sees, they’re also held to the same Core Web Vitals performance targets as any other page: LCP under 2.5 seconds, INP under 200 milliseconds, and CLS under 0.1, since a slow or visually unstable landing page undermines conversion goals regardless of how well the messaging is written. A landing page is where keyword-driven traffic actually converts: the keyword and intent research covered above only pays off if the page it sends visitors to is built to close the loop, which is why landing-page fundamentals are covered here rather than left as a separate, disconnected topic.

Latent Semantic Indexing (LSI)

Latent Semantic Indexing is a mathematical technique from natural language processing, originally developed in the late 1980s for analyzing relationships between terms and documents in small, static collections. It’s also one of the most persistent pieces of misinformation in SEO discourse: the term “LSI keywords,” widely used in SEO content and tools to mean “related terms” or “semantically relevant keywords,” does not describe anything Google actually uses. Google’s own John Mueller has stated directly that there’s no such thing as LSI keywords in the context of how Google ranks pages. What people are actually describing when they talk about “LSI keywords” and see results from including them is genuinely useful: writing that naturally incorporates related terms, synonyms, and topically relevant vocabulary tends to produce more comprehensive, better content, which real semantic ranking systems like BERT and neural matching are built to recognize. The benefit is real; the LSI explanation for why it works is not.

Before building a keyword list from a tool’s raw output, run the intent-check described under Keyword Research (examining what’s actually ranking) on your top 10-15 candidates. It catches mismatches a volume-and-difficulty score alone won’t show.

Common keyword intent categories

Intent type	What the searcher wants	Example query
Informational	To learn something	"how does keyword density work"
Navigational	To reach a specific site or page	"google search console login"
Commercial investigation	To compare options before deciding	"best keyword research tools"
Transactional	To complete an action, often a purchase	"buy ahrefs subscription"

Sources cited: SEO by the Sea: does Google use Latent Semantic Indexing, Google Search Central: keyword stuffing spam policy, Google Search Central: Knowledge Graph and structured data