How to Do Keyword Discovery (2026 Guide)

What keyword discovery is (vs research vs tracking)

Keyword discovery is the process of generating a broad list of potential keywords that a target audience might use to find information related to a specific topic or business. It focuses on quantity and breadth, identifying many keyword candidates from various sources.

This differs from keyword research and keyword tracking.

• Keyword research involves analyzing a smaller, more refined list of keywords to understand their search volume, competition, and potential value. Its goal is to prioritize keywords for content creation or optimization.
• Keyword tracking is the ongoing monitoring of a predefined set of keywords to observe their ranking performance over time. It measures the effectiveness of SEO efforts.

In essence, discovery generates the raw material, research refines it, and tracking monitors the outcomes.

Seed keyword to expanded set: the funnel

LemRank's keyword discovery process begins with a single "seed keyword." This seed keyword represents the core topic or product you want to explore. From this starting point, the platform expands outward to generate a comprehensive list of related keyword candidates.

Seed keywords are typically short-tail terms, such as "coffee" or "digital marketing software." The system then uses various techniques to identify long-tail variations, questions, and semantic relatives that people search for. The initial expansion can turn one seed keyword into 200–2000 candidate opportunities. This large set is then progressively filtered and scored to identify the most promising prospects.

The 12 harvest sources LemRank taps

LemRank's keyword discovery engine gathers keyword suggestions from a diverse array of sources. This multi-source approach ensures a comprehensive initial crawl, capturing keywords that might be missed by relying on a single data provider.

The platform first queries a cache that harvests from 12 free and publicly available sources to generate an initial candidate pool. These sources include:

• Google Suggest
• Bing Suggest
• Reddit
• YouTube
• Wikipedia
• Amazon
• People Also Ask (PAA)
• Google Related Searches
• Google Autocomplete
• Keyword Planner Idea (general data)
• eBay
• Quora

Should the initial harvest from these free sources not meet a sufficient threshold, LemRank then intelligently falls back to paid third-party APIs from providers like Semrush, Ahrefs, and Moz for additional keyword data. This layered approach maximizes the breadth of the initial keyword set.

Filtering the raw expansion set

After generating a large set of keyword candidates from the harvest sources, the next step in the discovery process is to filter this raw expansion set. This filtering helps refine the list and remove irrelevant or low-value terms.

LemRank allows users to apply various filters based on specific criteria. Common filtering options include:

• Minimum/Maximum Search Volume: Exclude keywords that have very low or exceedingly high search volume, depending on your strategy.
• Keyword Length: Focus on long-tail keywords by setting a minimum word count.
• Presence/Absence of Specific Words: Include or exclude keywords containing certain terms. For example, you might exclude brand names if you are focusing on generic keywords.
• Keyword Difficulty: Filter out keywords that are too competitive.

This stage significantly reduces the initial raw list of 200–2000 candidates, preparing a more manageable set for detailed analysis.

Scoring by KGR, KDI, and intent

Once the raw keyword set has been expanded and filtered, LemRank applies a scoring mechanism to prioritize the most actionable opportunities. This involves evaluating each keyword against three key metrics: Keyword Golden Ratio (KGR), Keyword Difficulty Index (KDI), and searcher intent.

• Keyword Golden Ratio (KGR): KGR is a data-driven metric used to find long-tail keywords that are underserved on the internet. It is calculated by dividing the number of Google results containing the keyword phrase in their title tag (using an allintitle: search) by their monthly search volume. A KGR below 0.25 typically indicates an excellent opportunity, suggesting that the keyword is being searched for, but few websites are directly targeting it.

  For example, if a keyword has 100 searches per month and an allintitle: count of 10, its KGR is 0.10 (10/100). This would be considered a strong KGR.

• Keyword Difficulty Index (KDI): KDI is a proprietary metric provided by LemRank that estimates how difficult it would be to rank in the top organic search results for a specific keyword. It takes into account various factors such as domain authority of competing sites, content quality, and relevance. The KDI is presented on a scale, typically from 0 to 100, where lower numbers indicate easier ranking potential.

• Searcher Intent: LemRank categorizes the likely intent behind a search query. Understanding intent helps align content with what users are truly looking for. Common intent categories include:

  • Informational: Users seeking answers to questions or general knowledge (e.g., "how to make coffee").
  • Navigational: Users trying to find a specific website or page (e.g., "LemRank login").
  • Transactional: Users looking to buy something (e.g., "buy noise-canceling headphones").
  • Commercial Investigation: Users researching products or services before making a purchase (e.g., "best espresso machine reviews").

LemRank prioritizes keywords where all three — search volume, KDI, and intent — pass a predetermined threshold. For example, a keyword must have a minimum search volume, a KDI below a specific level (e.g., KDI < 40), and an identified intent category that aligns with content strategy (e.g., Informational or Commercial Investigation). Only keywords meeting these combined criteria are presented in the actionable list, typically narrowing the initial large list down to 20–100 high-potential opportunities.

From discovery to prioritised backlog

The final output of the keyword discovery process in LemRank is a prioritized backlog of actionable keywords. This backlog is a refined list of 20–100 keywords that have successfully passed through the expansion, filtering, KGR, KDI, and intent scoring stages.

Each keyword in this backlog is presented with its relevant metrics, such as search volume, KDI, KGR, and inferred intent. This data enables strategic decision-making for content creation and optimization.

The prioritized backlog serves as a direct input for your content strategy. You can use this list to:

• Plan new content: Create blog posts, articles, or landing pages specifically targeting these high-potential keywords.
• Optimize existing content: Identify opportunities to update or enhance current pages to rank for these terms.
• Inform product development: Understand user needs and pain points reflected in transactional or commercial investigation queries.
• Cluster related topics: Group similar keywords to create comprehensive content pillars, addressing a broader range of user queries within a single, authoritative piece of content.

This systematic approach ensures that content efforts are focused on keywords with the highest probability of generating organic traffic and achieving business objectives.

Related

• Keyword Golden Ratio
• Long-Tail Keywords Guide
• Keyword Clustering Explained