Keyword Difficulty (KD)

Keyword Difficulty (KD) is a score that SEO tools give a keyword to estimate how hard it would be to rank on the first page of Google for it. Most tools use a 0 to 100 scale, where higher means harder. It is one of the traditional metrics, alongside search volume and CPC, used to prioritise which keywords to target.

How it works

Each tool calculates KD its own way. Semrush describes KD% as a percentage from 0 to 100 based on factors such as the median number of referring domains pointing to the ranking URLs, their dofollow/nofollow ratio and authority, and SERP features. Ahrefs bases its KD mainly on the referring domains of the top-ranking pages. Because the formulas differ, a keyword can score 30 in one tool and 55 in another, and the numbers should only be compared within the same tool.

How it’s used in SEO and marketing

• Prioritisation. Combine KD with volume, intent and trends to find realistic opportunities.
• Cluster-level planning. After keyword clustering, calculate average KD per cluster to see which topics are competitive and which are open. MLforSEO’s KeyBERT tutorial suggests charting sum of volume and average difficulty per cluster.
• Dashboard filters. Filter a keyword universe in Data Studio to low-difficulty clusters that match a given intent or SERP feature.
• Sequencing. Start with lower-difficulty topics to build relevance before tackling the hardest head terms.

Example

You cluster 2,000 keywords with KeyBERT and add search volume and KD. Cluster A has the highest total volume but an average KD of 70. Cluster B has half the volume, an average KD of 25 and mostly informational intent that matches your blog. Cluster B is the faster win, while Cluster A goes on the roadmap once you have supporting content and links.

Limitations and things to watch

• KD is an estimate built mostly from backlink data. It ignores how relevant and authoritative your own site is on the topic.
• It does not account for SERP features such as AI Overviews that reduce clicks even when you rank.
• Long-tail and new queries often have no KD data at all.
• MLforSEO’s semantic approach treats KD as one input among many: relevance to the brand, the user’s information needs and your ability to produce the content matter just as much.

Learn it on MLforSEO

• How to do keyword clustering with KeyBERT
• Keyword Cluster (KeyBERT) Data Studio Dashboard
• What Does a Good Semantic Keyword Universe Actually Look Like?
• Semantic AI-powered SEO Keyword Research course