{"id":48570,"date":"2025-05-04T16:38:44","date_gmt":"2025-05-04T16:38:44","guid":{"rendered":"https:\/\/hosthelp.net\/blog\/?p=48570"},"modified":"2025-05-06T21:38:57","modified_gmt":"2025-05-06T21:38:57","slug":"cosine-similarity-copywriting","status":"publish","type":"post","link":"https:\/\/hosthelp.net\/blog\/2025\/05\/04\/cosine-similarity-copywriting\/","title":{"rendered":"Cosine similarity: write relevant and effective copy like Ogilvy"},"content":{"rendered":"\n

Here\u2019s a method to boost the relevance and effectiveness of your content: cosine similarity<\/a><\/strong>.
It\u2019s not just keyword research \u2014 it\u2019s much more than that. Easy to use, compatible with large language models (LLMs), and widely adopted by these models to analyse content.<\/p>\n\n\n\n

Let\u2019s take it step by step <\/p>\n\n\n\n

If you have questions, feel free to leave a comment \u2014 I\u2019ll be happy to respond.and help you<\/p>\n\n\n\n

As I\u2019ve already shared, I\u2019m back in study and research mode. For 2025, I\u2019ve promised myself to dedicate more time to learning and staying up to date.<\/p>\n\n\n\n

With all these constant changes, I\u2019m diving deeper into SEO<\/a> and everything that comes with it: social media strategy, digital PR, alternative search engines, and new optimisation approaches.<\/p>\n\n\n\n

I\u2019ve already started shifting my mindset, working more closely with other communication and digital professionals on shared projects.<\/p>\n\n\n\n

I find this way of working much more fulfilling.<\/p>\n\n\n\n

Now, back to the point: I\u2019ve created an article to calculate content relevance for a specific keyword, generate related keywords (even across different semantic fields), include them in your copy, and assign a relevance score.<\/p>\n\n\n\n

Use my method, test cosine similarity<\/strong>. it with different keywords, drop a comment, and let me know if you found it useful. I\u2019m waiting to hear from you.<\/p>\n\n\n\n

\n\n\n\n

What is cosine similarity<\/strong><\/strong><\/h2>\n\n\n\n
\n

Cosine similarity measures the cosine of the angle between two non-zero vectors in a multi-dimensional space. It helps determine how similar two documents or pieces of content are based on their vector representations. The resulting similarity score ranges from -1 to 1:<\/p>\n<\/blockquote>\n\n\n\n

Cosine similarity is a mathematical concept that finds practical application in SEO. It\u2019s especially useful for:<\/p>\n\n\n\n

    \n
  • Improving content relevance<\/li>\n\n\n\n
  • Helping Google recognise and rank your content more effectively<\/li>\n<\/ul>\n\n\n\n

    Cosine similarity can be used to measure how relevant a keyword is within a document or across multiple documents.
    It\u2019s different from keyword research because it doesn\u2019t just detect keyword presence \u2014 it also evaluates their semantic relationship and distribution within the text.<\/p>\n\n\n\n

    That makes it a perfect choice for enhancing content quality and relevance in advanced SEO or document analysis.<\/p>\n\n\n\n

    \n

    Using cosine similarity allows SEO professionals to:<\/p>\n\n\n\n

      \n
    • Detect <\/strong>duplicate content : By comparing vector representations of web pages, businesses can identify and resolve duplicate content issues that might harm their rankings.<\/li>\n\n\n\n
    • Optimise Internal Linking<\/strong>: Group similar content pages together, improving site navigation and\u00a0user experience.<\/li>\n\n\n\n
    • Audit Content Overlaps<\/strong>: Regular content audits with\u00a0cosine similarity measures\u00a0can prevent keyword cannibalisation.<\/li>\n<\/ul>\n<\/blockquote>\n\n\n\n
      \n\n\n\n

      Cosine similarity and large language models<\/h2>\n\n\n\n

      Large Language Models (LLMs) are advanced AI technologies designed to understand and analyse text. While they can read content, they need a system to interpret it.<\/p>\n\n\n\n

      To analyse content, they use embeddings<\/strong>, which convert text into numeric values. Cosine similarity is one of the methods used to measure how similar those values are.<\/p>\n\n\n\n

      Let\u2019s say we want to identify the most relevant keywords for the focus keyword “step per dimagrire”<\/strong> using cosine similarity<\/strong>.<\/p>\n\n\n\n

      Here\u2019s the process using Claude AI <\/p>\n\n\n\n

        \n
      1. Enter this prompt:
        Identify top keywords related to “step per dimagrire” to achieve best cosine similarity.<\/strong><\/li>\n\n\n\n
      2. In the same chat, use this prompt:
        Using these keywords, write a 400-word SEO-oriented article focused on “step per dimagrire”.<\/strong><\/li>\n\n\n\n
      3. Then, use this prompt:
        Evaluate the cosine similarity score for the keyword “step per dimagrire” within this text.<\/strong><\/li>\n<\/ol>\n\n\n\n

        The first prompt gives you semantically related keywords.
        The second prompt creates an optimised article using those keywords.
        The third prompt assigns a cosine similarity score.<\/p>\n\n\n\n

        The score should be close to 80%<\/strong> to be considered effective.<\/p>\n\n\n\n

        Use my method . to:<\/p>\n\n\n\n

          \n
        • Detect semantically relevant keywords<\/li>\n\n\n\n
        • Generate content with those queries<\/li>\n\n\n\n
        • Test the effectiveness of the content using cosine similarity scores<\/li>\n<\/ul>\n\n\n\n
          \n\n\n\n

          <\/p>\n","protected":false},"excerpt":{"rendered":"

          Here\u2019s a method to boost the relevance and effectiveness of your content: cosine similarity.It\u2019s not just keyword research \u2014 it\u2019s much more than that. Easy to use, compatible with large language models (LLMs), and widely adopted by these models to analyse content. Let\u2019s take it step by step If you have questions, feel free to […]<\/p>\n","protected":false},"author":1,"featured_media":48571,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"rank_math_lock_modified_date":false,"footnotes":""},"categories":[2],"tags":[],"class_list":["post-48570","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog"],"_links":{"self":[{"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/posts\/48570","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/comments?post=48570"}],"version-history":[{"count":2,"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/posts\/48570\/revisions"}],"predecessor-version":[{"id":48573,"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/posts\/48570\/revisions\/48573"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/media\/48571"}],"wp:attachment":[{"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/media?parent=48570"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/categories?post=48570"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/hosthelp.net\/blog\/wp-json\/wp\/v2\/tags?post=48570"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}