Analyse TF-IDF scores across your content and up to 3 reference/competitor texts. Identifies terms you overuse, underuse, or are missing entirely relative to your corpus. Shows normalised TF-IDF scores for each term with colour-coded comparison bars.
Paste the text of your page or article in the "Your Document" field. This is the content you want to analyse and improve. Minimum 100 words is recommended for meaningful TF-IDF calculations.
Paste up to 3 competitor pages or reference documents in the additional fields. These form your comparison corpus — the baseline for calculating how distinctive your terms are relative to the competitive landscape. Leave them empty to analyse your document alone using estimated IDF values.
Terms are ranked by TF-IDF score. High-scoring terms (frequent in your document, rare across the corpus) are your distinctive topic signals. Terms appearing frequently in competitor documents but absent or underused in yours are highlighted as content gaps — topics you should consider adding for better topical coverage.
TF-IDF stands for Term Frequency–Inverse Document Frequency. It is a numerical statistic measuring how important a term is to a document within a collection. TF (Term Frequency) = (number of times term appears in document) ÷ (total words in document). IDF (Inverse Document Frequency) = log(total documents in corpus ÷ number of documents containing the term). TF-IDF = TF × IDF. Terms that appear frequently in your document but rarely across other documents score high — they are topically distinctive. Common words appearing in all documents (like "the", "and") score near zero.
Keyword density only measures how often a term appears within your document. A 2% density could mean the term is highly topical or it could be a common word that appears everywhere. TF-IDF accounts for how common or rare a term is across a broader set of documents. If "machine learning" appears frequently in your document AND infrequently across the web, its TF-IDF score is high — it is a genuine topical signal. If "information" appears frequently in your document AND frequently everywhere, its TF-IDF score is low — it is not a distinctive topical signal. TF-IDF is therefore a more meaningful measure of topical relevance.
Compare your content against top-ranking competitor pages for your target keyword. Identify terms with high TF-IDF scores in the competitors' documents that are absent or underused in yours — these are content gaps and semantic terms you should add to improve topical coverage. Identify terms you are overusing relative to the corpus — these may be diluting your content's topical focus or contributing to a keyword-stuffed feel. The goal is not to match competitors exactly, but to ensure your content covers the topic comprehensively from multiple semantic angles.
Google has never confirmed using raw TF-IDF directly as a ranking signal, but the concept of term frequency and document significance is foundational to information retrieval and clearly influenced early search engine development. Modern search algorithms use far more sophisticated approaches — semantic similarity, BERT language models, entity recognition, and contextual embeddings — that go well beyond simple TF-IDF. However, TF-IDF analysis remains a practical proxy for identifying: semantic terms to include for topical coverage, overused terms to vary with synonyms, and content gaps relative to top-ranking pages.
TF-IDF scores are relative, not absolute — they only have meaning in comparison to other terms within the same corpus. A high TF-IDF score for a term in your document means it is both frequent in your text and rare across the reference documents, making it a strong topical signal. There is no universal "good" number — instead, compare the relative ranking of terms. Terms in the top 20% of TF-IDF scores are your most distinctive topical signals. Terms near zero are either stop words or very common words that add little topical value.
No — TF-IDF analysis and keyword research serve different purposes. Keyword research identifies what terms people are searching for and how competitive they are — it is outward-facing. TF-IDF analysis measures the topical coverage and term importance of existing content — it is inward-facing. They are complementary: use keyword research to identify what to target, then use TF-IDF analysis (comparing against top-ranking pages) to ensure your content covers the topic comprehensively with the right term distribution. TF-IDF analysis is most useful for content optimisation of existing pages.
