> ## Knowledge Base Index
> Fetch the complete knowledge base index at: https://support.seotesting.com/sitemap.xml
> Use this file to discover available pages before exploring further.
> Pure-Markdown content can be obtained by appending a '.md' suffix to the content URLs listed in the sitemap (without the trailing slash).

# Content Evaluator tool

## **What is the Content Evaluator?**
The **Content Evaluator** scores how well a page answers the queries generated by the [Query Fan-Out Generator](https://support.seotesting.com/en/article/using-the-query-fan-out-tool-hqh1v7/). 
Rather than checking for keyword matches, it compares your content against the full query set by meaning, so you can see which questions a page covers, which it covers only partially, and which it misses entirely.

## **Why is this useful?**
A page can rank for its main keyword and still leave most of a topic unanswered. If you don't know which surrounding questions, comparisons, and intents your content is missing, it's hard to know what to add or where a rewrite will have the most impact. The Content Evaluator turns "this page might need work" into a specific list of covered, partial, and missing queries - so you can prioritise the gaps that matter.

## **How does the Content Evaluator work?**
The tool uses embeddings and semantic similarity to compare your page content against the query set generated by Query Fan-Out. Each query is matched against your content by meaning rather than exact wording, and classified as covered, partially covered, or missing. Because the comparison is semantic, a query can count as covered even when your page doesn't use the same phrasing.

Unlike the Persona Page Evaluator, the Content Evaluator doesn't need a page URL as a key - it evaluates against the saved Query Fan-Out results you upload, not against saved questions looked up from a database.

## **Before you start**
|| You need a Query Fan-Out export first. The Content Evaluator works against the JSON file produced by the Query Fan-Out Generator. Generate the query set for your topic and download that JSON before running an evaluation.

## **How to run an evaluation**
* From your dashboard, click the 'select your website' dropdown at the top of the page and choose the site you want to use the tool for. 
* Select **Labs** from the left-hand menu, and scroll down to find **Content Evaluator**.

![](https://storage.crisp.chat/users/helpdesk/website/-/3/4/3/d/343dcfd35d951400/screenshot-2026-07-22-122418_1wmpfqh.png =917xauto)

* **Upload** the JSON file you generated and downloaded from the [Query Fan-Out Generator](https://support.seotesting.com/en/article/using-the-query-fan-out-tool-hqh1v7/).
* Provide the content to evaluate, in one of two ways:
    * **Live URL** - the page at your chosen URL is fetched and evaluated.
    * **Pasted content** - paste plain text or markdown directly, so you can test draft or edited content before it goes live. To turn a live page into markdown to paste in, use our free[ __URL to Markdown Converter__](https://support.seotesting.com/en/article/url-to-markdown-converter-1i5elf2/) tool.

![](https://storage.crisp.chat/users/helpdesk/website/-/3/4/3/d/343dcfd35d951400/screenshot-2026-07-22-115533_1m8n60q.png =891xauto)

* Click **Evaluate Content Coverage**. The tool compares your content against the uploaded query set (this takes a few seconds).
* Review your results, described below.

## **Understanding your results**
The evaluation breaks the query set into three groups:
* **Covered** - queries your page answers well.
* **Partial** - queries the page touches on but doesn't fully answer.
* **Gaps** - queries the page doesn't address.

![](https://storage.crisp.chat/users/helpdesk/website/-/3/4/3/d/343dcfd35d951400/screenshot-2026-07-22-120011_1gy61ok.png =902xauto)

Use the partial and missing queries under the ‘**Query Coverage**’ tab as your to-do list (you can filter by ‘gaps only’ and ‘partial only’). They're the questions your audience (and AI search systems) expect the page to answer that it currently doesn't. 

\*Because you can also paste content in, you can draft additions, re-evaluate, and confirm the gaps are closing before you publish.

* The ‘**Persona Relevance**’ tab gives you an analysis of how well your content serves each target persona based on their high-priority queries.

* The ‘**Content Issues**’ tab will show you any content chunks that are not relevant to the topic queries.

* Use the ‘**Chunk Analyzer**’ tab to paste content in - to find which queries it best matches based on cosine similarity.

||| Tips for getting the most out of it:
* **Start from a Query Fan-Out you trust.** The evaluation is only as complete as the query set you upload, so generate the fan-out from a well-chosen seed topic.
* **Iterate on pasted content before publishing.** Paste a draft, evaluate, add to the weak areas, and re-evaluate - a fast loop for closing gaps without touching the live page.
* The **‘Extracted Content’** tab shows the markdown content extracted from the page that was analyzed - use this to verify the content was fetched and parsed correctly.
* **Prioritise by intent, not volume.** A handful of missing decision-stage or commercial queries often matters more than a long tail of minor ones.
* **Set up an SEO or LLM test** in SEOTesting to demonstrate the impact