Why does Effort Score matter to AI models?
Completeness signals are the oldest quality proxy search engines rely on: a page with a table, visual material and a current date more often results from thorough work on the topic than a page made of five paragraphs. Search engines treat it as a strong hint of quality.
What will you see in the report?
In the report you will find a checklist of met and missing items, with a hint next to each gap about exactly what to add. Beside it sits a format comparison with competitors: if six of them have a table and you do not, you see it directly.
We report freshness separately: the publication date, the update date and the age of the content in months.
One limitation applies here: images, video and dates are read from the page code, so an audit of content pasted from an editor rather than fetched from a URL does not cover those criteria.
Sample recommendations
Fragments of a report from an audit of a gaming mice category page in an electronics store.
Problem: The content is clearly shorter than the Top 10 average for this query (1,460 words) - competitors cover the topic more widely.
Four short descriptive sections under the product list.
Add the missing threads - a wired versus wireless comparison and a model ranking - until the length of the content reaches the competitor average.
Problem: The page shows neither a publication nor an update date, the simplest freshness signal there is.
Copyright © 2025. No datePublished or dateModified in the markup.
Add a visible update date plus datePublished and dateModified in JSON-LD, and refresh it whenever the offer changes.
How do we measure Effort Score?
The checklist covers up to eleven criteria and an algorithm alone computes it, with no language model involved. Some are absolute, some are relative to competitors - the length threshold is the Top 10 average for your phrase, not a fixed word count.
The criteria adapt to the page type: a table of contents is only checked for articles and encyclopedia entries, and a visible date is skipped on landing pages.
What the checklist covers
| Criterion | What it covers |
|---|---|
| Absolute length | Content over 700 words |
| Length against competitors | Content above the Top 10 average for this phrase |
| Visual material | At least four images |
| Video | An embedded video |
| Lists | Bulleted or numbered enumerations |
| Tables | Data presented in a table |
| Heading hierarchy | Correct H1, H2, H3 structure |
| Emphasis | Key terms in bold |
| Table of contents or summary | Articles and encyclopedia entries only |
| No walls of text | Paragraphs broken up and formatted |
| Freshness | A visible publication or update date |
What raises and what lowers the score?
Raises
- Content longer than the Top 10 average for the phrase
- Four or more pieces of visual material
- Data presented as tables and processes as lists
- A visible publication and update date in the page code
- A table of contents on longer articles
Lowers
- Content clearly shorter than the Top 10 competitors
- No visual material at all
- Running text with no lists, tables or emphasis
- No date - or a date visible only on screen, with no marker in the code
- An incorrect heading hierarchy
Frequently asked questions
Does longer text always raise the score?
No - the threshold is the Top 10 average for your phrase, not a fixed word count. On phrases where competitors write concisely, expanding the text does not raise the score and lowers it on the other dimensions.
Why does the audit not see my images and video?
Because those signals are read from the page code. An audit of content pasted from an editor instead of a URL has no access to information about images, video or a publication date.
Does a missing table of contents always lower the score?
No. We only check it on articles and encyclopedia entries. On a listing, product page or landing page the criterion is skipped.