> ## Documentation Index
> Fetch the complete documentation index at: https://docs.surnex.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Compare audits

> Diff two crawls to see what you fixed and what regressed.

**Audits → Compare** puts two crawls side by side. It's how you prove a fix worked — and catch the regressions a single score won't show you.

## Running a comparison

Two dropdowns at the top: pick an **older** audit and a **newer** one. Results appear as soon as both are selected.

Until then you'll see **Select two audits to compare** — "Choose an older and newer audit above to see how your site has improved".

## Stat cards

| Card                | What it shows                                             |
| ------------------- | --------------------------------------------------------- |
| **Score Change**    | The difference between the two audit scores               |
| **Old Score**       | The earlier audit's score, for reference                  |
| **Resolved Issues** | Issues present in the old audit and gone from the new one |
| **New Issues**      | Issues in the new audit that weren't in the old one       |

**Resolved** and **New** are the cards that matter. Score change is a net figure and hides offsetting movement: a crawl that resolves 30 issues and introduces 28 barely moves the score while telling you something important about your deploy process.

## Category breakdown

A table with one row per [issue category](/audits/overview#issue-categories):

| Category | Before | After | Change |
| -------- | ------ | ----- | ------ |

This is where a fix is confirmed or contradicted. Work on meta tags should show `meta_tags` falling and everything else flat. If `links` rose at the same time, the change that fixed your titles broke something else.

## Comparing fairly

The comparison is only meaningful if the two crawls are alike:

* **Same page limit.** A 50-page crawl against a 500-page crawl reports hundreds of "new" issues that were always there and simply weren't crawled before.
* **Same target URL.** Crawls starting at different paths cover different pages.
* **Enough time between them.** Comparing two crawls from the same afternoon measures crawl variance, not progress.

When the numbers look wrong, check the page limits first. Nearly every confusing comparison is two crawls of different scope.

## A working loop

1. Run an audit and note the score.
2. Fix the errors from the [issue list](/audits/issues), starting with sitewide patterns.
3. Run a second audit with **the same target URL and page limit**.
4. Compare them.
5. Read **Resolved** and **New** together, then check the category rows.

Repeat. Because Surnex keeps every audit, you accumulate a record of what changed and when — useful in a report, and useful when a score drops and you need to find which deploy caused it.

## Related

* [Audit issues](/audits/issues) — what each category contains
* [Run an audit](/audits/run) — matching scope between crawls
* [Reports](/reports/add) — putting audit progress in front of a client