---
name: clickd-content-refresh
description: Finds decaying or outdated pages on a website using Search Console traffic trends and clickd page checks, diagnoses whether clicks dropped from position loss, CTR decay or stale facts, updates the page in code, and verifies the refreshed page live with clickd. Use when asked to refresh old content, recover lost organic traffic, update outdated guides or year-specific pages, or fix stale facts and broken source links.
metadata:
  author: subinternal
  version: "2.3.0"
---

<!-- subinternal:header -->
```
╭─────────────────────────────────────────────────────────────────╮
│        ▄                                                        │
│      ▄ █    clickd                                              │
│    ▄ █ █    content-refresh                                     │
│  ▄ █ █ █                                                        │
│                                                                 │
│  skill 2.3.0  ·  by subinternal  ·  clickd.com/skills/content-refresh │
╰─────────────────────────────────────────────────────────────────╯
```

Print this header once, exactly as it is, in a code block, as the first thing you show when this skill starts. Then go on.
<!-- /subinternal:header -->

# Content refresh

Recover slipping pages and keep strong pages current by diagnosing why a page lost clicks, updating the actual facts, examples, headings and schema in the repository, and verifying the refreshed page on the live site with clickd.

What this does not do: bumping a `dateModified` timestamp without changing the substance of a page is a spam signal that search engines ignore or penalize. Only update a page's modified date when you make real, visible improvements to its content.

Connect to `https://clickd.com/mcp/all` to use `clickd_traffic`, `clickd_queries`, `clickd_tells` and `clickd_links` alongside `clickd_pages`, `clickd_check`, `clickd_rescan` and `clickd_changes` (also at `https://clickd.com/mcp`). With no MCP setup, scan a site at `https://clickd.com/scan/example.com`.

## 1. Find pages losing clicks or carrying stale facts

Identify which existing pages need a refresh first from traffic trends and page findings.

- Call `clickd_traffic({ site })` and `clickd_queries({ site })` (on `https://clickd.com/mcp/all`) to spot pages where clicks or average position slipped over recent weeks, or pages with high impressions where click-through rate lags.
- Call `clickd_pages({ site, path })` on each candidate page to see its failing rules, heading outline, word count and outbound links.
- Read the page's source file in the repository to flag outdated version numbers, retired APIs or products, dead external links, old screenshots or pricing facts that have changed elsewhere on the site.

Verify: list each page to refresh with its current path, the traffic or rule signal that flagged it, and the specific sections that are stale.

## 2. Diagnose why the page slipped

Separate a snippet problem from a substance problem or an internal link problem before editing.

- **Impressions held steady, clicks dropped:** the page still ranks, but its `<title>` or `<meta name="description">` looks dated or less specific than newer results around it.
- **Impressions and average position dropped for its main query:** rivals now answer a sub-question, show a worked example, or reflect a newer rule or version that this page lacks.
- **A newer page on the same site started taking impressions for the same query:** the two pages compete. Either fold the weaker page into the stronger one with a `301` redirect, or give each page a clearly distinct job.

Verify: state the diagnosis (snippet decay, substance gap, or internal overlap) for each page before changing the file.

## 3. Refresh the substance, headings and links in code

Update the page so a reader landing today gets accurate, complete answers immediately.

- Replace obsolete facts, commands, prices or screenshots with current ones verified against the site's own data or primary sources.
- Rewrite the opening paragraph under `<h1>` and each `<h2>` question so the first 40 to 60 words give the direct answer first.
- Add the missing section or comparison table that searchers now ask for (seen in `clickd_queries` or rival pages), and strip any AI writing tells flagged by `clickd_tells({ url })` (on `https://clickd.com/mcp/all`).
- Check every external source link (`curl -sI`) to replace `404` or redirected citations, and add internal links from newer pages on the site via `clickd_links({ site })` (on `https://clickd.com/mcp/all`).
- Keep the existing URL slug intact. Never change a working URL just to remove or change a year in the path unless you also wire a permanent `301` redirect and update every internal link.

Verify: diff the page's source file to confirm real factual and structural improvements were made before updating `dateModified` in visible text and `Article` JSON-LD.

## 4. Check the refreshed page live with clickd

Deploy the update and confirm the live page passes every on-page and schema check.

- Call `clickd_check({ url, kind: "agent" })` to confirm the new headings, updated facts and `dateModified` appear in the server-rendered HTML.
- Call `clickd_rescan({ site, reason: "refreshed decaying pages" })`, poll `clickd_changes({ site, job })` until done, and log the refresh with `clickd_log_work`.

Verify: `clickd_changes({ site, job })` shows the refreshed page passing its rules on the live domain.

## Report back

```
Site: <domain>
Page refreshed: <path>
Diagnosis: <snippet decay | substance gap | internal overlap>
What changed: <facts updated, sections added, dead links fixed, internal links added>
dateModified updated: <yes, YYYY-MM-DD | no>
clickd live check: <passing | remaining findings>
```

## Sources

- [Creating Helpful, Reliable, People-First Content](https://developers.google.com/search/docs/fundamentals/creating-helpful-content): why updating substance for readers matters and why changing dates without real updates fails Google's self-assessment.
- [How to Specify a Canonical with rel="canonical" and Other Methods](https://developers.google.com/search/docs/crawling-indexing/consolidate-duplicate-urls): consolidating competing pages cleanly when two routes split impressions for one query.
- [Article Structured Data](https://developers.google.com/search/docs/appearance/structured-data/article): the `datePublished` and `dateModified` ISO 8601 properties on `Article` and `BlogPosting`.
- [Search Analytics: Performance Report](https://support.google.com/webmasters/answer/7576553): comparing clicks, impressions, CTR and position across date ranges in Search Console.
