Content Pruning & Refresh Strategy

A site is a portfolio. The pages you keep should compound; the pages you can’t make compound should be pruned, redirected, or rewritten. Most teams publish their way into a slow decline because they treat new posts as the only growth lever and ignore the 70% of pages already losing money.

TL;DR

• Every URL is a liability until it earns its place. Crawl budget, link equity, and Google’s site-quality model are diluted by every page that doesn’t pull weight. The best move on stale content is often deletion or 301 to a stronger sibling.
• The keep/update/redirect/delete matrix is the workflow. Quarterly, score each URL on traffic, link equity, intent fit, and freshness, then route to one of four actions. This single discipline is worth more than most “publish faster” strategies.
• Refresh cadence is industry-specific. News and finance need monthly cycles; evergreen B2B SaaS can be 6–12 months; YMYL medical content needs a documented review every 12 months minimum. Pick a cadence and enforce it in the calendar.

The mental model

Pruning is like an orchard: a healthy tree sheds dead branches every season. Leave them on, and they stress the trunk, harbor pests, and shade the productive branches. The 1,200 pages on your site that earn 0–3 clicks a month aren’t free — they’re the dead branches. Each one consumes crawl budget, dilutes site-quality scores, and competes internally with the pages that actually rank.

The four-quadrant decision matrix is the pruning tool. Plot each URL on two axes — performance (traffic, conversions, rankings) and strategic value (does the page serve the brand’s topical authority?). The four cells:

• High performance, high value: keep and refresh on schedule.
• High performance, low value: usually a brand mismatch — keep, but consider whether to retarget the page toward your strategic focus.
• Low performance, high value: rewrite or update aggressively. The intent is right; the execution is failing.
• Low performance, low value: prune. Redirect to the closest relevant page if there’s link equity; otherwise delete and let it 404.

The hardest call is the third quadrant. Most teams either ignore it (publishing more new content while old strategic pages decay) or panic-rewrite without diagnosing why traffic dropped. The disciplined answer is to read the page, read the SERP, and decide whether the gap is intent fit, freshness, depth, or E-E-A-T — then fix that specific gap.

Deep dive: the 2026 reality

The Helpful Content system, baked into Google’s core algorithm in March 2024, scores sites at the domain level, not just the page level. The leaked Yandex documents (January 2023) and the leaked Google Content API documents (May 2024) both confirm that aggregate page-quality signals influence how individual pages rank. The implication: a long tail of zero-traffic, low-quality URLs hurts your good pages by dragging the site-level signal down.

This was demonstrated in the August 2024 core update, when sites that had aggressively pruned thin pages — HouseFresh, NerdWallet, several large publishers — recovered partial visibility while sites that left their archives intact continued to decline.

Decay is normal. A 2024 Animalz study of 100,000 SaaS articles showed that the median article peaks 3–6 months after publish, then declines 10–25% per year if not refreshed. Industries with rapid information turnover decay faster: software comparisons (3 months), tax/finance (6 months), health (12 months), historical reference (3+ years).

Identifying decay is straightforward in Google Search Console:

1. Set the Performance report to last 6 months vs. previous 6 months.
2. Filter to URLs with ≥100 clicks in the prior period.
3. Sort by clicks change ascending — biggest losers first.
4. The losers cluster in three groups: stale freshness (date-based queries), intent drift (the SERP changed), and ranking competitor takeover.

Refresh winners is the highest-ROI work in SEO. The Animalz data and our own production audits both show a 30–60% traffic recovery in 60 days when a top-of-funnel post is rewritten with current data, current SERP fit, and a refreshed dateModified. Cost: ~20% of producing new content.

Consolidation beats deletion when two thin posts target similar intent. Merge the better one into the canonical and 301 the weaker. Watch for keyword cannibalization signals: two URLs from your domain alternating in the same SERP position is the canonical symptom.

Refresh cadence by industry (rule of thumb based on SERP volatility):

Vertical	Min refresh cadence	Reason
News, current events	Continuous	SERP regenerates daily
Finance (interest rates, tax)	Monthly	Underlying numbers change
SaaS / software	Quarterly	Pricing, features, competitors
Travel	Quarterly	Seasonal + price volatility
Health (YMYL)	12 months minimum	Reviewed by clinician annually
Legal	6 months	Regulation churn
Reference / historical	24–36 months	Stable but periodic
Recipes	24 months	Visuals decay; method stable

For AI surfaces, dateModified is read by Perplexity, ChatGPT Search, and AI Overviews as a signal of whether to cite. A 2023 article competing with a 2026 article on a fast-moving topic loses by default unless its content is patently still correct.

Visualizing it

flowchart TD
  Start["Quarterly URL audit"] --> Score["Score: Traffic, Links, Intent, Freshness"]
  Score --> Q{"Decision"}
  Q -->|"High traffic + High value"| Keep["Keep + refresh on cadence"]
  Q -->|"Low traffic + High value"| Update["Rewrite: fix intent, depth, freshness"]
  Q -->|"Low traffic + Low value + has links"| Redirect["301 to nearest stronger page"]
  Q -->|"Low traffic + Low value + no links"| Delete["410 or remove from sitemap"]
  Update --> Republish["Republish, update dateModified, request reindex"]
  Redirect --> GSC["Submit redirect map to GSC"]
  Delete --> GSC
  Keep --> Cadence["Schedule next refresh"]

Bad vs. expert

The bad approach

The amateur leaves everything live and adds dateModified manually.

<!-- /best-crm-software-2022/ -->
<article>
  <h1>Best CRM Software 2022</h1>
  <p class="meta">Updated: 2024-08-12</p>
  <p>The best CRM software for 2022 includes Salesforce, HubSpot, and Pipedrive...</p>
  <!-- Body still references 2022 pricing, 2022 features, and now-defunct competitors -->
</article>

The URL is dated, the H1 is dated, the body is dated, and only the meta line is “fresh.” Google’s freshness signals catch this — the TextBlooper signal and the freshness gauge inside Navboost both compare the date claim against the actual content vintage. Worse, the URL /best-crm-software-2022/ will keep ranking poorly for “best CRM software 2026” no matter how much body text you change. This is when the right answer is consolidation: 301 to /best-crm-software/ (undated URL) and rewrite the canonical.

The expert approach

The expert separates URL, content, and date concerns and treats refresh as an editorial product.

<!-- /best-crm-software/ — undated canonical URL -->
<article>
  <h1>Best CRM Software (May 2026)</h1>
  <div class="meta">
    Published: 2021-03-04
    | Last reviewed: 2026-05-01 by
    <a href="/team/maya-chen/" rel="author">Maya Chen</a>
    | <a href="#changelog">View changelog</a>
  </div>
  <p>We re-tested all 14 CRMs in April 2026 with fresh accounts and updated
  pricing. Three changes since the January 2026 review: Pipedrive raised
  Professional tier to $49/seat, Salesforce launched Einstein Co-Pilot for
  free below 10 seats, and HubSpot deprecated Marketing Hub Starter...</p>

  <section id="changelog">
    <h2>Changelog</h2>
    <ul>
      <li>2026-05-01: Re-tested 14 tools, updated pricing</li>
      <li>2026-01-15: Added Attio, removed Insightly (acquired)</li>
      <li>2025-09-04: Added AI feature comparison column</li>
    </ul>
  </section>
</article>

{
  "@context": "https://schema.org",
  "@type": "Article",
  "datePublished": "2021-03-04",
  "dateModified": "2026-05-01",
  "author": { "@type": "Person", "name": "Maya Chen", "url": "https://example.com/team/maya-chen/" }
}

# Old dated URLs collapse to the canonical
location ~ ^/best-crm-software-(2020|2021|2022|2023|2024|2025)/?$ {
  return 301 /best-crm-software/;
}

The URL is undated and stable. The H1 carries the current month/year as a CTR signal. A visible changelog earns trust with humans and gives crawlers a signal that updates are real. The 301 collapses old dated variants into the canonical. The dateModified in JSON-LD matches the visible last-reviewed date.

Issue	Bad	Expert
URL strategy	/best-crm-2022/ permanent	Undated canonical + 301s
Refresh evidence	Manual date bump only	Body rewrite + visible changelog
Author	Anonymous	Named reviewer with date
Schema	None or stale	datePublished + dateModified + Person
Old variants	Compete with canonical	Redirected and removed from sitemap

Do this today

1. In Google Search Console, open Performance > Search results. Set date range to last 16 months (the maximum). Compare last 6 months vs. previous 6. Sort URLs by clicks change ascending. Export.
2. In Ahrefs Site Explorer > Top Pages, sort by organic traffic. Tag your top 10% as “must keep” and your bottom 50% (those with fewer than 5 monthly visits and no backlinks) as “prune candidates.”
3. Pull Internal Backlinks and External Backlinks for prune candidates. Any URL with a referring root domain or strong internal-link weight gets a redirect, not a delete.
4. Build a decision spreadsheet with columns: URL, monthly clicks, monthly impressions, referring domains, intent fit (yes/no), last updated, recommended action (keep, update, redirect, delete), redirect target, owner, due date.
5. For each update decision, run the target query in incognito Chrome and compare your page to the top 3 ranking results. If your H2 structure, data points, or examples are stale, that’s the gap. Apply the Module 15 content creation process to the rewrite.
6. In your CMS (WordPress, Sanity, Contentful), set the rewritten article’s dateModified to today. Add a visible “Last reviewed” line in the byline pointing to the editor’s author page.
7. For redirects, implement a 301 in your reverse proxy (nginx, Cloudflare, Vercel). Avoid plugin-based redirects when you have more than 50 — they slow the response. Submit a fresh sitemap to GSC after the change.
8. For deletions, return 410 Gone on URLs that should be permanently removed. Remove them from the sitemap. 404 is fine but 410 signals intent more strongly to Google’s crawler.
9. Set up a quarterly recurring meeting on the calendar titled “Content Audit Q[N].” Assign one editor to run the workflow against the entire site (or one cluster per quarter for big sites).
10. Track refresh ROI in a separate sheet: URL, refresh date, clicks 28-days-pre vs. 28-days-post, impressions delta, position delta. After a year you’ll know which kind of refresh (data update, depth expansion, intent retarget) yields the most lift in your domain.