Skip to content
NagentNagent
Log inSign upHire your AI Growth Team
NagentNagent
Log inSign up

How to update old blog posts based on performance data

8 Minutes read
Published:
Pull 12 months of Search Console and conversion data per post, sort each into refresh, rewrite, merge or retire, fix the specific weakness the data shows, then measure at fixed windows. Here is the full method with thresholds and checks.
PB
Pratap BeheraNov 8, 2026·8 min read
How to update old blog posts based on performance data

To update old blog posts from performance data, export 12 months of Search Console clicks, impressions, position and click-through rate per URL, plus conversions from analytics. Sort each post into refresh, rewrite, merge or retire. Then fix the specific weakness the numbers show: a weak title for low click-through, missing subtopics for a page stuck on page two, outdated facts for a slow decline. Keep the URL, change the date only for real changes, and measure at fixed windows.

We make Nagent, whose CREA and DRIS agents do parts of this, so they get their own section near the end. Everything else here works with free tools.

Which performance data should you pull first?

Build one sheet with a row per post and these columns:

  • Clicks and impressions, last 3 months against the same 3 months a year earlier, from Search Console.

  • Average position and click-through rate for the post's top five queries.

  • Top queries, so you can see whether the post ranks for what you meant it to.

  • Conversions from analytics: sign-ups, demo requests or downloads that started on the post.

  • Backlinks and internal links pointing to it.

  • Last updated date.

  • AI citations, if you track them: whether assistants such as ChatGPT or Perplexity cite the post for its target question.

Twelve months is enough to see seasonality. Less than three months is too noisy for most blogs.

How do you decide whether to refresh, rewrite, merge or retire a post?

Use the data to sort, not gut feel.

  • Refresh when the topic still matters and the post ranks on page one or two, roughly positions 4 to 20. Small changes move these most.

  • Rewrite when the post ranks poorly or no longer matches what searchers want. Rebuild the structure and argument on the same URL.

  • Merge when two or more of your posts compete for the same queries. Keep the stronger URL, fold in the best parts of the others, and redirect them. Google recommends "a permanent server-side redirect whenever possible" (Google).

  • Retire when a post has no traffic, no links and no purpose. Redirect it to the closest relevant page rather than leaving a dead URL.

AirOps' refresh guide uses a similar split and suggests treating a traffic fall of 20% or more over a quarter as a decay signal. Pick your own threshold and write it down, so the choice is consistent.

Which posts should you update first?

Score each candidate on three things, one to five each:

  1. Upside: impressions on its target queries and how close it sits to the top three.

  2. Value: how close the topic is to a buying decision, and whether the post already converts.

  3. Effort: how much work the update needs. Score in reverse, so small fixes score high.

Work the highest totals first. In practice, the first batch is usually posts with high impressions and weak click-through, and posts sitting in positions 4 to 10 on queries that matter to your buyers.

What should you change, based on what the data shows?

Match the fix to the symptom.

  • High impressions, low click-through: the title and meta description do not match what searchers want. Rewrite them to state the answer or the specific benefit. Change one element at a time so you can tell what worked.

  • Ranks for the wrong queries: the post drifted. Re-aim the opening paragraph and headings at the query you want, or accept the query it earns and adjust the angle.

  • Stuck in positions 8 to 20: competing pages answer more of the question. Compare their headings with yours, then add the missing sub-questions as new sections.

  • Slow decline over months: facts, prices, screenshots or examples are out of date. Replace them and remove anything you can no longer support.

  • Traffic but no conversions: the post answers the question but gives no next step. Add a relevant call to action and links to the product or comparison pages that match the reader's stage.

  • Not cited by AI assistants: the answer is buried. Put a direct answer in the first paragraph and keep each section self-contained, so a passage can be quoted on its own.

  • Two posts swapping positions: cannibalisation. Merge them.

What should you never change when updating a post?

  • The URL, unless you are merging. A changed URL throws away links and history. If you must change it, redirect the old one and update internal links.

  • The date, without real changes. Google says dates "must describe the publication or update date of the page" and that the visible date and the structured dateModified should match (Google on publication dates). Update the "Last updated" line when the content genuinely changed.

  • Sections that earn links or citations. Check which passages other sites quote before cutting them.

  • Everything at once on a high-traffic page. Big rewrites make it impossible to tell what helped.

What is a step-by-step refresh workflow?

  1. Export the data into the sheet above, with a baseline for each post.

  2. Sort every post into refresh, rewrite, merge or retire.

  3. Score and pick five to ten posts for this cycle.

  4. Write a short brief per post: the symptom, the fix, the target query, the sources for new facts.

  5. Update the post, keeping the URL. Add a visible "Last updated" date.

  6. Review: a second person checks every changed fact against its source.

  7. Republish and request indexing in Search Console's URL Inspection tool.

  8. Update internal links from newer posts to the refreshed one.

  9. Measure at fixed windows, then re-score.

How do you measure whether an update worked?

  • Two to four weeks: impressions, position and click-through for the target queries in Search Console.

  • Thirty to sixty days: AI citations for the target question, if you track them.

  • Ninety days: conversions and pipeline from the post.

Compare against the baseline in your sheet, and against a few posts you did not touch, so a site-wide or seasonal shift is not mistaken for your fix.

What mistakes make content updates backfire?

  • Refreshing posts nobody needs. Cost: hours spent on pages with no upside. Fix: score first.

  • Date-only updates. Cost: credibility with readers and no gain. Fix: change dates only with real changes.

  • Merging without redirects. Cost: lost links and broken internal links. Fix: redirect and re-link the same day.

  • New facts with no source. Cost: a wrong claim on a page that already ranks. Fix: a reviewer checks every changed fact.

  • No baseline. Cost: you cannot tell whether the work paid off. Fix: record numbers before you touch the post.

What do content tools offer for refreshes?

  • AirOps: refresh agents watch Search Console, analytics and AI citations, propose updates, route them for review and track results at 30, 60 and 90 days (AirOps).

  • Jasper: its content refresh guide suggests screening posts older than 12 months and using its Content Improver with a separate SEO audit tool.

  • Writesonic: its platform ranks actions by visibility and citation impact and measures fixes at day 14 and day 28 (Writesonic).

Where does Nagent fit?

Nagent is multiplayer AI for growth: people and AI agents in one shared context, delivering the number. Two of its agents cover parts of this method.

  • Finding what slipped: DRIS, on the Marketing team, reads Google Search Console. Its Rank movers view lists queries whose position moved past their usual band against their own last seven days and queries whose ranking page changed, and marks moves made on a day the whole category moved. Its Queue ranks fixes, and with Search Console connected, a page missing the clicks it should earn at its position rises.

  • Fixing titles and metadata: approved DRIS fixes, such as page titles, meta descriptions and FAQ sections, reach your site through the Site connection, which is in Beta.

  • Avoiding duplicates: in CREA's content studio, approving a new topic first checks your sitemap and, if a live page covers it, suggests updating that page instead.

  • Editing: a published post can be edited and saved in place, and changes go live immediately.

What it does not do: it does not rewrite and republish old posts on its own; a person approves every change. Answer-engine and content proposals wait in DRIS's Backlog rather than its Queue today. There is no direct WordPress publishing.

Frequently Asked Questions

How often should you update old blog posts?

Review the data monthly and update in cycles of five to ten posts. Your highest-traffic and highest-converting posts deserve a check every few months; others when the data shows a slip.

Does updating old content improve SEO?

It can, when the update fixes what the data shows: a weak title, missing sub-questions or outdated facts. Changing the date alone does not.

Should I change the URL when I update a blog post?

No, unless you are merging posts. If you must, use a permanent redirect and update internal links.

How do I find blog posts that are losing traffic?

Compare the last three months with the same period a year earlier in Search Console, per page, and sort by the largest drop in clicks.

See DRIS spot what slipped and CREA's team plan the update: book a demo.

Continue learning

Related agents

Agents that match this read.

Browse all 200+ agents
Select Category
Content

Why does ChatGPT lose the thread when writing a full ebook?

ChatGPT loses the thread on a full ebook because a long chat dilutes its attention, the outline lives only in the scroll, and each reply is written without re-reading the book. Here are the causes and the fixes that work.

Content

How to create an ebook from a PDF of our existing content using AI

Turning an existing PDF into a new ebook with AI starts with one test: can the AI actually read the text? How to check and fix the PDF, decide whether to reformat, condense or expand it, and run the AI steps without losing accuracy.

Content

How do I get AI to write blog posts grounded in our own product knowledge and case studies?

Give the AI a curated source of truth, make it write only from that, and check every claim before publishing. Here is what to put in the knowledge base, how to prepare case studies, which tools support it and how to audit a draft.

Marketing AI

Turn Ad Performance Data Into the Next Creative Brief

The most useful output of any campaign is a single sentence for the next brief. Ad performance data tells you which angle worked, for which buyer, and with what proof. Here is how to close the loop.

Content

How much does a blog post cost to outsource?

For a 1,500-word post, published rates run from about $60 to $240 on content platforms, about $375 to $675 from professional freelancers at EFA survey rates, and more for specialist B2B writers and agencies. Editing, images and review add to every quote.

Content

Nagent CREA vs AirOps for blog content

AirOps fits content operations teams that build workflows and publish at volume into their CMS. CREA fits teams that want a content team drafting only from their own material, with a person approving every post. A fair side-by-side, checked 24 October 2026.