When to refresh content and when to leave it alone
The refresh decision is not about whether content is old. It is about whether the investment changes your position in search or AI answers.
You have 140 published posts. Which ones do you refresh this quarter?
The refresh decision is not about whether content is old. It is about whether the investment changes your position in search or AI answers. I run this audit for most SME clients within the first two weeks, and the same mistake shows up almost every time. Teams refresh the wrong pages for the wrong reasons, or they refresh nothing because the backlog feels overwhelming.
Here is the decision framework I use.
Refresh candidates that earn their time
Three signals tell me a page deserves refresh work before anything else on the site.
The page ranks between positions 5 and 15. It already has authority. It is already being crawled. A focused update can move it onto the first page within weeks, while a brand-new post on the same topic will take months to reach the same position. Positions 5 through 20 carry the highest refresh ROI because they sit at the edge of visibility.
The page gets impressions but no clicks, or clicks but no conversions. Impressions without clicks mean the title or meta description no longer matches intent. Clicks without conversions mean the page body no longer delivers what the title promised. Both are fixable with a targeted refresh. Neither requires a rewrite from scratch.
The page used to rank or get cited, and now it doesn't. Traffic decay over 90 days is the clearest signal that a page is losing relevance. If you had a position and lost it, you can usually get it back faster than you can earn a new one on a different URL.
Pages that do not need a refresh yet
I see refresh backlogs full of pages that do not move the business forward. Refreshing them burns time you could spend on the pages that do.
Pages that rank on page one and still convert. If the page is pulling its weight, leave it alone unless the information is factually wrong. A position-one page does not need the publish date updated just because it says 2024. If it is still getting cited by ChatGPT and Perplexity, the engines do not care that it is 18 months old.
Pages with no impressions, no backlinks, and no conversion history. If a page has been live for 12 months and Google has shown it to almost nobody, refreshing it will not fix the underlying problem. The page either targets a query with no volume, or the site does not have enough topical authority for Google to rank it. In both cases, your time is better spent building authority elsewhere or retiring the page.
Pages ranking past position 50. A page in position 62 is not close to visibility. Moving it from 62 to 48 changes nothing for the business. If the topic still matters, write a new post that targets the current intent and redirect the old URL to it. If the topic does not matter, delete it and move on.
The mechanics of the decision
I score refresh candidates on a simple grid. Five factors, each weighted.
| Factor | Weight | What it measures |
|---|---|---|
| Position (5–15 is high) | 30% | Distance to page-one visibility |
| Traffic trend (90-day decay) | 25% | Whether the page is losing ground |
| Conversion value | 20% | Whether the page drives leads or revenue |
| Effort required | 15% | How much work the refresh will take |
| AI citation gap | 10% | Whether the page ranks but is not cited |
A page in position 11 with declining traffic and real conversion value scores higher than a position-four page that still converts. The decision is not about perfection. It is about where the next hour of work has the highest return.
In practice, I rarely refresh more than 10 to 15 pages per quarter for an SME with a two-person marketing team. That is enough to move the metrics that matter without turning refresh into a full-time job.
What refresh work actually looks like
When a page scores high enough to earn the time, the refresh itself is targeted. I am not rewriting the entire post. I am fixing the specific weakness that is keeping it from ranking or being cited.
For a page that ranks but is not cited by AI engines, the refresh is structural. Add a 40-word direct answer under the H1. Rewrite the intro to match the way people ask the question in ChatGPT or Perplexity. Add FAQ schema if it is missing. Update the visible modified date and request re-indexing. That is usually enough.
For a page with declining traffic, the refresh is about recency and relevance. Replace outdated statistics. Swap 2023 examples for 2026 examples. Check whether the SERP intent has shifted since you published. If Google now shows comparison listicles instead of how-to guides, you either restructure the page to match or you retire it.
For a page with impressions but no clicks, the refresh is on-page. Rewrite the title tag to match what the top three results are doing. Rewrite the meta description to include a clear benefit or outcome. That is it. The body might be fine.
I covered the step-by-step execution in how to write a content brief that gets cited in AI answers and building topic clusters when you only have one writer. This post is about the decision that comes before the execution.
The pages you should delete instead
Not every old page deserves a refresh. Some deserve a 301 redirect and a deletion.
Delete pages with no traffic, no backlinks, and no strategic value after 12 months. Delete pages that duplicate a stronger post on the same topic. Merge two weak pages into one strong one and redirect both old URLs to the new canonical. A site with 80 strong posts will outrank a site with 140 mixed-quality posts on the same topic.
Pruning is part of the refresh decision. If a page does not score high enough to refresh and it does not earn its place on the site, cut it.
The rhythm that works for small teams
The founders and marketing leads I work with do not have time to audit 140 posts every quarter. The cadence that works is simpler.
Review your top 30 pages by impressions and conversions every 90 days. Score them using the grid above. Refresh the top five to seven. Delete or redirect anything that has been live for a year with no traction. That is the whole system.
For high-value pages, pages that drive leads or revenue directly, I run a lighter check every 30 days. Verify that the statistics are still current. Check whether the page is still being cited by AI engines on your core queries. If it is, leave it alone. If it is not, that is your signal to refresh.
I help SME teams set this up as part of the 90-day content marketing cadence or through a consultancy engagement when they want the audit and the execution handled together. The decision framework itself takes 20 minutes to learn and about an hour per quarter to run once it is in place.
The refresh decision is not about discovering hidden gems. It is about directing your limited time toward the pages that will actually move your visibility in the next 90 days.