The Content Refresh Playbook: How to Update Pages That Are Losing Traffic

This playbook shows you how to refresh pages that are losing search traffic: how to choose between refreshing, rewriting, merging and removing, how to rank a long list of candidates, what to change on each page, how to handle dates and recrawling the way Google documents it, and how to tell afterwards whether the work paid off. It assumes you already know which pages are slipping; if not, start with what content decay is and how to detect it.

A content refresh is a deliberate, scoped update to an existing page, based on a diagnosis of why it is underperforming. It is not changing the date and a sentence, and it is not rewriting a page that was fine. Done well, it is often the fastest way to recover organic traffic, because the page is already indexed, already linked and already has a ranking history.

Why refresh before you publish more

A new page starts with no links, no history and no proof that it serves the query. A declining page already has those, and usually lost ground for a specific reason you can find and fix. Updating it removes that friction and keeps everything it built up.

The best-known public data point comes from HubSpot’s own blog, which reported that 76% of its monthly blog views came from posts published before that month, and that updating and republishing old posts raised their monthly organic search views by an average of 106%. That is one company’s blog, measured years ago, so read it as a direction rather than a benchmark. The underlying point holds for most mature sites: the pages you already have usually carry most of the traffic, so maintaining them is not a side task.

None of this replaces new content. New topics still need new pages. The question is allocation: if almost all of your team’s time goes to new pages and almost none to the ones that bring in most of the traffic, moving even part of that time to maintenance is usually the better bet.

Refresh, rewrite, consolidate or prune

The first decision for every candidate is the type of work. Defaulting to “refresh everything” wastes effort on pages that need a new format; defaulting to “rewrite everything” throws away what still works.

For consolidation, keep the strongest URL, merge the best material into it, and use a permanent (301 or 308) redirect: Google treats a permanent redirect as a signal that the target should be the canonical URL. For pruning, remember Google’s core updates guidance calls deleting content a last resort, for pages that cannot be salvaged. The consolidation guide covers both in depth.

Finding refresh candidates

Combine three kinds of evidence. Pages that show up in two or more are the strongest candidates.

• Search Console trends. Impressions down 20% or more year over year, average position worse by two or more places over three months, CTR down 25% or more at a similar position, or fewer queries with impressions. These are rules of thumb; the decay guide explains each signal and how to rule out seasonality and ranking updates first.
• Quality gaps on pages that still earn traffic. A page with real demand and weak scores on effort, originality or accuracy is a refresh waiting to happen.
• Results-page gaps. What do the top three results cover that you do not, in what format, and how recently were they updated? See competitor benchmarking.

Hold back pages that dropped in a step change inside a ranking update window. For a large drop, Google’s core updates guidance starts with a broader look at whether the site as a whole is helpful, then at the most affected pages; for a small one, it advises against drastic changes. A page-by-page refresh is not the first move in either case.

Ranking the candidates

With 30, 50 or 200 candidates, gut feel picks the pages with the biggest traffic or the ones the team likes. A simple weighted model does better. Score each page 0 to 10 on four factors, multiply by the weight and add up. The weights are a starting point, not a standard; adjust them once and keep them for the cycle.

1. Business value (40%). Key events or pipeline from the page if you have them. Otherwise, a traffic value: clicks multiplied by the cost per click of the main query. For example, 500 clicks a month on a query with a €4 CPC is about €2,000 a month in equivalent ad spend.
2. Decay velocity (25%). Month-over-month change in clicks or impressions over the last three months. Fast decliners score high: the window for an easy recovery is closing.
3. Recovery probability (20%). High for a freshness gap on a well-linked page; low for a thin page whose query is now dominated by tools.
4. Effort, inverted (15%). A facts-and-links update scores high; a full restructure scores low. This keeps quick wins next to the big bets.

Worked example

Three hypothetical candidates, with illustrative scores:

Page B goes first even though A has more traffic: it is losing ground faster, the fix is cheap and the cause is clear. Page C needs a decision about whether you want to build a tool at all, which belongs in planning, not in the refresh queue. Override the model when context demands it, for example a page tied to next month’s product launch.

dash.morrison.app/lumenpay.com/chat

Illustration

lumenpay.com/Refresh priorities

What do you want to know about lumenpay.com?

Ask about your website content...

Search ConsoleSerpApiContent Refresh Brief Builder

Claude Sonnet 5.5

The refresh, step by step

1. Capture the baseline, including what to protect

Before editing, record the last 90 days from Search Console (clicks, impressions, average position, CTR) and from GA4 (sessions, engagement rate, key events). Save the page as it is and screenshot the results for the main query. Then list what is working: the queries that bring most clicks, any section that holds a featured snippet, and the paragraphs that match valuable long-tail queries. That list is your “do not break” list. Google’s core updates guidance makes the same point from the other side: avoid changing content that is already performing well.

2. Diagnose the decline

• Stale facts: outdated statistics, prices, product names, dead links. See outdated content finder.
• Competitive gap: competitors now have original data, examples, tables or templates that you lack.
• Intent shift: the results page moved from explanations to comparisons or tools, or the reverse.
• Search features: CTR fell at a stable position because more sits above the organic results.
• Technical erosion: lost internal links, a canonical or redirect change, a slower template.
• Thin from the start: the page never had the depth to hold its place once competition arrived. See thin content identification.

3. Read the results page

Search the main query and study the top five: format, rough length, last update, subtopics covered, media and tables, and any search features. This is calibration, not imitation. If every top result has a comparison table and yours does not, that is a gap. If they were all updated this year and yours was last touched three years ago, freshness is part of the story. Google’s ranking systems guide describes “query deserves freshness” systems for queries where fresher content is expected, so how much recency matters depends on the query.

4. Write the refresh brief

One short brief per page, before anyone edits. It should name:

• The diagnosis, in one sentence
• Facts to replace, with the new source for each
• Sections to add, based on the results-page reading
• Sections to cut or merge where the page has bloated
• The “do not break” list from step 1
• New title, description and heading changes
• Internal links to add, both to and from the page
• Who reviews accuracy, and by when

A scoped brief prevents both failure modes: changing too little to matter, and rewriting sections that were earning clicks. The guide to content briefs covers brief writing in general.

5. Update the content

Make the page more accurate, more complete and more useful than it was and than what outranks it. Running text through a paraphrasing tool is not a refresh. Google’s helpful content questions are a good checklist: original information, complete coverage, clear sourcing, evident expertise. Changes that tend to matter:

• Replace dated statistics with current ones and cite the primary source
• Add the subtopic the top results cover and you do not
• Update products, prices, features and screenshots
• Add a worked example, a table or a template where the results page favors them
• Add first-hand detail or expert input that competitors cannot copy
• Name the author and their expertise if the page does not already

6. Update metadata and structure

• Title and description:they should describe the refreshed page and compete for the click against today’s results. See title and meta optimization.
• Headings: a logical, scannable hierarchy, with headings for new sections.
• Internal links: link out to newer related pages, and link in from strong pages on your site. See internal linking at scale.
• Structured data: keep Article markup accurate, including dateModified. Do not add markup for features Google no longer shows: How-to rich results were removed in 2023, and FAQ rich results stopped appearing in Search from 7 May 2026. See schema markup opportunities.

7. Dates: change them only when the content changed

Google’s self-assessment questions include whether you change the date of pages to make them seem fresh when the content has not substantially changed: a sign of content made for search engines first. When the update is substantial, follow Google’s byline date guidance: show a visible, clearly labelled date such as “Last updated”, and make dateModified in your structured data match it. In the sitemap, lastmod should reflect the last significant update; Google says it uses the value only if it is consistently accurate, and treats changes to main content, structured data or links as significant, while a new copyright year is not. For a change-by-change table and CMS rules, see what counts as a real content update.

8. Help Google find the update

For a handful of important pages, use URL Inspection in Search Console to request indexing. Google’s recrawl documentation says crawling can take from a few days to a few weeks, there is a quota for individual requests, and asking again for the same URL will not make it faster. For many pages at once, rely on an accurate sitemap with correct lastmod values.

9 and 10. Measure, then log

Then wait, and do not keep editing: if you change the page again during the observation window, you cannot tell which change did what. Log what you changed, when and why, so the next refresh starts from evidence rather than memory.

Seasonal and evergreen pages need different clocks

Evergreen pages(“how to write a business plan”) decay slowly. Refresh them when signals say so, and otherwise on a review interval that fits their type. Seasonal pages(“Black Friday deals”, “tax filing checklist”) must be refreshed before the demand peak, not during it. Since recrawling can take days to weeks, a lead time of four to six weeks before the peak is a sensible rule of thumb. See seasonal content planning.

Measuring whether it worked

Compare three windows for every refreshed page: the 90 days before the refresh (the declining state), a recovery window of roughly 30 to 90 days after, and the page’s best 90 days (the target). Track clicks, impressions, average position, CTR and key events. Beating the baseline shows the refresh had an effect; closing the gap to the peak shows how much of the loss you won back.

On timing, Google’s own statements set expectations: recrawling takes days to weeks, and the core updates guidance notes that some changes show effects within days while others take several months to be recognized. A practical rule of thumb is to read early movement after two to four weeks and judge the result at eight to twelve weeks. Checking after one week and declaring failure is as unhelpful as never checking.

To separate the refresh from everything else going on, compare refreshed pages with a few similar pages you did not touch in the same period. If both groups rose, the market moved; if only the refreshed group did, the work did. At portfolio level, report pages refreshed, the share that recovered to within 80% of their peak, median time to recovery and total clicks recovered. See content migration planning and the measurement guide.

Making it a habit

One round of refreshes gives one bump. A cadence compounds. Each quarter, rerun the candidate search, re-rank with the model, and feed the top of the list into the editorial calendar alongside new content. Give every important page an owner and a review interval. Split the work by skill: the SEO lead owns the model and the results-page reading, subject experts own the substance, editors own quality checks and publishing. See content refresh prioritization and content decay detection.

Common mistakes

1. Over-editing what works.Rewording the section that holds a featured snippet or matches a valuable query can lose it. Protect the “do not break” list.
2. Polishing the wrong format. Updating facts on an essay when the results page now rewards a tool or a table will not close the gap.
3. Changing the date without changing the content. Google lists it as a warning sign. If the substance did not change, the date should not either.
4. Not logging changes. Without a log, every result is a guess.
5. Refreshing easy pages instead of valuable ones. Short pages are tempting. The model exists to stop that.
6. Treating refresh as a one-off project. Without a cadence, the same pages decay again.

Where Morrison fits

Everything here works with Search Console exports, a spreadsheet and careful reading. Morrison takes out the reading and joining work at scale. Its content analyzer scores every crawled page on eight dimensions from Google’s helpful-content and spam guidance, each score opens to the checklist behind it with what to change, and one click puts a prompt in the chat composer that asks for a rewrite plan from the page’s latest scan. In chat you can paste a competitor’s URL and Morrison scores it the same way, pull Search Console queries for the page, and check the live results through SerpApi; the Content Refresh Brief Builder Skill turns that into a brief covering what to update and why. For a whole section, use a workflow: for example, pull each URL’s top queries from Search Console and live results from SerpApi, then have an AI step draft a new title and description, with each result written into a column in the URL grid. After publishing, an automation can report on the refreshed pages every week (weekly or monthly on Starter, up to daily on Pro, up to hourly on Business), reading the previous run so it says what changed. The edits themselves stay with your team, in your CMS, with your reviewers.

Key takeaways

• Diagnose first. Refresh, rewrite, consolidate and prune solve different problems.
• Rank with a model. Value, velocity, recovery probability and effort, weighted and written down.
• Protect what works. Record the queries and sections earning clicks before you edit.
• Handle dates honestly. Change the visible date, dateModified and lastmod only for substantial updates, and keep them consistent.
• Be patient, then measure properly. Days to weeks for recrawling, eight to twelve weeks to judge, with untouched pages as a comparison.
• Make it a cadence. Quarterly re-ranking, owners and review intervals turn maintenance into routine.