A page at position 12 gets almost no clicks. A page at position 5 gets several percent of every search. The gap between those two outcomes is usually one afternoon of work on a page that already exists, which makes page-two rankings the cheapest traffic in SEO and the most commonly ignored.
That is what a content refresh is for. Not rewriting old posts for the sake of it, but finding the specific pages Google already half-trusts and giving them the thing they are missing. Here is the process, from the Search Console filter to the date rules.
Why refreshing beats writing, once you have rankings
A new URL starts with nothing. Google has to discover it, crawl it, decide whether to index it, and then work out where it belongs. On a young domain that first step alone can take weeks.
This site is a live example. After moving hosts in September, Search Console listed 11 of its 12 URLs as unknown to Google, a week after the sitemap was submitted. None of those pages can be refreshed into page one yet, because none of them has a position to improve.
A page sitting at position 12 has already cleared all of that. It is indexed, it has a crawl history, whatever links it earned are counted, and Google has matched it to a set of queries. The refresh only has to close the gap between "relevant" and "best answer on the page", which is a much smaller job than getting from zero to relevant.
The rough click-through curve makes the payoff concrete. Position 1 typically takes a quarter or more of the clicks, positions 4 and 5 somewhere around 5 to 8 percent, position 10 around 2 percent, and page two well under 1 percent. Moving a page from 12 to 6 is not a small improvement. It is often the difference between no traffic and real traffic.
Find your striking-distance keywords
Search Console already holds the list. In the Performance report:
- Set the date range to the last 3 months and turn on all four metric tiles.
- On the Queries tab, open the table filter and add Position greater than 3.9, then Position smaller than 20.1.
- Sort by impressions, highest first.
- Click the top query, then switch to the Pages tab to see which URL is ranking for it.
Write each result into a list grouped by page, because you refresh pages, not queries. A page that ranks at positions 7 to 15 for six related queries is a much better refresh candidate than a page with one query at position 9.
| Page | Queries in positions 4 to 20 | Impressions (3 months) | Best position | Priority |
|---|---|---|---|---|
| /blog/invoice-late-fees | 6 | 4,200 | 7.4 | Refresh this month |
| /blog/freelance-tax-checklist | 2 | 1,900 | 11.8 | Next month |
| /pricing | 1 | 350 | 16.2 | Leave it |
Ignore anything with a handful of impressions. On a young site, a floor of about 50 impressions over three months keeps the list honest. The weekly version of this review is report 1 in the five Search Console reports post.
Diagnose why each page stalls
Refreshing without a diagnosis is how people spend a day on a page and see no movement. Each stall has a typical cause, and the fix depends on which one you have.
| What you see | Likely cause | Fix |
|---|---|---|
| High impressions, CTR well below what the position should get | Title and description do not match what the searcher wants | Rewrite the title around the query's specific hook |
| Ranks for queries you never targeted | The page answers them in passing | Add a dedicated H2 section that answers each one directly |
| Slid from the top 5 to position 12 over several months | Content decay: stale numbers, competitors updated | Replace data and examples, add what the new top results cover |
| Two of your URLs swap places for the same query | Cannibalization | Merge them into one page and 301 the other |
| Sits at 8 to 15, and the page looks complete | Authority gap against the rest of page one | Internal links from your strongest pages, then external links |
The last row is worth reading twice. If the weakest domain on page one is far stronger than yours, no amount of on-page work will close the gap. That is an authority problem, and the DR roadmap covers how long it takes to fix. Check it before you spend the afternoon.
The refresh recipe
For a page that stalls on content rather than authority, work through this in order:
- Search the main query and read page one again. Note what the top three cover that you do not: a section, a table, a calculator, a more recent number.
- Add a two to three sentence answer at the top. The direct answer to the main query, stated plainly, before any scene-setting.
- Give each striking-distance query its own H2. If the page ranks at position 11 for "how to charge late fees in the UK", a section with that heading and a real answer is usually what moves it.
- Replace every stale number and example. A 2023 statistic on a 2026 page is a reason for Google to prefer a newer page.
- Add an FAQ block of four to six questions, taken from the query list and from People Also Ask. It answers the long-tail variants directly, and it is the part AI answer engines lift most readily.
- Add two or three internal links from pages that already get traffic, with anchor text that matches the query.
- Keep the URL, the H1 topic and anything that already ranks. A refresh adds; it does not start over.
- Set the updated date, then request indexing once in URL Inspection.
Before you touch the page, write down its position for each target query and its clicks for the last 28 days. Without a before number, you will not be able to tell whether the refresh worked or whether the whole site simply moved.
The date rule
Keep the original publish date. Add a separate updated date, and only set it when the change is material: a new section, replaced data, a rewritten answer. Fixing a typo is not an update.
On this site the updated date is one frontmatter field, and it drives three things at once: the visible "Updated" line under the title, dateModified in the Article schema, and lastmod in the sitemap.
date: "2026-06-30"
updated: "2026-09-28"The reason to be strict is that Google compares your claimed dates with what actually changed. A sitemap where every URL claims to change on every deploy teaches Google to ignore your lastmod entirely, including on the pages where it is true.
Measure the move
Give it two to four weeks. Then compare the 28 days after the refresh with the 28 days before, filtered to the page, at the query level: position, impressions and clicks for each query you targeted.
Keep a short refresh log so the lessons add up:
| Date | Page | Queries targeted | Position before | Position after 28 days |
|---|---|---|---|---|
| 2026-06-12 | /blog/invoice-late-fees | late fee wording, uk late fees | 11.2 | 5.8 |
After five or six rows you will know which kind of fix works on your site, which is worth more than any general benchmark.
If a page has been refreshed twice with no movement, stop. Either the intent does not match (the results are tools and you wrote an article), or it is an authority gap. Both are reasons to spend the next afternoon elsewhere.
Put it on the calendar
Refreshing works best as a fixed slot rather than something you get to when there is time. Once you have a dozen posts with impressions, swap one new post a month for a refresh of the best candidate on your striking-distance list. The cadence stays weekly; only the output type changes, as described in the weekly content cadence post.
It is the same loop behind the promptbuilder.cc case study: Search Console shows which pages are stuck on page two, and those pages get worked on before the next new one is written. And when you pick the next new page instead, pick it the same way, from queries a young domain can actually win.
Want the full system, from keyword research to measurement, in one place? Grab the free guide, and follow the weekly breakdowns on X.
Frequently asked questions
- What are striking distance keywords?
- Queries where one of your pages already ranks between roughly position 4 and position 20. Google has decided the page is relevant but not yet the best answer, so a targeted update can move it onto page one or into the top three far faster than a new page could rank from nothing. You find them in the Search Console Performance report by filtering queries on average position.
- How often should you refresh old blog posts?
- Review the list monthly and refresh one or two pages per month once you have a dozen or more posts with impressions. Refresh a page when it sits in positions 4 to 20 with real impressions, when its position has slid over several months, or when its numbers and examples are out of date. Refreshing on a fixed timer, with nothing new to add, does nothing.
- Should I change the publish date when I update a post?
- Keep the original publish date and add a separate updated date, shown on the page and used for dateModified in your Article schema and lastmod in your sitemap. Only change it when the update is material: a new section, replaced data, a rewritten answer. Bumping the date on a post you barely touched is the kind of freshness claim Google learns to ignore.
- How long after a content refresh do rankings change?
- Expect two to four weeks. Google has to recrawl the page, reprocess it and re-evaluate its position against the rest of the results, and requesting indexing once in Search Console shortens the first part. Compare the 28 days after the refresh with the 28 days before it, at the query level, before deciding whether it worked.
- Is it better to update old content or write new content?
- Update when a page already ranks on page two for a query you want, because the crawl history, links and relevance are already in place. Write new when no page you have matches the query's intent, or when the results are gated by authority your domain does not have yet. Most sites past their first twenty posts get more from one refresh a month than from one extra new post.