Updating one statistic in an existing page is usually the cheaper test. You change a fact, keep the URL, and watch whether impressions and clicks move. Writing a new article is a different bet: a new URL with no history, competing for the same query. The decision should rest on signals you can actually observe in Search Console and your analytics, not on a hunch about "fresh content". This piece is a measurement explainer. It walks through the numbers you can see, the ones you cannot, and how to run the comparison without guessing.
A content refresh is any edit to a live page — a statistic, a date, a section rewrite — made to improve how it performs for the queries it already targets. The old blog post is the page already indexed. A new article is a new URL with no ranking history. Put simply, a refresh keeps the URL and its history; a new article starts from zero.
Why does one statistic count as a refresh?
A refresh doesn't have to be a rewrite. Google's Search Essentials documentation says pages should be created primarily to help people, and that thin or copied pages add little value. Updating a stale figure is a small, low-risk way to make a page more useful without touching its structure or URL.
The measurement value comes from keeping everything else constant. If you change one number and nothing else — no title change, no URL change, no internal link changes — any movement in impressions or clicks is more likely to be tied to that edit. That is the closest you get to a controlled test in SEO.
There is no public Google statement that a single-statistic edit will improve rankings. So treat the edit as a hypothesis, not a promise. The point is to make the hypothesis testable.
Which signals can you actually observe after an update?
You can observe four things with reasonable confidence: impressions, clicks, average position, and click-through rate (CTR) in Google Search Console. You can also observe on-page engagement in your analytics — scroll depth, time on page, and conversion events — if those are set up.
What you cannot observe directly: whether Google re-crawled the page, how it weighted the change, or whether a competitor's change affected your position. Those are invisible. The crawl part you can infer from server logs. Google's Search Console Help documents the crawl stats report, and your own log files show when the Googlebot user-agent requested the URL.
| Signal | Where you see it | Observable? | What it tells you |
|---|---|---|---|
| Impressions | Search Console → Performance | Yes | Whether the page appears for more or fewer queries |
| Clicks | Search Console → Performance | Yes | Whether the snippet and title still attract clicks |
| Average position | Search Console → Performance | Yes | Whether ranking moved, but it is an average across queries |
| CTR | Search Console → Performance | Yes | Whether the result is more or less appealing at the same position |
| Re-crawl date | Server logs | Inferred | When Googlebot last fetched the page |
| How Google weighted the edit | Nowhere public | No | Not measurable; treat as unknown |
This table is the core of the explainer. The left column is what you can put in a report; the right column is what you have to admit you don't know. A refresh decision built only on the observable column is honest and repeatable.
What does a new article cost you in time and history?
A new article starts with no impressions, no clicks, and no link equity pointing at its URL. It also needs its own internal links, its own metadata, and its own place in your sitemap. A refresh inherits all of that from the existing page.
Cost is not just writing time. A new article needs keyword research, a brief, a draft, an edit, a publish step, and then a waiting period before any data is meaningful. A refresh of one statistic needs a fact-check and a publish step. The difference in hours is large. The difference in risk is also large, because a new URL can fail and a refreshed URL keeps its history.
There is no public benchmark for how long a new page takes to rank. I would not invent one. What I can say is that the existing page already has a crawl history and possibly links, so the refresh path has less to prove.
When should you write new instead of refreshing?
Write new when the query is different. If your existing page targets "how to update a blog post" and you want to target "content refresh checklist for B2B SaaS", that is a different intent and a different page. Refreshing the old post to cover both usually weakens both.
Write new when the old page has no meaningful history: near-zero impressions over a full quarter, no internal links, no external links. In that case there is little history to protect, and a new page with a clearer angle can be the better bet.
Write new when the topic is time-sensitive and the old page is structurally wrong — for example, it was written for a product you no longer sell. A statistic update won't fix a page whose premise is outdated. The decision framework is: same query + existing history = refresh; different query or no history = new.
How do you run the comparison without fooling yourself?
Pick one page. Record its Search Console metrics for the 28 days before the edit: impressions, clicks, average position, CTR. Make the single-statistic change. Record the same metrics for the 28 days after. Compare the two windows.
Do not compare a 7-day window to a 28-day window, and do not compare December to January — seasonality will swamp the signal. Google's Search Console Help notes that performance data can be incomplete for recent days, so leave a few days of lag before you read the after-window.
Also check the crawl log: did Googlebot fetch the URL after the edit? If it did not, the change has not been seen yet, and any flat result means nothing. If it did, and the metrics moved, you have a candidate cause — not proof, but a candidate. That is the honest limit of what a single-page test can tell you.
What can't you measure, and why does it matter?
You cannot measure how Google's ranking systems weighted the edit, whether a competitor changed their page at the same time, or whether the query mix shifted. Any of those can move the numbers without your edit being the cause.
You also cannot measure the "freshness" boost as a separate quantity. Google's Search Central documentation discusses freshness as one of many signals, but does not publish a formula. Treat freshness as a factor, not a dial you can turn.
This matters because a single-page test is weak evidence. It is still better than no test, and it is cheaper than a new article. The right reading is: if the refresh shows a clear move and the crawl log confirms a fetch, repeat the pattern on two more pages before you call it a strategy.
How often should you refresh versus publish?
There is no public cadence rule from Google. Any "refresh every 90 days" advice is a convention, not a standard. Set cadence by content type instead: evergreen explainers get a fact-check every six months; product pages get a check when pricing or specs change; news-style posts get archived rather than refreshed.
Publishing cadence is a separate decision. A steady flow of new articles builds topical coverage; a steady flow of refreshes keeps existing coverage accurate. Most B2B sites need both, but the refresh work is cheaper per page and easier to measure.
Key Facts at a Glance
- A content refresh keeps the URL and its history; a new article starts from zero.
- You can observe impressions, clicks, average position, and CTR in Google Search Console.
- You cannot observe how Google weighted the edit, competitor changes, or query-mix shifts.
- Check server logs for a Googlebot fetch before reading any metric change.
- Compare 28-day windows before and after the edit; avoid seasonality traps.
- No search engine publishes a refresh cadence; treat fixed intervals as conventions.
Should you update old blog posts or write new ones?
Update when the page already ranks for the query you want and has some history. Write new when the query intent is different or the old page has no meaningful impressions or links. The observable signals — impressions, clicks, position — tell you which case you are in.
How often should you refresh content for SEO?
No search engine publishes a cadence. A practical default is a fact-check every six months for evergreen explainers, and on-change checks for product or pricing pages. Treat any fixed interval as a convention, not a rule.
Does updating old content hurt rankings?
It can, if the update changes the page's topic or removes content that matched the query. A single-statistic edit is low-risk because it changes little. A full rewrite is higher-risk because it changes more at once, and you cannot isolate the cause if traffic dips.
How do you know which posts to refresh?
Look for pages with stable impressions but falling CTR, pages ranking on page two for a target query, or pages with outdated figures. Search Console's Performance report shows all three. Pick one page, run a single-statistic edit, and compare 28-day windows before and after.
What is the difference between a content refresh and a rewrite?
A refresh edits part of a live page and keeps the URL. A rewrite replaces most of the content, often with a new title and structure. A rewrite carries more risk because more variables change, so it is harder to attribute any movement to a single cause.