AXENOR
Publish no faster than you can also keep your existing pages current. Publishing frequency is not a ranking signal, as Google's John Mueller said in 2018, and Google's guidance still says adding pages to look fresh will not help. A library of 30 pages on a quarterly refresh cycle needs about two refreshes a week before anything new is written. Axenor Consulting publishes twice a week, on Tuesday and Thursday, because that is the rate our refresh load leaves room for.
The advice to publish every day is everywhere, and the arithmetic behind it is nowhere. Publish daily, the advice runs, and the traffic follows. Nobody who gives it says what happens in month seven, when the library you built needs maintaining and the same team still has clients to serve.
We nearly made that mistake ourselves. The plan for this site was one article every weekday. We changed it on 22 September 2026, before the second article went out, once we had done the sum below.
Is daily blogging good for SEO?
No, not as a cadence. Publishing frequency is not a ranking factor, and on this question Google's answer is plain and old.
In March 2018 Google's John Mueller said publishing frequency is not a ranking signal, adding that "a site isn't a machine that pumps out content at a fixed rate." Google's guidance on creating helpful content (updated December 2025) asks whether you are adding a lot of new content mainly to make the site seem fresh, and answers: "No, it won't."
The sharper edge is policy. Google's spam policies define scaled content abuse as pages "generated for the primary purpose of manipulating search rankings and not helping users", however they were made. Twenty pages that each answer a different question with checked sources sit outside that definition. Twenty rewrites of one keyword page sit inside it.
So cadence earns nothing on its own. Each page has to deserve its place first.
What earns rankings and AI citations instead?
Three things, and volume is not one of them.
Distinct questions. More pages help only when each one answers something the others do not. Could you write a different, specific question at the top of every page without two colliding? Google's guide to generative AI features (July 2026) names the trap: separate content for every variation of how people search, made mainly to manipulate rankings, breaks the scaled content abuse policy.
Depth on one subject. Semrush tracked 50,000 brands across 1,094 categories in ChatGPT and reported in July 2026 that 85% of categories have no clear owner. Domain strength barely predicted who did: the owner had the higher Authority Score in 52.5% of pairs, close to a coin toss. What a small firm controls is a complete cluster, one main guide with the narrow questions answered around it, as set out in our guide to topical authority and topic clusters.
Freshness, on a cycle. This is the part that changes the arithmetic. Ahrefs found in July 2025 that AI assistants cite pages 25.7% fresher than Google's organic results. AirOps reported in December 2025, without publishing its method, that pages left more than three months without an update were over three times more likely to lose visibility. A page you never touch again decays, as we explain in our guide to content decay and the 13-week refresh.
How long should a blog post be for SEO?
As long as the question needs, and no longer. Google's helpful-content guidance asks whether you write to a word count because you heard Google prefers one, and answers: "No, we don't." Its generative AI guide adds: "There's no ideal page length."
The citation data agrees. Ahrefs analysed 174,048 pages cited in Google AI Overviews (December 2025): the median ran 1,115 words, 53.4% were under 1,000, and the correlation between length and citation was 0.04, close to zero. Orbit Media's September 2026 survey of 1,042 content marketers puts the average post at 1,312 words, down from 1,427 in 2023. It also links six or more hours on a 2,000-word post to roughly double the odds of strong results: effort pays, but length is not shown to cause anything.
Our articles run 800 to 1,500 words before the questions at the end, because that is what one question and its evidence take. A post padded to 2,000 words also takes longer to refresh.
How often should you blog for SEO? Start from the refresh load
Blog at the rate your refresh load leaves room for. Here is the calculation that set our own cadence, the one the daily-blogging advice never shows.
Take the quarterly refresh cycle as given, because the citation evidence above points that way. Every page you publish joins a library that needs revisiting every thirteen weeks. So a library of thirty pages is about two refreshes a week, for as long as the pages stay live, before you write anything new.
Now publish daily. Five new pages a week is about twenty-two a month. After six months you are carrying 130 pages: ten refreshes a week on top of the five new ones. At 60 to 90 minutes a refresh, that is 10 to 15 hours of maintenance a week before any new writing. No small team does that, so the refreshing quietly stops, the library ages, and the daily output is spent replacing ground you already held.
| Cadence | New pages a week | Library after six months | Refreshes due a week | Pages to handle a week |
|---|---|---|---|---|
| Every weekday | 5 | 130 | 10 | 15 |
| Twice a week | 2 | 52 | 4 | 6 |
| Once a week | 1 | 26 | 2 | 3 |
| Twice a month | about 0.5 | 12 | about 1 | about 1.5 |
The table starts from an empty blog and refreshes every page once a quarter. It is arithmetic, not a measurement, and you can redo it with your own numbers.
At twice a week the library takes fifteen months, not six, to reach 130 pages. That gap is what keeps the rate sustainable: it leaves time for the quarterly review below to merge or retire the pages that earn nothing, so the refresh load grows more slowly than the archive. That is the whole argument. It is not about discipline or ambition. A publishing rate you cannot maintain turns, within a few months, into a slowly decaying archive.
What do we publish, and how do we judge it?
Two articles a week, on Tuesday and Thursday at 09:00 Cyprus time, each answering one question a Cyprus business owner asks, with something in it we measured ourselves. No two answer the same question, and if a piece is not ready the slot stays empty. The time serves distribution, not ranking: the familiar best-time-to-publish advice comes from a 2017 CoSchedule round-up of older US studies of traffic, comments and social shares.
Each article is researched and drafted with AI assistance, then edited and fact-checked by the person whose name is on it. Every statistic links to its source and gives its date, and a figure we cannot trace to a primary source does not run. That rule exists because we broke it: four of our June guides quoted a freshness statistic with no traceable origin, and the refreshes replace it with numbers you can open.
Clicks alone are the wrong yardstick now. Ahrefs measured a 58% lower click-through rate for the top-ranking page when an AI Overview appears (December 2025 against December 2023), so a page can be quoted to a buyer and never clicked. We judge each article on five numbers instead:
- impressions and average position in Search Console;
- impressions inside AI Overviews and AI Mode, from Search Console's generative AI report;
- citations in a fixed panel of 25 buyer questions put to ChatGPT and Perplexity;
- days from publishing to the first impression;
- enquiries that mention an article.
An article that earns nothing gets one of three outcomes at the quarterly review. At positions 11 to 30, it is refreshed with a sharper direct answer or a missing sub-question. With nothing after 60 days, it is merged into the guide it supports and redirected. If it duplicates a page that does the job better, it goes. A site is judged on its weakest pages as well as its best.
How do you set your own blogging cadence?
Start from the library, not the calendar:
- Count the pages you intend to keep alive.
- Divide by thirteen. That is your weekly refresh load before anything new.
- Multiply by 60 to 90 minutes to see the hours it takes.
- Subtract those hours from the time you have each week.
- Divide what is left by the hours a new article takes. That is your real publishing rate.
For a professional firm in Cyprus that arithmetic tends to land between two articles a week and two a month, and the lower end is fine. Eight connected articles on one subject will out-pull fifty scattered ones. If you would rather have the structure mapped before you commit, that is the first week of an SEO and AI search engagement, and it starts with the questions your buyers ask.
Key takeaways
- Publishing frequency is not a ranking signal, and Google says adding pages to look fresh won't help.
- Distinct questions, depth on one subject and a quarterly refresh earn citations.
- Set your cadence from the refresh load: pages kept alive, divided by thirteen. Thirty pages means about two refreshes a week.
- Length follows the question. The median page cited in AI Overviews runs 1,115 words.
Frequently asked questions
Is daily blogging good for SEO? Not as a cadence. Google said in 2018 that publishing frequency is not a ranking signal. What helps is a larger set of distinct, useful pages, each answering one question well. Daily publishing is only worth it if every piece clears the same quality bar as a monthly one and you can still refresh everything you have already published.
Will Google penalise a site for publishing too frequently? Frequency itself is not penalised. Google's spam policies target scaled content abuse: pages produced mainly to manipulate rankings, with little value to readers. Articles with original data and checked sources sit outside that definition. Near-duplicates of one keyword page sit inside it, however they were produced.
What blogging frequency suits a small business? Work backwards from the refresh load. A quarterly cycle on a library of thirty pages is about two refreshes a week before anything new, so a small firm tends to land between two articles a week and two a month. Fewer pages, deeper and kept current, beat more pages left to decay.
How long should a blog post be for SEO in 2026? As long as the question needs. Ahrefs found the median page cited in Google AI Overviews ran 1,115 words and more than half were under 1,000, with almost no correlation between length and citation. Google says it has no preferred word count. Our articles run 800 to 1,500 words.
When should you update old blog posts? On a schedule set by how fast the topic moves. We refresh statistics, AI search and platform pages every 13 weeks, and stable definitions every 26 weeks. Change the visible date only when the content has changed: Google asks whether you change dates to make pages seem fresh when nothing substantial has changed.
Does the time of day you publish matter? There is no evidence it affects indexing or ranking. The familiar advice about weekday mornings comes from older US studies of traffic, comments and social shares, gathered in a 2017 CoSchedule round-up. Timing matters for distribution, so publish when your LinkedIn or email push goes out, and keep it consistent.
