How often to publish: what actually matters

There is no bonus for publishing weekly. Google's freshness systems favour recent pages on queries where recency matters, which is news, prices, releases and anything dated. For "how to fix a redirect chain", a page from last year that is still correct beats one published this morning that is not.

• A site that publishes weekly and a site that publishes monthly are judged on the pages, not the pace.
• What a rhythm gives you is compounding: more pages means more searches you can appear on, eventually.
• What it costs is attention: every page you publish is a page you now have to keep accurate.

So the question is not how often you can publish. It is how often you can publish something that deserves to exist and that you will maintain.

A page that answers nothing is not neutral. It is a URL Google has to crawl, judge and store, and a signal about what the rest of your site is like.

The page that will earn you the most clicks next month is usually one you already published. It ranks on page two, or it ranked last year and slipped. Refreshing it is faster than writing a new page and the outcome is more certain.

1. 1

   Find the falling pages

   In the Performance report, compare the last three months to the three before, sorted by the difference in clicks. The pages at the bottom lost something.

2. 2

   Separate lost demand from lost position

   Open the page's queries. Impressions fell too means the search itself shrank. Impressions steady with worse positions means the page fell.

3. 3

   Fix the facts first

   Screenshots, prices, product names, the interface that moved. A stale page loses to an accurate one whatever the writing.

4. 4

   Then the gaps

   Read the results page again. New questions appear over time, and a page that does not answer them is not the best result any more.

Count a refresh as a published page in your rhythm. It takes similar time and it usually returns more.

Pick the cadence you can hold in a bad month, not the one you can hold in a good week. A site that publishes twice a month for two years has forty-eight pages. A site that publishes eight in January and nothing after has eight.

Nothing breaks if you skip a month. There is no streak to protect and no algorithm counting. The only thing that decays is the accuracy of what you already published.

There are cases where volume is the answer, and they have one thing in common: each page holds something of its own.

• A glossary built from the terms your guides already use, where each entry has a definition you can make more precise than the pages that rank.
• Comparison and alternatives pages for tools your buyers actually weigh, where each holds a real judgement about who should choose what.
• Documentation, where every page is a task somebody has to perform.
• A family of pages built on data you hold and nobody else does.

The failure mode is the same list without the substance: two hundred pages generated from a template with the names swapped. That is the pattern the scaled content policy describes, and the pages that survive it are the ones that would have been worth writing one at a time.

Do not judge a page in its first month. Discovery and indexing take days to weeks for a new site, and ranking on anything competitive takes months.

• Month one: the page is indexed. Check the Page indexing report or inspect the URL.
• Month two or three: impressions appear on queries you did not target, which tells you what the page is actually about.
• Month four onward: clicks, if the search has demand and the page earns the position.
• Every month: the pages you published six months ago should be gaining, not flat. Flat after six months means the page answered nothing anybody searches.