SEO guide · Chapter 5 of 6 · 5 min read
Publish and connect it
A page is published when Google can reach it, the right pages on your site link to it, and you have written down what changed. A page that loads perfectly in your browser can still be blocked, duplicated or wrong in what it sends to Google.
Updated
You leave with: A completed publication checklist and a change-log row.
What Google needs before it can index a page
Google's technical requirements come down to three things: Googlebot is not blocked from the page, the page returns a working response (HTTP 200), and it has content Google can index. Meeting them makes a page eligible. It does not make it rank.
Two settings get mixed up often. A robots.txt rule stops Google crawling a page, but a blocked page can still be listed if other sites link to it. To keep a page out of results, use a noindex tag and let Google crawl it so it can see the tag.
A sitemap tells Google which pages you consider important. It is a hint, and Google may still skip pages in it. Links from your own pages matter more, as long as they are normal links with descriptive text (crawlable links).
Publish in this order
Fix the final address
Decide the URL once. Check that the page's canonical points to itself.
Check the page
Title, description, H1, headings, image descriptions, facts and links, on a phone as well as a laptop.
Link to it
From two or three of your pages that already cover the topic and get visits, with words that describe it.
Link from it
To the service, the proof and the next thing the reader needs.
Confirm access
Not blocked in robots.txt, no stray noindex, included in your sitemap.
Publish after approval
The person who owns the facts signs off.
Inspect the live URL
In Search Console's URL Inspection, run a live test, then request indexing. A request is not a promise, and there is a daily limit.
Record it
Add a change-log row with the date and exact change (chapter 6).
Check again in a week or two
Confirm the page is indexed instead of assuming it is.
Your publication checklist
| Check | Who | Status | Evidence |
|---|---|---|---|
| Final URL and canonical are correct | Not started / passed / blocked | The live URL | |
| Title, description and H1 are unique and useful | Screenshot or page source | ||
| Every claim is in the claim check | Claim check | ||
| Images have useful descriptions | Page review | ||
| Related pages link to it | Their URLs | ||
| It links to the right service and next step | Destination URLs | ||
| Allowed in robots.txt, no noindex, in the sitemap | robots.txt and sitemap URLs | ||
| Works on a phone, and the contact button works | Phone test | ||
| URL Inspection run | Inspection date | ||
| Change-log row written | The row |
What URL Inspection does and does not tell you
The report shows the version of the page Google has stored, and the live test fetches it now. “URL is on Google” means the page is indexed and can appear. It says nothing about where it ranks. A passing live test does not mean the page will be indexed either: it does not judge duplicates, quality or which version Google treats as canonical. The rendered HTML in the live test is worth opening: it shows what Google actually received, which is not always what your browser shows you.
Worked example · rankrhq.com, raw HTML served to GPTBot ·
Our page looked right, and told crawlers the opposite
Our homepage animated two figures counting up from zero. In a browser they ended on the real values, 1 and 4. In the raw HTML sent to crawlers, including OpenAI's, both read 0, so any AI reading the page took away the opposite of our main claim. The page had shipped and looked fine to every person who visited. We fixed it to send the real figures in the HTML.
What this does not prove
- We found it by reading the raw HTML. Nothing in Search Console or our analytics flagged it.
Worked example · rankrhq.com, our change log ·
Two more that shipped wrong
On 29 September our 404 error pages were telling Google that the homepage was their canonical version. On 4 October we found every blog post's share card returning an error, so links shared on social media had no preview. Both pages worked in a browser. Both were caught by checking what machines receive.
What this does not prove
- Neither had a measured effect on rankings that we could isolate. They were fixed because they were wrong.
Internal links: where to add them
Find your pages that already mention the topic and add a link with descriptive words. A quick way is to search Google for site:yourdomain.com topic, then link from the results. Google asks that every page you want found is linked from at least one other page on your site (link best practices). When we crawled our own site on 20 September, the two pages carrying 43% of our impressions had only 2 internal links each, against a median of 5 for our blog posts. We fixed how related posts are picked. Internal links help Google find and understand pages; on their own they will not lift a weak page to page one.
Common mistakes
- Treating “deployed” as “indexed”.
- Blocking a page in robots.txt to hide it.
- Publishing an orphan page that nothing links to.
- Forgetting to record the date, so later changes cannot be explained.
Your 20-minute task
Run the publication checklist on your newest page, and inspect it in Search Console.
Sources
- Google:Google Search technical requirements
- Google:Introduction to robots.txt
- Google:Block indexing with noindex
- Google:Learn about sitemaps
- Google:Sitemaps report
- Google:URL Inspection tool
- Google:Make your links crawlable
Ahrefs sells an SEO tool. We cite it for practical methods and check its numbers against the live results.