SEO Blog Publishing Checklist

Twenty posts, no traffic, and nothing obviously wrong

The archive looks respectable. Twenty relevant articles, competent writing and clean templates. The audit finds a different story: no older page links to the newest posts, three articles use near-identical headings for different intents, full-resolution images delay rendering, and two useful posts carry a noindex instruction left over from staging.

None of those failures began in the draft. They appeared between approval and retrieval. The CMS was treated as a publish button rather than the place where titles, URLs, headings, links, images, index controls, canonical signals, structured data and measurement become one page.

Publishing hygiene cannot make demand exist for a query nobody uses, rescue an unhelpful article or guarantee ranking. It can make strong content findable, legible and connected—and prevent avoidable technical settings from cancelling the work.

What publishing actually is

Blog publishing is everything between an approved draft and a page that users and authorised search crawlers can find, read, navigate and measure. That job can be delegated because it has observable stages:

1. receive the approved copy, target intent, owner, sources, assets and destination;
2. build the entry in the correct CMS template without losing semantic structure;
3. prepare the title element, visible heading, description, slug, author, dates and categorisation;
4. optimise and place images with meaningful alternative text where appropriate;
5. add and validate contextual internal and external links;
6. check canonical, robots, sitemap and structured-data behaviour as relevant;
7. review desktop and mobile rendering, accessibility basics and interactions in staging;
8. publish under change control, inspect the live URL and submit or monitor through the established search workflow; and
9. record the URL, target cluster, links, date and baseline for later review.

The publishing owner does not silently rewrite strategy, legal claims or subject-matter conclusions. They surface conflicts to the editor or marketing lead. A clean workflow separates approved content changes from mechanical implementation.

The pre-publish SEO checklist

Keep the checklist in the work system, not in one person’s memory. Record pass, fail, not applicable and the reviewer.

Intent, title and visible structure

• Confirm one primary query intent and the page’s distinct job in the site.
• Use one clear, visible H1 that accurately names the page; check that the CMS template does not add a second H1.
• Write a unique, descriptive HTML <title> aligned with the page and visible heading. Google’s current title-link guidance explains that title links are generated automatically from several sources, including the title element, H1, prominent text and links. The entered title is an input, not a guaranteed display.
• Arrange H2 and H3 headings as a readable outline. Headings should answer the reader’s questions, not repeat keyword variants mechanically.

Description, URL and page identity

• Write a unique meta description that accurately summarises the value of this page. Do not enforce a mythical universal character count: Google’s current snippet guidance says there is no fixed meta-description length limit and snippets may be generated from page content and truncated for the device.
• Use the approved, readable slug; avoid changing an established live URL without a redirect and migration decision.
• Set the correct author, original publication date and modified date according to editorial policy. Do not change a date when the substance did not change.
• Check category, breadcrumb and archive placement so the page is not isolated.

Images and rendering

• Export each image close to its intended display dimensions in an appropriate modern format supported by the platform; preserve a quality master separately.
• Declare width and height to reduce layout movement, use responsive sources where the CMS supports them and avoid lazy-loading the main above-the-fold image blindly.
• Write descriptive alt text for images that convey information. Use empty alt text for genuinely decorative images so assistive technology can ignore them.
• Use meaningful filenames, captions when they help comprehension and rights/source records where required.
• Inspect the page at narrow and wide widths for overflow, unreadable type, cropped content, broken tables and intrusive overlays.

Links, sources and trust

• Add contextual links to relevant existing pages using descriptive, concise anchor text.
• Identify at least two older, relevant pages that will be edited to link into the new article after publication.
• Link factual claims to current primary sources. Open every external link and confirm it supports the nearby statement.
• Check redirects, fragments, protocol and final status. Remove draft, tracking or preview URLs.
• Confirm the author, editorial review and contact routes appropriate to the subject.

Technical search controls and measurement

• Confirm the live page returns the intended successful response, is not protected by staging authentication and is not blocked unexpectedly.
• Inspect robots meta and HTTP headers, canonical URL, sitemap inclusion and the rendered HTML.
• Add only structured data that matches visible page content and the supported type. Validate syntax and required properties.
• Confirm analytics and consent behaviour under the organisation’s approved configuration; avoid publishing duplicate tracking.
• Save the target cluster, live URL, publication date and internal-link actions in the content register.

The two checks people skip

Indexing instruction: inspect robots meta tags and HTTP headers on the live response. Google’s official noindex documentation says a page with a discoverable noindex rule is removed from Google Search after crawling. It also warns that if robots.txt prevents crawling, Google cannot see the noindex rule. “Blocked” and “not indexed” are different controls.

Canonical signal: confirm a self-referential canonical for the approved primary URL unless the architecture intentionally consolidates the page elsewhere. Google describes rel="canonical" as a strong signal, not a command, and advises internal links to use the canonical URL in its canonical guidance. A template pointing every post to the blog hub can cause serious consolidation confusion.

Internal linking as a publishing task, not a project

Every new article should leave the archive better connected. At publication, link out to two or three genuinely helpful existing resources in the body. Then edit at least two relevant older pages to link into the new post. Those numbers are a manageable editorial rule for this workflow, not a Google ranking formula.

The reverse-link step is the one most teams omit. A new article can mention established pages because the publisher sees them during production; older pages do not update themselves. Without retro-linking, the new URL may exist only in a feed, category or sitemap and receive little contextual support.

Google’s current link best-practices documentation says crawlable links help Google find pages, that descriptive anchor text helps people and Google understand them, and that every important page should have a link from at least one other site page. Use ordinary <a href> links, descriptive anchors and the canonical destination. Do not replace useful prose with “click here” or create dozens of unrelated keyword links.

Keep a small link map with source URL, destination URL, anchor, section, date added and reason. Review broken or redirected destinations monthly. When a page is merged or retired, update important contextual links rather than relying indefinitely on redirect chains.

The refresh cycle that compounds

Run a quarterly content review using comparable Search Console periods, editorial change logs and business priorities. Candidate signals include declining impressions or clicks, queries the article partly answers, outdated facts, broken citations, changed products, overlapping pages and weak internal-link support. Seasonality, reporting delays and market changes can also move data, so diagnose before editing.

1. Recheck intent. Inspect the current query set and actual search results. Decide whether the page still serves the same reader and whether another site page now overlaps.
2. Verify facts and dates. Replace stale platform, law, price, personnel or statistical claims with current primary sources. Remove unsupported claims instead of cosmetically rephrasing them.
3. Close content gaps. Add sections only when they answer a demonstrated user question or complete the process. Preserve useful depth rather than rewriting for novelty.
4. Repair structure and media. Improve headings, accessibility, examples, images and mobile presentation where evidence shows friction.
5. Refresh links. Add relevant newer resources, improve anchors and create new inbound links from appropriate pages.
6. Publish a genuine revision. Record what changed, update the modified date under policy and inspect the live page.

Changing the date alone is fake freshness. It does not make the article more correct or useful and corrupts the change record. If the page remains accurate and performs its job, leave it alone.

Measuring publishing, not just content

Build a publishing-health view before attributing traffic changes to a checklist. Track published URLs against eligible, indexed canonical URLs; posts with unexpected noindex or canonical destinations; successful sitemap discovery; broken links; image and mobile issues; and the share of priority posts with at least two contextual inbound internal links.

Use Search Console’s Performance report by page and query cluster. Google’s official metric definitions describe impressions as how often a link was seen under result-specific rules, clicks as navigation from Google to the site, click-through rate as clicks divided by impressions and position as a relative value. Average position is complex and should be interpreted as a trend with its filters, not as the exact rank a colleague sees today.

Compare page groups, query clusters, devices, countries and periods where sample size permits. An increase in impressions with flat clicks may indicate broader visibility, weaker snippet appeal, different queries or changed result features. A click decline may reflect demand or ranking, not a publishing error. Pair search data with index inspection and the change log.

The publishing report should show process measures too: posts released on schedule, checklist exceptions, retro-links completed, live-page defects, time to correction and refresh backlog. These are controllable. Rankings and revenue have more causes.

Fixing what is already published

Audit the archive in an order that protects retrievability before polish.

1. Indexing and canonical errors first. Identify valuable URLs excluded or consolidated unexpectedly, verify intent, correct the cause and inspect the live response. Do not bulk-index thin or private pages.
2. Thin, duplicate and obsolete-content decisions. Keep and improve, consolidate with a mapped redirect, intentionally noindex, or remove under the site’s policy. Decide one URL at a time according to user value and architecture.
3. Internal links. connect important orphaned or weakly linked pages, repair destinations and make anchors useful.
4. Refreshes. update candidates with a clear intent, evidence and realistic opportunity.

Fixing accidental exclusion can be a higher-value hour than drafting the next article because it restores access to work already paid for. It still does not guarantee indexing or ranking; Google ultimately chooses what it indexes and displays. The dedicated guide to fixing Google Search Console errors covers technical diagnosis beyond this publishing layer.

Building the workflow so it survives a busy month

Maintain one calendar showing draft approval, asset readiness, staging, review, publication, retro-link and refresh dates. Give one person authority over the checklist even when writers, designers, developers and subject experts own individual inputs. Unowned checks become optional under deadline pressure.

Use a staging step that closely represents production without being publicly indexable. Define who can approve copy, claims, SEO fields and technical changes. The definition of done should include a live-page inspection, working links, intended robots and canonical state, mobile check, measurement record and completed retro-links—not merely a CMS status of “published.”

An external publishing assistant can build entries, prepare assets, run checks, maintain the link map and report exceptions. The marketing lead supplies approved strategy, copy, claims, brand rules, priority and access to decision-makers. Developers own template, rendering, performance and integration defects outside the publisher’s permissions. The separate content publishing assistant guide explains that role and editorial-calendar interface.

Issue a monthly report with URLs published, exceptions, unresolved technical dependencies, inbound links added, indexed-status review, posts selected for refresh and owner decisions due. A process survives when blocked work is visible rather than silently skipped.