WordPress metadata problems are rarely limited to empty descriptions. Agencies also encounter duplicate titles, stale social text, canonicals pointing to old URLs, unexpected noindex values and templates that create the same output across hundreds of pages.
A WordPress metadata audit for agencies should identify what affects important pages, separate technical risk from editorial improvement and verify the rendered result after correction.
Define the audit inventory
Collect the URLs the site intends to manage, including:
- pages and posts;
- products and product categories;
- relevant custom post types;
- public archives;
- multilingual versions;
- campaign and location pages;
- known redirects and removed URLs.
Compare the WordPress inventory, sitemap and crawlable public site. Differences may reveal drafts in the sitemap, orphaned pages, unexpected archives or canonical variants.
Capture the right fields
For each URL, record:
- HTTP status;
- indexation intention;
- HTML title;
- visible H1;
- meta description;
- canonical URL;
- robots meta;
- Open Graph title, description and image;
- Twitter/X metadata where used;
- schema type and main entity;
- language and
hreflangrelationships; - last meaningful content update;
- internal-link count or context where available.
The saved WordPress field is not enough. Themes, plugins and filters can change the public output.
Run the audit in six passes
1. Accessibility and status
Confirm intended pages return a successful status and removed URLs use the approved redirect or removal response.
2. Indexation
Compare noindex, canonical and sitemap choices. Flag contradictions, especially valuable pages excluded from search or operational pages advertised as indexable.
3. Titles and headings
Find missing, duplicate, generic, obsolete or misleading titles. Compare them with the visible H1 and page purpose.
4. Descriptions
Identify empty, duplicated or keyword-list descriptions. Prioritise pages where a concise summary would help the reader understand the result.
5. Social and structured metadata
Check that shared links use an appropriate title, description and representative image. Confirm schema describes the page accurately without unsupported claims.
6. Language and templates
Inspect translated metadata, canonicals and template-generated output. A single flawed template can affect an entire content type.
Prioritise by impact and risk
Use a simple order:
- Critical indexation and canonical errors.
- Important pages with missing or misleading titles.
- High-value pages with duplicate or weak descriptions.
- Broken social images and incorrect schema.
- Long-tail improvements and low-impact consistency work.
Do not spend the first week rewriting descriptions for pages that should be consolidated or excluded.
Review titles in context
Google may use the HTML title, visible heading, prominent content, Open Graph title, anchor text and other sources when generating a title link.
Check whether the page has one clear subject and whether those signals agree. A good title should be specific, current, readable and supported by the page.
Avoid solving duplication by adding a location or brand token to otherwise identical pages. If the content serves the same intent, the underlying duplication may need consolidation.
Review descriptions as summaries
Google primarily creates snippets from page content and may use the meta description when it describes the page more accurately.
Descriptions should:
- summarise the actual page;
- distinguish it from neighbouring pages;
- use verified information;
- avoid keyword lists;
- make sense outside the website;
- remain useful even if search engines choose another snippet.
For structured catalogues, programmatic descriptions can be appropriate when the source data is reliable and the output remains varied and human-readable.
Audit canonicals deliberately
A canonical is a preference signal for duplicate or similar URLs, not a general repair tool.
Flag:
- canonicals to redirected or non-existent URLs;
- cross-domain canonicals without a documented reason;
- translated pages canonicalised to the wrong language;
- pagination or filtered pages using inconsistent rules;
- sitemap URLs whose canonical points elsewhere;
- multiple canonical tags in the rendered head.
Google recommends consistency across canonical signals and self-referencing canonicals on canonical pages.
Treat robots settings as high risk
An accidental noindex can remove an important page from search. Conversely, indexing account, search or incomplete-language pages creates low-value URLs.
Require explicit approval for bulk robots changes. Remember that crawlers must be allowed to access a page to read a page-level noindex directive.
Correct findings through controlled batches
Group issues by cause:
- one page needs editorial work;
- a content template needs correction;
- a plugin configuration affects the site;
- a multilingual rule affects one language;
- old URLs need redirects;
- content should be consolidated.
Fix the cause before editing hundreds of individual records. Apply a bounded batch, verify public HTML and retain rollback evidence.
WpAgencyKit SEO includes metadata controls, content auditing, internal-link analysis, 404 monitoring and sitemap functionality. Use it to support the workflow, while keeping page purpose and approval as editorial decisions.
Report portfolio progress without hiding risk
For agencies managing several sites, report operational coverage separately from search performance.
Useful audit measures include:
- URLs inspected by content type;
- critical contradictions resolved;
- high-value pages manually reviewed;
- template defects corrected;
- findings awaiting client approval;
- pages intentionally consolidated or excluded;
- next scheduled review.
Do not report “100% metadata complete” when automated fields exist but have not been checked for accuracy. Completion should mean that the required fields and indexation decisions were reviewed to the agreed standard.
Final verification
For each corrected group:
- inspect representative public URLs;
- confirm one valid title, description and canonical;
- verify robots and sitemap alignment;
- check social metadata and image;
- validate the intended schema;
- review language-specific output;
- record the change and reviewer;
- monitor recrawling and performance over an appropriate period.
Turn the audit into maintenance
Run a full baseline when onboarding a site, then repeat smaller checks after launches, migrations, template changes and significant content work.
The value of an audit is not the number of rows exported. It is the number of important, evidence-backed decisions implemented without creating new contradictions.