WordPress diagnostic guide

WordPress audit: from public evidence to the setting to check

By Antoine Le Louët · Estimated reading : 8 min · References

An unexpected noindex, an unchanged description or a slow banner: start with the HTML a visitor receives, then trace it to the relevant WordPress setting. This guide helps you choose a check and verify the correction.

Choose a service page, a post and, if relevant, a WooCommerce product. Record URLs, theme, relevant plugins, date and cache state. You or your developer carry out admin operations; Argos observes one public URL at a time, with no wp-admin access or plugin to install.

Analyse my WordPress page

Four findings, four WordPress procedures

A page intended for Google contains noindex

View the URL’s source as a signed-out visitor and find meta robots. Also inspect X-Robots-Tag in response headers. Keep the exact directive and final URL: an admin setting alone does not describe the response delivered.

Next check: In Settings → Reading, check “Discourage search engines from indexing this site”. Correct it if the exclusion is unintended; then review the page setting in your SEO plugin, theme and server configuration. Purge the relevant caches and reread the anonymous response. Keep intentional exclusions, including on staging. Removing noindex does not guarantee indexing.

Identify SEO signals in the public response

The saved description does not appear

Compare the meta description in the initial HTML with the saved value for that same page. Check the final URL, language and duplicate tags before concluding that the cache is stale.

Next check: Save the page and identify active layers: cache plugin, hosting and CDN. Purge the affected URL with your developer and reload as a signed-out visitor. If the mismatch remains, identify which theme or plugin generates the tag; avoid two competing SEO generators. Verify the text in a fresh response rather than the editor preview.

Follow the stale-description example below

The main banner appears late

Identify the LCP element in a browser trace before changing images. Record its request, dimensions and any deferred loading. A large image, late HTML and an element hidden by a script need different fixes.

Next check: Find the block or theme template rendering that image. In staging, test appropriate dimensions and, if this image is the LCP, discovery in the initial HTML without lazy loading. Compare repeated loads with the same conditions and cache state. Do not give every image this priority.

Diagnose LCP delay and compare before/after

A plugin or the theme seems involved

A script path containing wp-content/plugins is a lead, not proof of a conflict or a complete inventory. First describe a reproducible failure: page, action, message and expected result.

Next check: On a backed-up staging copy, compare the failure with a reference theme and reactivate plugins one at a time. Health Check troubleshooting mode changes only your session: a signed-out public audit cannot see that same state. For an Argos comparison, use publicly reachable staging with a controlled configuration, no private data and indexing excluded.

Understand the scope of automated analysis

Fictional example · no customer data

Before/after example: a description served from an old cache

Fictional scenario, no WordPress installation or customer measurement
URL: https://atelier.example/services/
Saved value: Bicycle repairs in Nantes, by appointment.
HTML received by the visitor:
<meta name="description" content="Welcome to our website.">

Suggested correction: In this scenario, the same URL serves an old version. After checking the language and tag generator, purge its page cache and CDN cache if active. This illustration does not establish caching as the cause on your site.

Expected response after correction — illustration
<meta name="description" content="Bicycle repairs in Nantes, by appointment.">

Verify afterwards: Make a new anonymous request to the final URL, compare the initial HTML and check that one description matches the visible service. Keep before/after responses and repeat after the cache warms. Google may choose a different search snippet.

Read the public example report and its evidence

Complete the audit inside WordPress

  1. Make corrections verifiable

    For each finding, share the URL, steps, evidence before the fix, changed component and expected result. Check HTML and the journey after publishing; a saved edit is not necessarily what visitors receive.

  2. Check media and forms

    Inspect alternative text in the block actually displayed: a media-library value may not be used by the theme. Test field names, focus and form confirmation by keyboard. Separately verify delivery of enquiries with test data.

    Repeat the journey’s accessibility checks

  3. Assign maintenance to people with appropriate access

    Verify updates and backup restoration in a suitable environment. For HTTPS and headers, identify what depends on the host, server or a plugin. A public analysis checks neither your backups nor every vulnerability.

Limits of this diagnosis

  • No exhaustive scan of plugins, versions or vulnerabilities; no database or wp-admin analysis. WordPress detection remains a probable indication.
  • One URL at one point in time. An unavailable module is not a passed check. Public evidence does not automatically identify the responsible theme, plugin or host.
  • Corrections need human review and appropriate access. No SEO ranking or business improvement is guaranteed.
Understand scores and partial results

Analyse my WordPress page

Run a technical audit of one public URL across four domains: performance, security, SEO and accessibility. The report and its progress remain accessible through a private link.

No account · One public page analysed

Official references