Improving Crawlability with Technical SEO: Fixing Duplicate Content via Custom Canonical Logic

Project Overview

This case study documents a technical SEO problem that emerged on my own portfolio and blog sites — me.dwiyanti.com and dwiyanti.com — as a direct result of improving the permalink and breadcrumb architecture. The changes were intended to make navigation cleaner and more SEO-friendly. Instead, they introduced a duplicate content problem that required a custom PHP solution to resolve.

The issue is a common one in WordPress SEO: when you restructure permalink hierarchy or introduce parent-child page relationships, WordPress can begin generating multiple accessible URLs for the same content. Google sees these as separate pages with duplicate or near-duplicate content — and penalizes the site's crawl efficiency and ranking signals accordingly.

What makes this case study relevant is not just the problem itself, but the method of resolution. Rather than installing a premium SEO plugin to manage canonicals, I wrote the canonical logic directly in functions.php — giving complete control over how every URL type on the site outputs its canonical tag.

SEO Problem Identified

The problem began after implementing two structural changes to the portfolio site:

Change 1 — Parent/child category and page structure

Posts were organized under both a parent category and a child category. For example, a finance article under the Lifestyle parent would generate a URL like /lifestyle/finance/article-slug. This is verbose, hard to memorize, and — critically — creates URL variations that WordPress can expose simultaneously.

Change 2 — Permalink shortening with breadcrumb preservation

To fix the long URL problem, the permalink was simplified to /article-slug only. But the breadcrumb was configured to still display the full path: Home › Lifestyle › Finance › Article. This is the correct user experience — but without proper canonical handling, WordPress continues to make the longer URL pattern accessible alongside the short one.

Technical SEO Analysis

Diagnosing this problem required checking three separate layers: how WordPress was generating URLs, how the canonical tag was being output, and how Google was interpreting both.

• 1
  Google Search Console — Coverage Report Identified multiple URLs being indexed or flagged as duplicate. GSC's "Duplicate without user-selected canonical" warning confirmed Google was finding URL variations and was uncertain which to treat as the primary page.
• 2
  URL Inspection — Canonical Tag Audit Used GSC URL Inspection on affected pages to check what canonical tag was being output. In several cases, WordPress's default canonical was outputting the wrong URL — defaulting to the parent/child path rather than the clean short permalink.
• 3
  Site Crawl — Accessible URL Mapping Crawled the site to identify all URL variations that returned a 200 status code. Any URL that returns 200 is a URL Google can index — even if it's unintentional. Several parent/child combinations were still accessible.
• 4
  Breadcrumb vs. Permalink Structure Analysis Mapped the mismatch between breadcrumb display paths and actual permalink patterns. Confirmed that the breadcrumb function was building paths from page parent relationships — which were also influencing WordPress's internal URL generation logic.
• 5
  Plugin Canonical Output Review Checked what canonical output existing free SEO plugins would generate for the affected URL types. Confirmed that free-tier plugin logic did not support the custom conditional canonical behavior needed — this required a code solution.

SEO Strategy Implemented

The strategy had two components that needed to work independently of each other: a clean URL structure for Google to index, and a full breadcrumb path for users to navigate. These two things do not need to match — and separating them is the key insight behind the fix.

Component 1 — Custom Canonical Logic

WordPress's default canonical output was removed using remove_action('wp_head', 'rel_canonical'), then replaced with a custom function that outputs the correct canonical based on post type and page context. For posts, the canonical always resolves to /post-slug. For service pages, the canonical resolves to /page-slug regardless of any parent page relationship.

// Remove WordPress default canonical
remove_action('wp_head', 'rel_canonical');

// Output custom canonical based on page type
add_action('wp_head', function() {
  if ( is_single() ) {
    $canonical = home_url( '/' . get_post_field('post_name') . '/' );
  } elseif ( is_page() ) {
    $slug = get_post_field('post_name', get_queried_object_id());
    $canonical = home_url( '/' . $slug . '/' );
  } else {
    $canonical = get_canonical_url();
  }
  echo '<link rel="canonical" href="' . esc_url($canonical) . '" />';
});

Component 2 — Custom Breadcrumb Function

A separate breadcrumb function was written that builds the display path from page relationships and category assignments — independently from the permalink structure. The breadcrumb shows Home › Services › SEO Specialist in the UI, but the canonical and the actual URL remain /seo-specialist.

Implementation Process

• 1
  Remove Default Canonical Hook Added remove_action('wp_head', 'rel_canonical') to functions.php to stop WordPress from outputting its default canonical tag on all page types.
• 2
  Write Conditional Canonical Function Built the custom canonical function covering four conditions: single posts, pages, category archives, and the homepage — each resolving to the correct clean URL.
• 3
  Write Custom Breadcrumb Function Built a breadcrumb function that reads post categories, page parent relationships, and hardcoded service page slugs to generate the correct display path without depending on the URL structure.
• 4
  Validate Canonical Output per Page Type Used GSC URL Inspection on posts, service pages, category pages, and the homepage to verify the correct canonical was being output in each case. Tested edge cases including child pages and posts with multiple category assignments.
• 5
  Request Re-indexing of Affected URLs Submitted all previously affected URLs through GSC URL Inspection's "Request Indexing" to prompt Google to re-crawl and re-evaluate the canonical signals on each page.
• 6
  Monitor Coverage Report Tracked the GSC Coverage report over subsequent weeks to confirm the "Duplicate without user-selected canonical" warnings were resolving and the correct URLs were consolidating ranking signals.

Results and Improvements

After the custom canonical logic was deployed and Google re-crawled the affected pages, the duplicate content signals cleared progressively over several weeks. The ranking impact was significant and measurable.

On Bing, the site ranks #1 for "dwi yanti", "dwi yanti seo", and "dwi yanti seo specialist". Bing Copilot now surfaces Dwi Yanti by name in AI-generated answers for relevant SEO specialist queries — a direct result of clean canonical structure combined with correct Schema markup and entity signals.

The breadcrumb display remains fully intact: users still see the complete navigation path in the UI and in SERP breadcrumb snippets, exactly as designed. The URL in the browser and in Google's index is the clean short version. Both goals were achieved simultaneously.

SEO Insights from This Case Study

URL structure and breadcrumb display are independent problems

A common misconception in WordPress SEO is that the breadcrumb must reflect the URL structure. It does not. Breadcrumbs are a UX and Schema signal — they help users navigate and help Google understand content hierarchy. The URL is a crawlability and canonicalization signal. These two systems can and should be configured separately for optimal results.

Duplicate content is most often an architecture problem, not a content problem

Most duplicate content issues in WordPress are not caused by copying text — they are caused by URL architecture decisions that inadvertently create multiple accessible paths to the same content. The fix is always at the PHP/architecture level, not the content level.

Custom canonical logic gives more precise control than any plugin

Free SEO plugins apply canonical rules globally based on post type. They cannot apply conditional logic based on specific slug patterns, parent page relationships, or custom breadcrumb configurations. Writing canonical logic directly in functions.php allows for exactly the behavior each page type requires — nothing more, nothing less. This is part of the technical SEO methodology I apply across all projects.

Crawlability improvements compound over time

Eliminating duplicate content does not produce an overnight ranking jump. What it produces is a steady consolidation of ranking signals onto the correct URLs — which compounds as Google re-crawls, re-evaluates, and re-ranks pages with clear canonical authority. The SERP dominance visible on this site is the result of that compounding effect.

← Back to all case studies