
This playbook is for sites that already have content and are fixing structural SEO problems - not brand-new builds. The scenario below is hypothetical, used to illustrate the mechanics.
The Scenario: A Cape Town Retailer Whose Category Pages Live Four Clicks Deep
Imagine a Cape Town homeware retailer on WooCommerce. Its homepage links to 'Shop', which lists brand collections. From there, a shopper picks 'Kitchen', then 'Cookware', then a product - four clicks from the homepage. The site ranks for the brand name and little else. Paid search sends traffic to the product page, but the category pages that should rank for 'cast iron pots Cape Town' are orphaned in the internal link graph.
This is a site structure problem, not a content problem. The pages exist with copy and images, but Google's crawler discovers them through links, and the deeper a URL sits, the less internal link equity and crawling attention it receives.
Tip: Treat your URL hierarchy as a business decision, not a developer default. The depth of your money pages directly affects how much crawl budget they consume.
Why Site Structure Is the First SEO Decision - and the Last One Anyone Revisits
Site architecture is set during build, rarely revisited, and quietly compounds. As you add products, blog posts, and landing pages, the original hierarchy either scales or buckles. Because restructuring later means redirects, lost signals, and a nervous CMO, most teams avoid it. The result: category pages that should be revenue drivers sit behind three or four intermediate pages. Structure also determines which pages your paid team can use as landing pages and whether organic and paid funnels point at the same high-intent URLs.
Information Architecture: Blueprinting Hierarchy Before a Single Page Is Built
Information architecture (IA) is the blueprint: which categories exist, how they nest, and which pages are hubs versus leaves. The decision depends on catalogue size. For 1,000+ SKUs, a deep, narrow tree (category → subcategory → product) keeps each level focused; for under 200 SKUs, a flatter structure (category → product) avoids thin intermediate pages.
Choosing a topology
Two common topologies: hub-and-spoke (a central category hub links to spokes like guides and product pages) and flat (everything linked from a single index). Hub-and-spoke suits content-rich e-commerce because it concentrates link equity on the hub; flat suits small service sites where every page is a potential entry point.
URL Taxonomy for South African Sites: Why /category/ and /shop/ Keep Swapping Places
URL taxonomy is the folder convention that expresses hierarchy. South African sites often mix /shop/, /category/, and /collections/ because different plugins or themes impose their own defaults. Pick one convention and document it. /shop/kitchen/cookware/ is readable and expresses the path; /p/12345 is not. Avoid dates (/2024/05/) and database IDs in URLs for evergreen category pages.
| Convention | Example URL | Use when |
|---|---|---|
| Folder depth 2 | /shop/kitchen/ | Small catalogue, strong subcategories |
| Folder depth 3 | /shop/kitchen/cookware/ | Large catalogue, granular filters |
| Flat | /cast-iron-pots/ | Single-product service pages |
Orphan Pages: The Content Nobody Links To and Google Eventually Stops Visiting
An orphan page has no internal links pointing to it - a product removed from category listings, a blog post only linked from a sitemap, or a forgotten campaign landing page. Google can still find orphans via XML sitemaps or external links, but without internal links they receive little crawl attention and rarely rank.
Run an orphan audit: compare URLs in your XML sitemap against URLs discovered by crawling your site's internal links. Any page in the sitemap but not in the crawl is an orphan. For South African sites with thousands of SKUs, this often surfaces discontinued products that should be redirected or removed.
Faceted Navigation and Filters: Indexable, Canonicalised, or Blocked - Pick Deliberately
Faceted navigation (filters by colour, size, price) creates many URL variants, each a potential duplicate or thin page. Decide per filter type:
| Filter type | Treatment | Rationale |
|---|---|---|
| Colour/size (low search volume) | Canonical to parent category | Prevents index bloat; passes equity to parent |
| High-intent filter (e.g., 'cast iron') | Allow indexing with unique title/meta | Captures specific queries |
| Price range | Block via robots.txt or noindex | Infinite combinations, no search demand |
The trade-off is between capturing long-tail traffic and managing crawl budget. Test and monitor.
XML Sitemaps, robots.txt and Canonical Logic as One Coordinated Layer
These three signals must agree. A common contradiction: a page is in the XML sitemap (signaling 'index me') but has a canonical tag pointing elsewhere, or robots.txt blocks the crawler from reaching it. Google resolves conflicts, but you waste crawl budget and may not get the outcome you want.
# robots.txt example - block filter URLs with query strings
User-agent: *
Disallow: /*?filter=
# But allow the sitemap
Sitemap: https://www.example.co.za/sitemap.xmlRule of thumb: XML sitemaps should list only canonical, indexable URLs. Robots.txt controls crawling, not indexing - a blocked URL can still be indexed if linked externally. Canonical tags consolidate duplicate signals. Audit all three together.
How Structure Shapes Paid Media: Landing Page Depth, Quality Score and MER
Paid media buyers need stable, deep-linkable URLs for landing pages. If category pages are buried four clicks deep or use session-based URLs, you limit testable options. Landing page relevance also feeds Google Ads Quality Score - a clean, relevant category page can lower CPC and improve ad rank. Structure affects CPA and MER, not just SEO. For aligning paid and organic structures, see our CRO service.




