Validate category structure and compute taxonomy health rating

[emirbelkahia]

Context

Many e-commerce clients use category trees that are suboptimal for search and navigation. Misaligned taxonomy can degrade query classification, facet usability, and merchandising performance.

Opportunity

MCP can provide high value by:

1. Analyzing a client’s category tree to assess whether it follows Algolia’s best practices.
2. Validating key structural elements: consistent depth, parent-child coherence, and product coverage balance.
3. Scoring the taxonomy with a clear, qualitative rating.
4. Proposing actionable suggestions: e.g. flattening deep branches, merging redundant categories, or adding missing hierarchy levels.

This enables clients to improve user navigation, search relevance, and filtering quality at scale.

Proposed behavior

The MCP node should:

• Analyze the category attribute used for hierarchy or faceting (e.g., categories, hierarchicalCategories).
• Validate the tree against best practices:
  • Reasonable depth (ideally 3–4 levels).
  • Avoids duplication or circular logic.
  • Balanced product distribution across nodes.
• Compute a taxonomy health rating using a qualitative scale:
  • ❌ Poor – Likely to hinder performance
  • ⚠️ Needs improvement – Usable, but suboptimal
  • ✅ Good – Follows most best practices
  • 🏆 Excellent – Fully aligned with best practices and ready for scaling
• Output:
  • Breakdown of detected issues.
  • Suggestions per node (e.g., merge candidates, empty branches).
  • Optional: preview of an improved tree.

Example prompt

Can you audit my category structure and tell me if it aligns with Algolia’s best practices? If needed, suggest improvements.

Expected output

{
  "taxonomyHealthRating": "Needs improvement",
  "issues": [
    {
      "type": "ImbalancedDepth",
      "details": "Some branches go 6 levels deep while others stop at 2."
    },
    {
      "type": "RedundantNode",
      "details": "'Sneakers' and 'Trainers' contain 90% overlapping products."
    },
    {
      "type": "EmptyCategory",
      "details": "'Home > Fitness > Cardio > Rowers' has no products."
    }
  ],
  "suggestions": [
    "🔁 Merge 'Sneakers' and 'Trainers' into a single node.",
    "📏 Flatten branches deeper than 4 levels.",
    "🧹 Remove or archive empty categories."
  ]
}

Value

• Helps clients align their category structure with Algolia’s technical and UX guidelines.
• Improves faceting, filtering, and Query Categorization adoption.
• Drives more consistent and scalable taxonomy quality across implementations.

Notes

• This qualitative score is more actionable than a raw numeric rating, and helps drive conversations with clients.
• Future versions could allow mapping the rating to maturity stages (e.g. Early > Optimizing > Mature) for strategic accounts.
• Optional: let MCP suggest or apply a cleaner hierarchy using Algolia’s Data Transformation feature.
• Future version could integrate SKOS or RDF for semantic validation.