MADE.com: Unlocking 420 New SEO Landing Pages Through Attribute Enrichment

The Challenge

MADE.com was an online furniture and accessories retailer with a large and fast-growing product catalogue.

The business had more than 30,000 products and over 100,000 assignable product attributes, but more than 50% of assigned product attributes were missing within the PIM. This meant that many relevant products were not appearing within category filters, even when MADE.com stocked products that matched the customer's search.

This created issues for both SEO and user experience. Filter pages with no products were de-indexed from search results, meaning valuable landing pages could not rank. At the same time, filters with no assigned products were hidden from users, reducing their ability to refine the catalogue and find the products they wanted.

A manual enrichment process was considered too expensive and labour-intensive to be practical at this scale.

The project required us to:

• Enrich a large ecommerce catalogue at scale
• Identify missing product attributes across more than 30,000 products
• Improve product assignment across category and filter pages
• Unlock valuable SEO landing pages that were unavailable due to missing attributes
• Create a process that reduced the need for expensive manual tagging
• Improve the quality and usefulness of product data across SEO, PPC and user experience

The Work

Melt built a platform that used machine learning and advanced automation to suggest product attribute assignments.

The process combined image recognition, text mining and human validation. Product images were analysed using a convolutional neural network to identify relevant visual attributes, while product copy and user review content were mined to find additional attribute signals.

These recommendations were then passed into a validator application, giving reviewers a simple dashboard where they could accept or reject each suggested attribute. Once approved, the attribute was added back into MADE.com's PIM, allowing the product to appear within relevant filters and search-led product listing pages.

Our work included:

• Building a machine-learning platform to suggest product attribute assignments
• Creating an attribute universe of potential product attributes by family, category and product type
• Using image AI to recommend attributes based on product photography
• Using text mining to identify attribute signals within product page text and user reviews
• Assigning confidence scores to suggested product attributes
• Building a validator application for human review and approval
• Creating a workflow where accepted attributes were added directly into the PIM
• Mapping enriched attributes to previously unavailable filter and search pages
• Unlocking new landing pages where MADE.com had relevant products but insufficient product data
• Measuring the number of new attribute options added to SKUs
• Measuring the keyword and search volume opportunity unlocked by attribute enrichment

The Outcome

The attribute enrichment process allowed MADE.com to unlock significant SEO opportunity from products it already stocked.

By improving the completeness and accuracy of product data, MADE.com was able to assign products to a much wider range of useful filter pages. This helped turn previously empty or unavailable pages into valid SEO landing pages, while also improving the customer's ability to refine and discover products on-site.

The growth highlights:

• 236,000 attribute tags added
• 420 new filter pages unlocked
• Additional 570,000 monthly search volume targeted
• 60% increase in organic traffic
• 624% increase in top 3 rankings
• £870k monthly SEO revenue uplift from enriched attributes
• Improved product data positively impacted SEO, PPC and user experience
• A scalable enrichment process created for a catalogue where manual tagging was not practical
• A validator workflow built to combine AI efficiency with human review

The project demonstrated the value of treating product attribute data as a major ecommerce growth lever. By enriching existing catalogue data, MADE.com was able to unlock search demand, improve product discovery and create a scalable framework for surfacing more of its product range across organic search.