Melt Accelerators
Catalogue Enrichment Sprint
Turn the detail buried in your product descriptions into data you can use.
For ecommerce teams
Materials, styles, shapes, sizes and features are often written into product descriptions or visible in images, but not stored as attributes. Without them, filters are patchy, collections are hard to build, site search misses things and your Shopping feed is thinner than it should be. We use AI and data engineering, checked by people, to extract those attributes at scale, validate them, and hand back clean, import-ready product data.
Who it's for
- Retailers with large or fast-changing catalogues where attributes have been filled in inconsistently.
- Teams planning new filters, collections or a PIM project who need better data first.
- Brands whose Shopping feed or site search is held back by missing product detail.
What you get
- An agreed attribute set, chosen using search demand and fitted to your taxonomy.
- Enriched, validated product data for the SKUs in scope, ready to import.
- A quality report showing how accurate the attributes are.
- A summary of the collection, filter and feed opportunities the new data opens up.
How it runs
We assess your product data, propose the attribute set, enrich the SKUs in scope and validate a human-checked sample before handover. Nothing is handed over until you've seen the checked sample and agreed the accuracy. You approve at two points: the attribute set and the quality sample.
Scope: Agreed up front by the number of SKUs and attributes in scope. Image-based extraction can be added where product text is thin.
What we need from you
A full product data export, any taxonomy or PIM documentation, and one named contact.
What's not included
- Importing the data or configuring filters in your platform, unless scoped separately.
- Building the new pages the data makes possible (that's the Collection Expansion Sprint).
- Rewriting product descriptions.
Proof
- The Cotswold Company: more than 10,000 products enriched with new attribute data, supporting 900 optimised product listing pages and contributing to an 86% increase in SEO visibility.
- MADE.com: a machine-learning enrichment process that added 236,000 product attribute tags and made 420 new filter pages possible.
The check-in and what happens next
Around 90 days after the changes go live, we review the results against the baseline and agree the next step from your data.
Build pages on the new attributes with a Collection Expansion Sprint, keep data current with the Catalogue Growth Programme, or pause while your team maintains it.
About the Catalogue Growth ProgrammeOther Melt Accelerators
Collection Expansion Sprint
New collection pages for searches you already have the products for.
- Length
- Typically 8 weeks
- Changes ship
- New collection pages in your CMS, or import-ready files
Organic Shopping Sprint
Get your products showing up for the searches that sell.
- Length
- 4 weeks
- Changes ship
- Your Merchant Centre feed, plus product page fixes
Category Growth Sprint
Make the category pages you already have earn more.
- Length
- 4 weeks
- Changes ship
- Updates to your existing category pages
AI Shopping Readiness Sprint
Make sure AI assistants can find, understand and recommend your products.
- Length
- 4 weeks
- Changes ship
- Feed attributes, structured data and product content
Migration Assurance
Change platform without losing the organic revenue you've built.
- Length
- Set by your launch date
- Changes ship
- Redirect maps, specifications and QA; your build team implements
Organic search is a long game. We just move faster.
Tell us the problem you want to solve, and we'll tell you honestly which model fits.
Discuss this sprint