Farfetch is the leading global technology platform for the luxury fashion industry. The talk will show Farfetch’s experience in adopting a taxonomy in the fashion domain and how it was applied to improve semantic search methods in our e-commerce search. We’ll show the challenges we’ve had and how we’ve used the fashion taxonomy in our query expansion module to improve search results. In addition, analyzing search usage logs, we will show how we were able to suggest additional examples to the taxonomy development.
Improving search experience with a taxonomy in the fashion domain
Improving search experience with a
taxonomy in the fashion domain
George Cushen & Pedro Balage
Taxonomy Boot Camp 2019
Farfetch at a glance
Employees across 13 countries
Gross Merchandise Value
Brands available for consumers
Luxury sellers on the
AOV on Marketplace
> 2.9 Million*
Orders on Marketplace
Active Marketplace consumers
Size of personal luxury good
industry (Bain estimates)
*Correct for full year 2018 **As at Q1 2019
Marketplace language sites
Outfit available from
📖 Standardising with a unified semantic fashion taxonomy
🏷 Connecting these fashion entities with products in a graph via AI
🔍 Applying the taxonomy and graph to improve the search experience
We’re mapping fashion DNA to decode personal style
Data Science Data Science
Where Business Meets Fashion
A domain specific knowledge graph for fashion.
Business vs Fashion Entities
How many ways can you say “puffer jacket”?
What is a knowledge graph?
A knowledge graph can describe
● a collection of nodes (entities) representing business and fashion entities
● and with labeled relationships between the nodes
Language = “EN”
● each containing information (properties)
ProductID = 123
AI: A Multi-Modal Multi-Task Approach
Product Type: Dress
Building a fashion knowledge graph
Search Recommendations ...
Fashion Knowledge Graph
Associates fashion entities with business entities
AI Knowledge cleaning Entity resolution Schema mapping
Expert Knowledge Data-Driven Insights
What did we learn?
✅ Collaborate early with SMEs
✅ Incorporate data-driven insights
✅ Invest in defining a unified taxonomy/graph and
✅ Consider perspective: seller vs consumer
✅ Use and promote standards where possible
How to mix different set of products?
Query A Query B Query C Query D
Products Products Products
How to score different expansions?
By = λ1 + λ2 C+ λ3 D+ λ4A
● Each expansion gives a score for each product
● Combined score is a is a parameterized model
● Parameters may be trained using click logs and machine learning
● The threshold cut could also be learned