Out-of-Stock Experience, Turning a Dead-End into an Opportunity
User Problem
When users at Farfetch land on a Product Detail Page (PDP) for an out-of-stock item, they often abandon their session without browsing similar or related products. This behavior leads to missed conversion opportunities and disengaged users, as they leave without exploring the broader catalog. Given that around 15% of landing PDPs are for out-of-stock items, this was a significant gap in user experience and revenue potential.
Findings
High Drop-Off on Out-of-Stock Pages: Analysis revealed that a substantial portion of users who reached out-of-stock PDPs didn’t continue to engage with other products, leading to single-page visits.
Peer-to-Peer Insights: Through peer-to-peer research and role-playing exercises, simulating the experience as if interacting with a client in a luxury store, we gained a deeper understanding of customer expectations. We realized that by analyzing user behavior on the site, we could intelligently infer and recommend the next best item, mirroring the personalized service of a luxury sales associate.
Opportunities
Next-Best Recommendation Logic: Inspired by the luxury retail experience, we identified an opportunity to develop a recommendation algorithm, where we surface the several carrousels 1) Same item other colours 2) Same Brand plus Category (e.g. Burberry Mules) 3) Same Category plus Colour.
A/B/C Testing Opportunity: Given this page's wide flow of users, we could afford a wider test. Our approach was to test the second row against the thirds given we were confident that the same item yet a different color would be the best option to serve our users.
Impact
The rollout of smart recommendations on out-of-stock PDPs contributed to an increase in engagement and conversion:
Single Page Visits: Clear improvement, as users who encountered out-of-stock items were more likely to browse other suggested products.
Add-to-Bag Actions: Increase, Revenue Boost: GMV rose with a high confidence level.
Impact of Recommendation Variants: Variant C, which prioritized similar product recommendations over generic designer and category suggestions, drove a more impactful engagement rate:
Revenue Impact: GMV increased XXM during this test (can’t state the results :)) but they were good.
Notes consider:
We continued refining this project by updating the ‘Email me when it’s back in stock’ flow, allowing users to add out-of-stock items to their wishlist. This enabled Back in Stock on Wishlist notifications (see use case in Relevant Notifications).
Role
Lead Designer at Farfetch
Team
PM: Carlos Oliveira
Mapping the User Problem with Data Metrics to Kick-Start the Project
Assumptions and Solutions
From left to right: Mobile First
Design Close-up, Specs for Dev and Translations in Cyrillic
