At Farfetch, I led the team that built the Customer Experience Index to make the post-purchase journey measurable. We turned customer service conversations into evidence leaders could use to decide where to invest, then used machine learning to expand that capability across markets and languages.

AI-Powered Customer Insights at Farfetch


Transformed fragmented customer feedback into actionable intelligence, connecting machine-learning-powered sentiment analysis with experience quality and business performance.

The post-purchase blind spot

As Farfetch approached its IPO in 2018, the business could see how customers browsed and bought online. After checkout, much of the experience happened off-screen. Leaders had limited visibility into where the customer promise was breaking down, what was driving the problem, and which improvements warranted investment.

Customer Service had a view into those moments through the conversations advisors had every day. That knowledge was spread across individual interactions and teams, making it difficult to identify patterns or establish how widespread a problem was.

Building the evidence by hand

I led a team of data scientists and analysts to build the first version of a post-purchase journey measure. We mapped contact volumes to key points in the journey, then read hundreds of customer emails to understand the sentiment and anxiety behind those contacts.

Contact volume showed us where customers needed help. Reading the conversations helped us understand why they were reaching out and which problems were creating the most frustration. We could then investigate root causes and identify where a change in the journey could prevent the issue from recurring.

From early findings to investment

Executives used the early reports to guide investment in customer self-service and robotic process automation. The research helped distinguish the issues customers could resolve themselves from those that called for a change in how the service operated.

The Self-Service case study shows how that evidence informed one of those decisions, starting with online order cancellation.

Scaling the research with AI

We worked with Marketing and an external technology partner to apply machine learning to service emails and survey comments. The team could analyze thousands of interactions each week and extend the research beyond post-purchase into the broader customer lifecycle.

Working across 12 languages gave teams a way to compare what customers were experiencing in different markets. They could see which problems recurred across regions and which needed local investigation.

Connecting service performance to customer behavior

The Customer Experience Index brought together three kinds of evidence: operational performance, contact volumes, and customer sentiment. Teams could use it to see where service performance was slipping, how customers experienced it, and where to investigate the relationship with behaviors such as cancellation, repeat purchase, and retention.

That gave leaders a way to prioritize improvements and build business cases around specific journey problems. It also gave teams a way to track the experience over time as changes went into effect.


Scale & Adoption

12

12 Customer feedback languages monitored for global insights.

1000s

1000s of Customer interactions monitored for sentiment each week.

100s

100s of Business cases supported for customer-centric strategies & business initiatives.


Business Impact

15% reduced

15% reduction in cost to serve as a result of improved operational efficiency and mitigated negative customer friction.

0.6 pt increase

Meaningful improvement in the company’s Trustpilot score through targeted CX initiatives guided by the CXI metric.


In 2022, my direct report, Farfetch’s Senior Head of Voice of the Customer, was nominated for the AIconics AI Innovator of the Year award for his work with the team on the Customer Experience Index.

Industry Recognition

More Case Studies.

Previous
Previous

Scaling Self-service

Next
Next

Unlocking Loyalty with AI