Protecting profits: Trainline’s journey toward secure growth

Trainline is the world's leading independent rail and coach travel platform. For the UK’s largest single rail ticket retailer, protecting the rail industry’s revenue while ensuring customers can access legitimate refunds simply and seamlessly is something Nick Aiken, Director of Fraud and Payments at Trainline, and his team, are razor-focused on.

“As an industry, rail is heavily subsidized by taxpayers, so every penny counts. A chargeback results in a 100% loss of the entire ticket value. To earn that money back, you’ve got to sell a significant amount of tickets. A bad day or a bad hour can have a big impact in terms of losses,” explains Nick.

The impact of fraud on Trainline goes beyond financial. Trainline has worked tirelessly to build strong, sustainable relationships with their train operator partners. Trust in Trainline is something the business takes very seriously.

That’s why the Fraud and Payments Team needs to be at the top of their game at all times. Nick regularly reviews his tech stack to see what Trainline could bring in to improve performance.

So, in 2019, Nick knew it was time to become machine-learning-first in their approach to tackling fraud.

“Machine learning offered us a far better way of fine-tuning our decisioning to make sure we continue to drive down fraud losses, while at the same time accepting as many genuine customers as possible to support the businesses growth,” said Nick.

Trainline chose Ravelin to support them in the strive for continuous improvement.

At the time Ravelin’s solution was integrated, Trainline was offering rail travel for customers in 173 countries around the globe which naturally led to Trainline’s block rate (% of orders declined due to suspected fraud) coming under the microscope.

“If a new customer lands on our site and is erroneously prevented from buying, and then never comes back... Well, this is the worst-case scenario for us,” said Nick.

By introducing and continually refining the bespoke machine learning model created by Ravelin, Trainline was able to reduce its block rate while maintaining a consistently low fraud rate. In new markets, getting that first customer interaction right could mean the difference between instant churn or repeat business – which would be bad news for Trainline and the wider rail industry.

Moving away from rules and manual review to largely automated decisioning through machine learning has freed up capacity for the Trainline Fraud team to focus on other impactful tasks. More on that later.

From 2021 to 2024, Trainline’s monthly transactions increased considerably, so to enable such scale, focus and operational efficiency are absolutely crucial for the Payment and Fraud team.

Analysis and investigation are at the heart of what a fraud team does every day, whether it's comparing behavior across customer segments, trend analysis or conducting deep-dive investigations into suspicious cases. How easy and quickly teams can complete these tasks not only defines productivity, accuracy and effectiveness, but also makes for an overall more enjoyable experience for fraud analysts and leaders.

“I’ve used tools in the past that don't align with what we needed. With Ravelin, it’s quite the opposite: You have a really good UI; you're always iterating to make it better, easier to navigate. Ravelin has the right information in your eye level at the right time. You don't have to go hunting for data.

From my perspective, the UI being clean and easy to navigate is a super strength of Ravelin,” says Nick.

When investigating suspicious behavior, data is everything. Trainline’s analysts are industry-leading in their methods, leveraging social media platforms and the dark web, among numerous other techniques, to impressive effect.

One tool that provides invaluable ammunition is link analysis through graph networks, which connect known fraudulent actors with other accounts created, as fraudsters often create new accounts, use different email addresses etc in order to mask their true identities and continually defraud a business.

“Ravelin’s Connect graph network is a very powerful tool. It helps us visualize suspicious activity beautifully and gives us a very quick view of what's gone wrong. Having the graph network in our toolkit is incredibly powerful in detecting fraudsters,” says Nick.

Over time, Ravelin and Trainline have worked together to go beyond card-not-present fraud prevention and explore additional ways to protect rail industry revenue.

Trainline obsess over the customer experience they provide. One factor in that equation is enabling customers to easily request a refund. But with this comes challenges: In the last couple of years, Trainline had observed an increase in incidents of refund abuse and with swift action saw a quick drop of refund rates as customers changed their behaviors.

“We know that if we have customers who are buying tickets, traveling and then fraudulently refunding, we have a duty to our train operator partners to be able to weed them out and try and deal with them in order to guarantee revenue for the rail industry,” said Nick.

Trainline have been pioneers in identifying and thwarting refund fraud, enabling enormous savings for rail companies; over £10/$12 million for Northern and a 39% reduction for Great Western Railway, to name a few:

Discover more insights about Nick and his team’s efforts to eradicate refund abuse by watching a recording of our joint webinar, How to manage refund abuse with Trainline.

Indeed, partnerships are instrumental in Trainline’s success – from strong relationships with their rail industry partners to working together with solution vendors such as Ravelin to enable secure growth and support Trainline’s vision and goals.

“Being able to pick up the phone to a partner is incredibly important. Ravelin have always been very open to the fact that if something goes wrong, then it's all hands on deck to get things fixed. I measure people and partners in terms of what they do when the chips are down and there's a problem. Ravelin has always reacted brilliantly and that’s why I continue to put my trust in them,” says Nick.

Since working with Ravelin, Trainline has seen excellent year on year performance improvements.

They are accepting more transactions safely and with confidence, which means more customer orders: Trainline’s block rate is 75% down in the UK and 64% down in European markets, while maintaining impressively low fraud rates.

Thanks to Ravelin’s machine learning, automation has also freed up time and resources, with manual review rates down by 92% and the number of active fraud rules reduced by 30%.

Trainline’s success story showcases three pillars of Ravelin’s value proposition:

• Secure growth by allowing more transactions through, safely

• A focus on machine learning and automation

• Strong partnership and industry-leading support

If you’re also considering Ravelin’s solutions to fuel your company’s growth, we’d love to have a chat about it. Reach out to our team of experts today.

“Ravelin is the perfect partner for us. They work alongside us in line with our values and have the toolkit and people to help us be successful” – Nick Aiken, Director of Fraud and Payments at Trainline.

• Watch the RavCon presentation: How Trainline handles refund abuse: A delicate balance
• Webinar: How to manage refund abuse with Trainline
• Ravelin insights: Policy abuse
• Ravelin developer docs: Refund abuse integration
  
• Travel & Hospitality Merchant Report