Beyond HumanBig PictureCatalystsConnected WorldExchangeMarketing MixNew MoneyNew SchoolPeople SciencePulse

A response to the looming threat of online fraud

smart fraud detection startup smart fraud detection startup
Photo credit:

Perspecsys Photos

As the on-demand economy scales, so does the need for businesses to make smart, accurate fraud decisions, as Alex van Someren from Amadeus Capital found out.

The threat of cyber-security is an ever growing, ever topical problem within the tech sector.

As a direct result of business and consumer privacy concerns, many startups are entering the space to remedy the trust that this vertical is losing.

Ravelin is a smart fraud detection startup and prevention platform for businesses that need to make accurate fraud decisions in real-time.

It utilizes data science, anomaly detection and machine learning to identify bad actors more effectively online and has just secured £1.3m funding led by Amadeus Capital with Playfair Capital, business angels, including Errol Damelin and Paul Forster, and existing investors Passion Capital.

The Barclays Techstars alumnus operates within the on-demand economy by processing information instantly, allowing payments to be accepted with more confidence by vendors.

Amadeus Capital Partners is a technology investor working from London. Since its inception in 1997 the firm has raised over $1 billion for investment and backed more than 100 companies in software, mobile, internet, medical and cyber-security technology sectors.

Alex van Someren, Managing Partner of Amadeus Capital Partners’ early stage funds interviews Martin Sweeney, CEO of Ravelin on the rising threat of online payment fraud and how smart fraud detection startup is changing.



Alex van Someren: What motivated you and the Ravelin team to establish the smart fraud detection startup?

Martin Sweeney: Payment fraud is one of the biggest and often unanticipated risks for on-demand companies – businesses like transport apps, digital marketplaces and rapid-delivery retail.

As the VC money has flooded in to find the new Uber, so have the fraudsters who have a bunch of new businesses to target.

The cost of getting stolen credit card numbers is negligible, the risk of arrest is almost non-existent and, if it works, the rewards are enormous. The giants like Amazon and eBay know this and have teams to protect their businesses.

For newcomers, however, even very large ones, they need a different type of approach to keeping their businesses protected and to do so without the resources of a giant, that’s one of the benefits of a smart fraud detection startup.


AvS: One of things I often hear though is that ‘if you don’t want fraud – don’t accept any payments’. You worked at Hailo so you must have direct experience of this tension between adding customers and protecting against fraud. How do you intend to manage that balance for Ravelin users?

MS: That tension is very real and different companies at varying stages of their development have varying appetites for risk.

Our intention is to allow our users to operate at the level of tolerance that they decide is right for them. We primarily use machine learning models to score customers and orders on the probability of their being fraudulent.

Where those thresholds are set is up to our users. Over time, smart fraud detection startup models become more and more accurate, so the user will be able to accept more payments with less risk, which ends up keeping costs down for customers.


AvS: We’ve been focused on investing in machine learning for several years at Amadeus. My feeling is that the key is to analyse the data using the right tools, rather than having a fixed approach to how smart fraud detections startups deal with fraud; do you agree?

MS: I think so and we don’t intend to use a fixed approach. Take rule-based systems, for instance – which is how 90% of the industry today manages fraud.

Rules are simple to explain, reasonably transparent and can be quick to change. But they’re a terrible way to interrogate data, really hard to deploy and to manage at scale, and don’t take advantage of having access to aggregate trends across many customers.

However, we still intend to give our users access to a limited set of rules as, in certain scenarios, they’re a great way to prevent fraud. Our goal is to do whatever it takes to help businesses trade online safely.


AvS: It’s a fast-growing, global market opportunity, which inevitably means there’s plenty of competition. How does Ravelin intend to be different?

MS: As a smart fraud detection startup, our product architecture and technical approach have been built for the next generation of challenges: we’re not having to carry forward a product designed for a different era.

Our team has been in the trenches recently, dealing with the real issues, so we think we understand and are delivering what’s needed in the years to come.