Pat Phelan, Trustev: "What would a really modern anti-fraud system look like?"

With what vision did you start Trustev in 2013?

Just before Trustev, I was founder of a telecoms company and noticed online payment fraud was a huge problem. I looked at the major software packages to prevent it and found them very outdated: unusable and based on “rules” - complicated decision trees that need huge maintenance and still miss a lot of fraud while blocking far too many genuine customers. So we asked ourselves: what would a really modern anti-fraud system look like? With behavioural analysis, all data analyzed in one logic, machine learning (ML), and feedback loop. That’s what we built with Trustev.

How many good ecommerce orders get rejected and what do you consider the true cost of declining good orders?

It varies hugely, but it is far more than you think! As a general rule, 2 to 5% of revenues are turned away by blocking real customers -- because it’s not just that one transaction that’s blocked: you also waste the marketing spend to acquire the customer, and the future spending of that customer. Very damaging.

Could you please elaborate on how the “decision matrix” technology developed by Trustev works?

What we do is analyze thousands of disparate datapoints before, during and after a transaction online. Before, we look at not just device fingerprint, which we run on every page of every visit, but also website behaviour -- clicks and scrolls, velocity etc. - for mismatches with ‘normal’ behaviour.

Then, during a transaction, we examine IP data, e-mail validation, phone records, history, physical address, bot and VM detection and other data. This gets synthesized into a ‘single view’ of that customer’s context. Then we apply hybrid machine learning to spot outlier patterns.

After a transaction, chargebacks are fed back into the system automatically, to create a closed loop, which gets constantly better over time. Our philosophy is that you need to understand the whole customer interaction, end-to-end, to make a good decision.

The system takes around three days to deploy: drop in a code snippet on the website, and hook into our API. This surprises many companies, who are used to multi-month deployments. Because Trustev is light, we can offer ‘try before you buy’ no-cost pilots to prove the business case, which is also unusual in this industry.

SMEs selling goods internationally are more vulnerable than big companies that can invest in a professional fraud prevention team. Can SME’s benefit from you technology as well?

Mostly we sell to large enterprises, where the potential upside is literally tens of millions of USD or EUR, but we have SME customers too. The platform is cloud-based, so smaller companies can implement Trustev fairly quickly. Those customers tend to switch on automatic decisioning - no manual reviews at all. It is a great way for startups to expand internationally without worrying about payments.

Being a serial tech entrepreneur, what would you consider the next space for innovation?

It is a very exciting time to be in tech. We have seen a lot of innovation in finance and ecommerce recently: upstart vendors like Jet.com bringing new approaches, and new technologies like Bitcoin starting to go mainstream. I think the healthcare industry will soon see a lot more innovation, as consumers take more control and data from gadgets and platforms is fed into the system.

About Pat Phelan

Pat Phelan is the co-founder and CEO of Trustev. One of Ireland’s best-known serial entrepreneurs, he is recognised internationally as one of the most connected men in the technology industry. Over the past 15 years he has built a number of successful businesses in the internet, mobile and telecoms sectors.

About Trustev

Trustev is a new approach to preventing fraud online. It analyzes thousands of datapoints in real time to determine if a visitor is genuine or not. By looking at all data in one logic, in context -- instead of multiple siloes -- it can pattern-match for signs of fraud across site behavior, physical and e-mail addresses, phone numbers, IP addresses, proxies, devices, and many more signals. And by using machine learning, it constantly learns and improves. The company was named the EU’s “Top Technology Startup” by the European Commission in 2014, and won SXSW for big data and retail technology in 2014 and 2015 respectively.