How many times have you had your bank card declined just as you are about to take out some cash after arriving in a different country? Or an important internet purchase blocked for no apparent reason?

I was recently trying to buy a large quantity of liquid for a party – granted it wasn’t water – and the bill was rather large. However, I am a loyalty card holder, buy online and in store, and my shopping habits are well collected by this large retailer. They know me; they write to my home address; they send me regular emails – every morning at 6am in fact! The problem must have been my credit card, so I rang my bank: 'Nothing wrong with your card, everything is fine,' they told me. So undeterred, I tried once more. I re-entered all the wine and beer, only to have my transaction blocked again.

I have learnt a lot over the past year at Featurespace about fraud systems, and guess what? My party order was blocked because a fraud system had invoked a rule that said I was probably making a fraudulent order. But I wasn't. I'm a genuine, loyal and regular customer… only somehow their systems didn't know that. That retailer lost a very valuable order.

Marketers tell their risk departments on a daily basis that they are costing the company lost income, while risk departments say they are saving their companies millions by blocking 'potential' fraud. As we approach the most important weeks for retailers in the year, telling the difference between genuine and fraudulent customers could make all the difference between hitting targets and taking market share or anxiously waiting to see which retailer will confess that this Christmas didn’t quite live up to trading expectations. Thankfully we don’t need Harry Hill’s ‘fight’ sketch to resolve the inter-departmental conflict. The solution comes in the form of a revolutionary, home-grown, British technology, created by brilliant brains at Cambridge University.

A sea change in mathematics has been taking place across the world, which in turn has led to a fundamentally new approach to understanding human interaction through data and then accurately predicting outcomes. One of the world leaders in this area is Professor Bill Fitzgerald, one of the creators of predictive analytics. For 30 years, he has been working in Cambridge University’s Engineering Department, to identify the exact moment of change in data. He isn’t working with an army of scientists, but with a machine. The next generation of this approach was to combine machine learning with Dave Excell’s research in converting human behaviour into data. Together you have a revolution in what predictive analytics is capable of, and it’s called Adaptive Behavioural Analytics.

Although most companies have come to realise that they are sitting on gold in their data, they are seeking to address this through manually rather than by using Dave and Bill’s automated approach. The manual method requires hiring an army of data scientists who are special, brilliantly educated, highly sought after, expensive to hire, and require a new breed of talent management. They are essential to any modern organisation; however, their work can be of much greater value if data analytics is automated. Instead of being tasked with writing models, this will leave data scientists free to focus on more exciting projects.

The beauty of the automated system is that it unites the marketing and fraud departments by enabling both to monitor individual customers, in real time, and predict what they will do next. The marketers are happy because they can learn how to delight each customer with the same attention to detail as the owner of a corner shop in a small village. And it’s precisely because we know each customer intimately that we are able to say that their purchase is out of character, and predict with a very high probability if someone else (a fraudster!) is using their account or card. There is no guess work here – we simply know.

So as we approach the most important trading weeks of the year, the tension rises within retail organisations who are trying to find ways to delight every customer and win the race to out-perform peers. By this time next year, your business can resolve internal tensions and have the ability to spot fraudsters whilst taking genuine customers’ orders, especially those necessary liquid ones. Cheers!

http://www.featurespace.co.uk/