A unique map of card fraud losses in Europe was released by the analytic software firm FICO on Thursday.

Anyone seriously interested in card fraud – in reducing it, that is – should have a good look at the map and the stats for the last ten years that go with it.

The data come from Euromonitor International, with additional information from the UK Cards Association, and can be reached at www.fico.com/europeanfraud

FICO sees the big story as the increase in Card Not Present (CNP) fraud in many countries as other card fraud types – counterfeiting, card stolen or lost, ‘card stolen or lost in post’ and ID theft have been contained.

However it’s worth noting that in several countries – France and Spain for example – counterfeiting and lost card fraud continue unchecked or are even increasing.

The data show that card fraud losses for 19 European countries hit approximately €1.8 billion in 2016, a new high. (Russia, Ukraine and Turkey are included in these figures).

The UK saw the highest losses at £618 million, a 9 percent rise over 2015, topping the previous peak in card fraud, set in 2008 after the introduction of chip and PIN.

Card not present (CNP) fraud has gone from 50 percent of gross fraud losses in 2008 to 70 percent in 2016. Ten countries saw an increase in fraud losses, while eight saw a decrease.

Countries that have done better than most at reducing card fraud include the Netherlands, where it is down by 70% since 2008, and Germany, where it has fallen 31% since 2010.

FICO comments that good investment and control by the banks made the card fraud reduction in the Netherlands possible. Bringing dynamic authentication for some banks meant that they had better experiences for their customers, it says.

Germany’s hike in losses prior to 2010 is attributed to fraud migrating from the UK post a clampdown by banks in 2008.

But the UK’s roaring e-commerce market continues to place consumers and retailers at high risk.

In 2015 the UK’s card fraud rise was the highest in Europe. In 2016 two countries saw higher rises — Poland (+10 percent) and Sweden (+18 percent). The UK’s rise from 2015 to 2016 was just half of that from 2014 to 2015.

Some of the figure seem to be anomalous – no figures for CNP fraud are shown for France for example, yet FICO reports that France had the highest basis points at 8.9 (ratio of fraud losses to sales) among the 19 European countries, compared to 7 basis points for the UK.

Together, FICO says, the UK and France account for 73 percent of the total losses among the 19 countries in 2016, followed by Germany, Spain, Russia, Italy and Sweden.

That would seem likely to include a hefty CNP total for France, where card sales are at about half the UK level.

Commenting on the data, Martin Warwick, senior consultant for fraud at FICO, said:

“The growth in online spending and CNP fraud brings new challenges for banks and retailers, as criminals thwarted by chip & PIN have moved to a less risky channel. Hiding amongst the growth in online purchases is great from a criminal point of view, but finding and stopping fraudulent transactions just gets tougher. Spotting the ‘needle in a haystack’ requires new behavioural analytics and artificial intelligence, combined with enhanced information from outside the traditional data contained within a purchase.”

FICO says it is working with banks to advance the use of machine learning and artificial intelligence to identify fraud faster. The key, Warwick says, is to spot anomalies without putting friction into the transaction.

“It’s no longer just about identifying patterns that are unusual for the customer — we’re also looking at anomalies at the mobile device, IP address and merchant level,” said Scott Zoldi, FICO chief analytics officer. “All of these have ‘behaviors’ just as individuals do, and we’re using our 25 years of experience in artificial intelligence to identify those.”

Mobile analytics is an important area here, said Zoldi, who developed or co-developed half of the company’s 70 patents in artificial intelligence and machine learning. “FICO has developed archetype analytics that taps into the rich source of mobile context such as advanced geolocation, allowing us to use that information in FICO Falcon Fraud manager to make real-time decisions during a transaction,” Zoldi said. “These analytics draw on our patented work with customer behaviour archetypes.”

Banks and card issuers are also beginning to step up their use of real-time customer communication. “Contacting consumers early using automated two-way SMS is a key solution to making sure the transactions are valid,” Warwick said. “If this is fully automated and tied into the fraud solution — as it is with FICO Customer Communication Services and the FICO Falcon Platform — then cases can be closed without human intervention and consumers can be allowed to continue to spend when and where they want.”