Voice of the Industry

Old fraud, new approach: how to mine the value of false positives

Tuesday 6 February 2018 10:22 CET | Editor: Melisande Mual | Voice of the industry

Stefan Nandzik from Signifyd: Due to Guaranteed Fraud Protection merchants can test the border between fraudulent and legitimate transactions.

This editorial was first published in our Web Fraud Prevention and Online Authentication Market Guide 2017/2018. The Guide is a complete overview of the fraud management, digital identity verification and authentication ecosystem provided by thought leaders in the industry from leading solution providers (both established and new players) to associations and experts.

The marketing world is awash with buzzwords – those light-as-air, sophisticated-sounding, but often hollow collections of syllables that serve as a glue for the message you would like your audience to embrace, without exactly examining it.

In the business of online fraud protection, the phrase “false positives” is rising to the top of the buzzword heap. It’s a real thing, but the term is vague, confusing and often misused. In fact, I prefer the term ‘insult rate’ to describe the percentage of legitimate orders falsely declined. After all, withholding orders from your legitimate customers is one of the easiest ways to lose a customer.

That said, it’s time to give false positives some love.

Ecommerce fraud protection has evolved rapidly, requiring a shift in thinking. The biggest innovation in fraud prevention has been the emergence of Guaranteed Fraud Protection – the notion that liability for chargebacks and fraud costs shifts from merchants to companies that specialise in machine learning-based fraud management.

Disruptive technologies and business models require disruptive thinking. In the case of Guaranteed Fraud Protection, the required reset is the realisation that the goal of fraud protection is not to prevent every incident of fraud.

In fact, in order to achieve true fraud protection, merchants or their fraud vendors must be willing to ship orders that they are convinced are fraudulent. They need to test the bounds of what is fraudulent and what is legitimate. They need to go up to the line representing fraud and then step over it.

Automated fraud systems learn from their mistakes

How else are smart machines going to learn all the permutations of fraud – or all the different looks a legitimate order can display? The sophisticated criminal rings engaged in online fraud are constantly changing tools, targets and tactics. The only way for machines to keep up is to experience the new twists firsthand Think of the tactic as the fraud-prevention equivalent to a vaccine to prevent disease. A vaccine contains the very antigen that causes the disease. It’s how the body learns what to combat.

In the case of fraud, shipping orders that a machine deems fraudulent is how an enterprise learns what fraud looks like. In fact, if your orders never result in a chargeback, your model is likely holding back too many orders.

As it is, retailers are not doing a very good job of finding the line between fraud and legitimate orders. Business Insider reported that incorrectly declined orders cost merchants USD 8.6 billion in 2016.

And the situation isn’t getting better. Digital Transactions reported that 35% of the orders rejected by large retailers turned out to be legitimate – up from 25% in 2016.

Now, purposely shipping apparently fraudulent orders is hardly a sustainable business plan for an individual merchant. First, there is the financial cost to the business. Second, there is the problem of explaining to bosses and investors why you appear to be throwing money away.

This is where Guaranteed Fraud Protection comes in. The promise, which has been recognised by consultants and others as state-of-the-art fraud protection, shifts financial liability away from merchants and onto partner companies that are experts in fraud protection.

The guarantee means that the vendor will financially cover the merchant for any approved order that turns out to be fraudulent. A fraud protection vendor is able to take on that risk because it sees millions of transactions across thousands of merchants.

Shipping fraudulent orders is R&D

It also means that the vendor can afford to – and in fact needs to – allow some orders through that appear to be fraudulent. The cost is akin to investing in research and development. The machine models learn from the fraud and from legitimate orders that were originally believed to be fraudulent. The data the vendor relies on becomes more valuable in the process.

In the end it’s all about data. Signifyd, for instance, invests significantly in third-party data from world-leading companies that provide high-quality information on devices, physical and cyber addresses, real estate records and more.

Therefore, leading fraud-management innovators are willing to invest to develop the data yielded by shipping orders initially identified as fraudulent. It is, in fact, data that money can’t buy. It’s the new paradigm that Guaranteed Fraud Protection enables.

Think about it: Everybody in ecommerce talks about making better fraud decisions and about reducing or eliminating false positives. But how? How do you do that, exactly?

Fraud feedback must include unexpected outcomes

Fraud protection systems that don’t understand the value of allowing apparently fraudulent orders to ship are stuck pushing their learning machines in a conservative direction. If every bit of feedback a machine ingests is the result of a chargeback due to fraud, that machine is going to ratchet back with every new data point.

If you also feed the machine some surprising results – feedback from orders that shouldn’t have been allowed through, according to current models, but that were in fact legitimate – then you’re achieving true learning. Your revised model will know to allow more orders to ship.

Think of it as the investment you need to make in order to discover the positive in false positives.

About Stefan Nandzik

Stefan Nandzik is the head of marketing at Signifyd. His “what if” approach to business problem-solving constantly challenges conventional wisdom and means that he is never afraid to upend the status quo to lead change.



About Signifyd

Signifyd, as the world’s largest provider of Guaranteed Fraud Protection, delivers a 100% financial guarantee against fraud and chargebacks on every approved order. This shifts fraud liability, allowing you to increase sales and focus on your core business. Signifyd customers include Fortune 1000 and Internet Retailer 500

Free Headlines in your E-mail

Every day we send out a free e-mail with the most important headlines of the last 24 hours.

Subscribe now

Keywords: fraud, false positives, Stefan Nandzik, Signifyd, chargeback, fraudulent transactions, legitimate transactions, expert opinion, voice of industry
Countries: World