“A significant proportion of the ‘big players’ in Fraud Prevention are now offering machine-learning based tooling to their clients”
The majority of eCommerce organisations recognise that fraud is a major issue, directly hitting revenue and impacting customer perceptions. The way Businesses deal with the threat of eCommerce fraud differs depending on their customer demographics and transaction volumes, but there are some general themes across the eCommerce that can ensure adequate preparation against fraud risks:
- Improve Fraud-Screening Efficiency. Companies need to make sure only the most necessary of transactions are referred to manual review by automating as much of the fraud checking process as possible.
- Identify, Analyse and Manage Fraud by Channel. Organisations need to understand the different risks, behaviours and patterns associated with Mobile Commerce fraud in comparison with “traditional” eCommerce fraud.
- Train Fraud-Prevention Teams on the Latest Trends in eCommerce Fraud. A team who know what they are looking for, and who understand the current landscape, are always going to be more highly-skilled and adept at filtering out genuine transactions from fraudulent ones.
With that in mind, here are 3 key trends in Fraud Prevention for 2017 and beyond!
Social Media Fraud Review
More and more organisations are looking to social media for assurance of genuine transactions. As part of a 2016 Worldpay customer survey, 60% of organisations stated they are already using social media data in their manual fraud review processes. Customers with open and accessible Facebook, Instagram and Twitter accounts that corroborate their name and location build up an impression of an honest, genuine transaction. A solid social media presence is very much the ‘icing on the cake’ when deciding that a transaction should pass manual fraud review.
Manual checks on social media will continue to have a huge benefit, but cross-checking several social media platforms is time-consuming, and by no means foolproof. Many organisations will look for a more automated way to validate customers’ social media presence. The use of social sign-on for customer login is viewed as a simple way to increase customer data accuracy, and take advantage of a more authentic pool of customer data. In a 2016 survey by Rippleshot, only 35% of survey respondents offered social sign-on as part of their Customer account management process, but 56% acknowledged that they would place greater trust in customers who use social logins.
As more and more sales come via Mobile Commerce, companies are starting to take note of the need to analyse and treat mobile transactions differently. As mComm revenue increases, optimal fraud management on this channel becomes more important: customers browse and purchase differently on mobile devices, and likewise fraudsters have different techniques to exploit both businesses and customers on the mobile channel. According to the Worldpay survey, 53% of organisations track mobile transactions, but only 33% treating these transactions differently, despite almost two thirds of companies believing that Mobile Commerce comes with increased risk of fraud. To address this, companies are looking to:
- Track transactions that are placed on a mobile device, and separate them from ‘regular’ eCommerce transactions.
- Capture and analyse specific device IDs and mobile operating systems, in order to look for device-specific weaknesses and spot further patterns.
- Create and tune a specific set of mobile-only fraud rules that capitalise on the differences in customer behaviour, and protect businesses from Mobile Commerce fraud trends.
A significant proportion of the ‘big players’ in Fraud Prevention are now offering machine-learning based tooling to their clients, in the hope that it will reduce the amount of transactions that require actual human intervention. Manual review processes are one of the largest fraud management costs, and can be a threat to revenue and security if not done well. According to the CyberSource 2016 Fraud Report, 22.5% of eCommerce orders are still manually reviewed for fraud, with a small percentage of customers still choosing to manually review every single transaction for fraud risk.
By identifying important patterns from the past, machine-learning tools can help to make accurate predictions about the future, and spot any pattern anomalies. This makes machine-learning a great approach for preventing fraud.
For machine learning models to be effective and accurate, organisations need to have gathered a significant amount of transactional data over time. As eCommerce businesses are collecting more and more data than ever before — this is an area where organisations could make huge savings and reduce manual overhead in the future.
What to Do Next:
As fraud-screening techniques and solutions become more advanced, so too do the techniques used to commit fraud. The nature of eCommerce fraud is continually adapting, and businesses who wish to successfully deal with fraud need show awareness to fraud trends, both in terms of execution and prevention. Organisations can make sure they are prepared by asking themselves the following questions:
- Is a disproportionate amount of your Fraud prevention budget spent on maintaining manual review team(s)?
- Are you making the best use of both customer and transactional data within your organisation?
- Does your fraud screening solution offer you the flexibility to assess transactions by channel and consider behavioural patterns in your customers?
Contact us now to find out how we could help find & implement the right Fraud Screening solution for your business.