Page 37 - GS171202
P. 37

Education





            etc. all have different fraud pro-
            files. Knowing the type of device   AI by itself is not enough – the key is the addition
            allows you to screen accordingly.
                                                    of the human element to calculate specific
            3. Reduce manual reviews. Man-
            ual review of orders should be a   tolerance and risk levels for each specific business
            last line of defense when fight-         to provide real intelligence for retailers.
            ing online fraud. The process is
            slow, resource intensive and not
            scalable for company growth. In
            addition, making your customer    Some businesses think that blocking transactions from one particular region,
            wait for  approvals can cost  you   or stopping sales via one specific channel that has higher fraud rates than the
            sales and brand reputation. Au-   rest, will allow them to sidestep fraud. But this thinking will hurt them in the
            tomating  the  transaction  review   long run. Merchants need to find a solution that gives them the confidence to
            process can eliminate costs and   sell more products to more people in more places and expand revenue oppor-
            speed up order acceptance while   tunities – mitigating fraud, but boosting overall sales.
            maintaining a high-quality cus-
            tomer experience.                 Don Bush joined Kount as Director of Marketing in October 2010 and became Vice President of
            4. Use order linking and perso-   Marketing in December 2012. Previously, he was Director of Marketing at CradlePoint, a leading
            nas. Capabilities like order link-  manufacturer of wireless routing solutions in the mobile broadband industry. Don has worked in
            ing and personas use hundreds of   several management roles within the technology segment for over 20 years with both hardware/
            variables to construct a definitive   software manufacturers and as a partner in two top technology marketing agencies. He has led
            link to online purchase behavior   products launches and marketing programs for dozens of companies around the world such as
            ‒ either directly or indirectly ‒ to   Citi, HP, IBM, Kodak, Motorola and Weyerhaeuser, and he co-authored the seminar series Common
            help reveal fraudulent activity.  Launch Disasters and How to Avoid Them. Contact Don at don.bush@kount.com or visit www.
                                              kount.com.
            5. Optimize rules. Every busi-
            ness is different, and there is no
            universal setting for rules within
            your fraud platform. Your fraud
            platform should be able to un-
            cover hidden insights and opti-
            mize performance  of the fraud
            platform itself. After reviewing
            specific business goals and ob-
            jectives, merchants should make
            sure their fraud platform con-
            ducts an analysis of currently
            active rules and their overall ef-
            fectiveness to best make recom-
            mendations, along with an as-
            sessment of future performance
            gains.
        While merchants work to prevent
        fraud and its side effects, they also
        must ensure their systems aren't stop-
        ping legitimate sales in the process.
        The goal of most online (and brick-
        and-mortar) businesses is to grow
        revenue, increase market share and
        maintain an expanding base of hap-
        py customers. Expanding into new
        markets, adding new products or re-
        ducing  manual  reviews  can  help  or
        hinder growth, depending on what
        fraud strategy is behind each goal.



                                                                                                                37
   32   33   34   35   36   37   38   39   40   41   42