The Digital Economy Demands A New Approach To … Digital Economy Demands A New Approach ... The...
Transcript of The Digital Economy Demands A New Approach To … Digital Economy Demands A New Approach ... The...
Andrew Naumann
Vice President – CyberSource Risk and BI Products
The Digital Economy Demands A New Approach To Fraud Management
CONFIDENTIAL
Distribution beyond attendee’s company is prohibited.
Fraud solution strategyFraud Management v1.0
Reactive response to an immediate
threat
Overriding objective: minimize direct
fraud costs
Establish a deterrentthat forces fraudsters
elsewhere
Evolution of fraud
1980 1990 2000 2010 2015
Fraudsters Individuals Teams Local crime rings Global crime ringsGlobal crime rings with decentralized organization
Target Consumers Small retailers Larger retailers Banks processors Payment industry
Leading fraud types
Lost/stolen intercepted
Domestic counterfeiting/ skimming
Identity theft,phishing, rudimentary data compromise
Cross-border data compromise, CNP fraud,3D-Secure fraud,ATM fraud, ID fraud
Cross-border data compromise, CNP fraud,ATM fraud, ID fraud, pharming, hacking
Type of accounts targeted
T&E cardsPremiumcredit cards
Mass marketcredit cards
All types: Credit cards,debit cards, prepaid cards
All types: Credit cards,debit cards, prepaid cards, banking accounts
Necessary resources
OpportunismRudimentary knowledge
Technicalknow how
Audacity, technicalexpertise, insider information,global connections
Audacity, technicalexpertise, insider information,global connections
Source: Payments Cards & Mobile "Card Fraud Report" 2015 http://bit.ly/1QYPK8d
Balance
Minimize operational costs
Balancing multiple merchant objectivesFraud Management v2.0
EfficiencyMaximize automated decisioning
Streamline review process
Maximize data and system integration
Accurate detectionReduce fraud rate
Minimize chargebacks
Positive customer experienceReduce false positives
Increase acceptance rates
Review orders faster
GoalProfit optimization
High
Low HighAccepted Fraud Rate
Increasing
• Valid order rejection
• Manual review cost
Adverse profit impact
Profit optimization
(least loss)
Fra
ud
lo
ss | L
ost
sale
s | C
ost
of
op
era
tio
ns
Ad
vers
e P
rofi
t Im
pact Increasing
• Chargebacks/fees
• Credits due to fraud
Channels
Traditional two-stage eCommerce fraud model
Place
order
Reject
Accept
Digital
download
Deliver
Book trip / room
Business rules
4-8 hours queue
Automated Screening
Manual Review
1 2
0.9% Fraud rate
Current performance – North America
Source: CyberSource 2017 North America Online Fraud Management Benchmark Report
25%Manual review rate
2.9%Reject rate
OptimizeBalance
Minimize operational costs
3 months to quantify impact
of new strategy
Merchant CNP fraud rate
Proactive merchant credits, are issued to offset any chargebacks, representthe majority of direct fraud costs
Fraud chargebacks represent approximately one-quarter of thetotal merchant “fraud” losses
Additional merchant fraud costs include: Chargeback fees and fraud operations
Source: CyberSource North America Fraud Report 2016
28% 25% 27% 35% 32%
72% 75% 73% 65% 68%
Overall <$5M $5–25M $25–$100M $100+M
Fraud chargebacks Credit issued by merchants
Share of fraud claims
Annual online revenue (in US$)
Order rejection
Source: CyberSource 2017 North America Online Fraud Benchmark Report
2.9% Reject rate (on average)
forever
of rejects are likely valid10%
A customer may be lost
0.9% 0.9%
2.9%
6.8%
Percent of orders rejected due to suspicion of fraud
eCommerce fraud loss rate Order reject rate
Cross-border fraud2017 CyberSource fraud report
Percent of orders that were fraudulent
Source: CyberSource 2017 North America Online Fraud Benchmark Report Domestic Cross-border
0.9%
2.9%
Manual review
of operating budget**
25%
of North American businesses conduct manual review*
79%
of orders are manually reviewed* (on average)
Source: * CyberSource 2017 North America Online Fraud Benchmark Report** CyberSource 2016 North America Online Fraud Benchmark Report
46–52%
of manually-reviewed orders are accepted*89%
Current performance – Latin America
Source: CyberSource 2016 Latin America Online Fraud Management Benchmark Report
1.4% Chargeback
rate
29%Manual review rate
8.0%Reject rate
OptimizeBalance
Minimize operational costs
3 months to quantify impact
of new strategy
The integrated commerce experience
Anywhere Anytime Know me everywhere
Integrated commerce is the convergence of commerce channels to deliver a seamless customer experience
Brand names and logos are the property of their respective owners and the above-mentioned reference does not imply product endorsement or affiliation with CyberSource
eCommerce is driving global retail sales growth
Source: eMarketer Worldwide Retail Ecommerce Sales: The eMarketer Forecast for 2016 (excludes travel and event tickets)
2016F–2020F: Worldwide eCommerce and Face-to-Face (F2F) growth (US$T)
26.327.7
2019F 2020F
23.7
4.022.01.9
2016F
2.923.42.3
2018F2017F
20.1 22.9
3.4
22.021.1
24.9
F2F
eComm
US$ trillions
+16%
+3%
CAGR+6%
By 2020…
Digital is driving global spending growth
$2.2T $1.8T$24T
Sources: 1. eMarketer; 2. Juniper
ANNUAL GROWTH 2015-2020
3.5% 16.9% 17.2%
Face-to-Face POS1 eComm (excl. mobile)1 Mobile2
Global fraud is growing
Lost / stolen
Counterfeit
Card not present
2011–2016: Visa global fraud (US$B)
Cash
Other
+3%
+12%
+8%
CAGR
+12%
+3%
20162015
0.75
2014
1.271.02
1.62
0.911.48
2011 20132012
1.2
2.5
1.61.4
0.7
2.9
0.30.3
6.8
1.9
3.3
1.3
4.1
0.30.3
1.8
0.7
0.30.3
3.6
7.3
5.3
0.2
0.2
0.6
0.6
4.84.3
0.6
2.2
0.2
0.30.2
0.2
6.10.7
US$ billions
+9%
Integrated commerce is driving the evolution of fraud management
Consumers expect seamless checkout
experience
Drives account-on-file payments. Fraudsters
shifting focus to fraudulent account
creation and account takeover
Pace of change and pace of fraud attack morphing
Drives need to design/assess the impact of strategy
changes dynamically
Consumer expectations for “immediate
outcomes”
Challenges viability of manual review
Example: buy online, pick up in store models
Multi-device /multi-channel engagement
Complicates accurate customer validation
and anomalous behavior modeling
Integrated commerce is driving the evolution of fraud management
Consumers expect seamless checkout
experience
Drives account-on-file payments. Fraudsters
shifting focus to fraudulent account
creation and account takeover
Pace of change and pace of fraud attack morphing
Drives need to design/assess the impact of strategy
changes dynamically
Consumer expectations for “immediate
outcomes”
Challenges viability of manual review
example: buy online, pick up in store models
Multi-device /multi-channel engagement
Complicates accurate customer validation
and anomalous behavior modeling
6 AM Noon 6 PM Midnight
10%
5%
0%
Channel variances
Sources: 2014 comScore
Percentage of daily internet consumption by device
Good news…we are adapting
eCommerce Mobile
Sources: 1. CyberSource 2013 Annual Fraud Report
20121
mCommerce fraud rates are more than
50% higher than eCommerce
0.9%1.4%
0.9% 1.1%
Good news…we are adapting
eCommerce Mobile
Sources: 2. CyberSource 2014 Annual Fraud Report
20132
mCommerce fraud rates are more than
22% higher than eCommerce
0.9% 0.9%
Good news…we are adapting
eCommerce Mobile
Sources: 3. CyberSource 2015 Annual Fraud Report
20143
mCommerce fraud rates are the
same as eCommerce
0.9% 0.8%
Good news…we are adapting
eCommerce Mobile
Sources: 4. CyberSource 2017 Annual Fraud Report
20164
mCommerce fraud rates are
lower than eCommerce
But still room for improvement
Q: What percent of your annual eCommerce revenue comes from mobile devices?
‘eCommerce’ defined as any channel through which a customer can place a non-store order. This may be through your website or a mobile device.
**Mobile channel includes online orders placed via mobile optimized website or a mobile app, and does not include mPOS.
33%40%
49% 52%61%
8%
28%
38%
52% 49%
2012 2013 2015 2016 2017
Support mobile channel Support mobile channel & track fraud
Mobile channel adoption & fraud loss tracking
Integrated commerce is driving the evolution of fraud management
Pace of change and pace of fraud attack morphing
Drives need to design/assess the impact of strategy
changes dynamically
Consumer expectations for “immediate
outcomes”
Challenges viability of manual review
example: buy online, pick up in store models
Multi-device /multi-channel engagement
Complicates accurate customer validation
and anomalous behavior modeling
Consumers expect seamless checkout
experience
Drives account-on-file payments. Fraudsters
shifting focus to fraudulent account
creation and account takeover
Source: ThreatMetrix Q4 2015 Cybercrime Report, and 2017 CyberSource Fraud Survey
100% YoY growth
39% of merchants havetools to monitor
Account takeover fraud
Loyalty fraud
Sources: 1. "Loyalty points fraud: A real risk for a virtual currency", Ryan Yuzon, Director of Consulting,RFi Group, April 30, 2015. http://bit.ly/1NcWCqL; 2. Michael Smith, Managing Partner, Ai Group, Inc., http://bit.ly/1M3QR2p;3. Deloitte "Loyalty data security” http://bit.ly/1YyaYu0
US$238BLiability on the books of airlines, hotels, and other loyalty program owners
of loyalty program managers experienced issues related to fraud
72%
of consumers cancel membership if fraud occurs
26%
Integrated commerce is driving the evolution of fraud management
Pace of change and pace of fraud attack morphing
Drives need to design/assess the impact of strategy
changes dynamically
Multi-device /multi-channel engagement
Complicates accurate customer validation
and anomalous behavior modeling
Consumers expect seamless checkout
experience
Drives account-on-file payments. Fraudsters
shifting focus to fraudulent account
creation and account takeover
Consumer expectations for “immediate
outcomes”
Challenges viability of manual review
example: buy online, pick up in store models
28
Desktopand
laptop
Mobile
4+ hours
30–60minutes
B O O K I N G T O P I C K U P
27
Brand names and logos are the property of their respective owners and the above-mentioned reference does not imply product endorsement or affiliation with CyberSource
Multi-device /multi-channel engagement
Complicates accurate customer validation
and anomalous behavior modeling
Integrated commerce is driving the evolution of fraud management
Consumers expect seamless checkout
experience
Drives account-on-file payments. Fraudsters
shifting focus to fraudulent account
creation and account takeover
Consumer expectations for “immediate
outcomes”
Challenges viability of manual review
example: buy online, pick up in store models
Pace of change and pace of fraud attack morphing
Drives need to design/assess the impact of strategy
changes dynamically
Channels
Traditional two-stage eCommerce fraud model
Place
order
Reject
Accept
Digital
download
Deliver
Book trip / room
Business rules
4-8 hours queue
Automated Screening
Manual Review
1 2
Fraud management model for integrated commerce
31
Channels
Place
Order PaymentMethods
Credit / Debit
Gift / Paid / Loyalty
ACH / Direct Debit
AccountProtection
AutomatedScreening
Authentication
1 2 3 4
Review
******
Reject
Accept
Real-time strategy control and testingAnalyticsRules
Fulfill
Fraud
Feedback Into
Strategy / Models
1 2 3 4
Real-time Strategy Control and TestingAnalyticsRules
Account Protection
Jane Q
fraudster1234
555-555-1234
January
Male
1 1965
******
Monitor
• Same device accessing lots of accounts?
• Presence of malware on device?
• Same email address, multiple creation attempts?
• Device associated with spam?
Automated Screening
Patented Real-Time Fusion Modeling
Neural networks
Ensemble machine learning Static + self-learning techniques
Regression analysis Decision trees Proprietary classification
systemsSingle Score 0–99
Authentication
Business Rules
Allow risky transactions to be challenged
More transaction completed
Buy Now
Enter PIN code:583-800 Authenticate
Passively authenticate most transactions (no friction)
Review
Find Anomalies Fast
See Why
Suspended
Decision Manager Replay
Rule idea
Test on past transactions
Actual
Accept Reject Review
With proposed rule
Review
Accept
Reject
Accept
Review
Reject
Fraud management model for integrated commerce
31
Channels
Place
Order PaymentMethods
Credit / Debit
Gift / Paid / Loyalty
ACH / Direct Debit
AccountProtection
AutomatedScreening
Authentication
1 2 3 4
Review
******
Reject
Accept
Real-time strategy control and testingAnalyticsRules
Fulfill
Fraud
Feedback Into
Strategy / Models
1 2 3 4
Real-time Strategy Control and TestingAnalyticsRules
Other fraud sessions
Streamline your Manual Review Processes
Wednesday at 3:30 PM
3-D Secure: Present and Future
Wednesday at 4:25 PM
Reducing false positives: how to keep your legitimate customers
Thursday at 9:55 AM
Questions