insights »
Why Real-time Risk Decisions Require transaction AnalyticsReduce losses with sharper, faster risk prediction and action in customer management and collections
The double shock of recession-spurred delinquencies and new regulatory inroads on profit is
making creditors acutely aware of the need to make sharper risk distinctions among customers.
In both account management and collections, they need precise insights to guide more targeted,
timely actions.
Every bank with transacting accounts has the potential to achieve this higher level of risk precision.
The necessary deep insights into customer behavior can be drawn from credit card and debit card
account transaction data. It’s the most consistent and frequently updated source of customer data.
By using transaction analytics to analyze this data—and combining the results
with the FICO® Score and behavior scores—banks more cleanly separate
customers by risk level, especially in moderate risk bands and early lifecycle
accounts. Sharper risk separation and real-time or frequent scoring enables
banks to quickly and precisely target line changes and timing, and other
treatments, to reduce losses and improve profitability. Better insight into risk
also enables banks to focus collections resources in ways that reduce charge-offs.
This white paper:
Demonstrates how combining transaction scores with traditional behavior scores and FICO® •
Scores (credit bureau risk scores) increases accuracy of risk predictions
Discusses the advantages of deploying transaction analytics in real-time mode to accelerate •
awareness of developing risk and enable early intervention to mitigate losses
Presents results of three studies showing the increased predictive accuracy, segmentation •
granularity and monetary value gained by adding transaction scoring to current methods
Number 29 —January 2010
www.fico.com Make every decision countTM
FICO® Transaction Scores, generated with every transaction, can be used instantly by FICO® TRIAD® Customer Manager to make real-time decisions
www.fico.com page 2
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
“Transactions will become more important as demand for reduced latency in business intelligence and monitoring require more granular knowledge and timely tracking at the transaction level. Real-time intelligence will be needed in every FSI vertical and across every transaction type from retail deposits and savings to loans, investments, and payments.”
TowerGroup, July ‘09
While top banks have recently increased their use of transaction data, even industry leaders have
yet to realize its full potential for real-time detection of changing risk.
Transaction analytics detect changes in risk as they occur, and generate fresh scores with each
transaction. Yet most banks are processing the scores in batch mode, usually once a month. This
batch processing prevents promotion or intervention with clients at point-of-sale when early
signs of improving or deteriorating performance occur. In addition, many banks are using models
that capture only a portion of the insights that could be extracted from their transaction data.
The means to make sharper risk distinctions in real time
are available today. Banks can take advantage of better
modeling techniques as well as software that enables
deployed models to deliver accurate scores in the
milliseconds it takes for authorization of a credit card
purchase or an ATM withdrawal.
Transaction scoring is the key to making accurate real-
time risk decisions. As TowerGroup recently pointed out,
demand for reduced latency in business intelligence and
monitoring will require creditors to pull more insight from
their transactions.1
The combined use of traditional behavior scores and the FICO® Score to achieve a more accurate
assessment of customer risk has been standard industry practice for many years. The predictive
models that generate these scores analyze different data sources and, consequently, provide
different risk perspectives. Using both perspectives to make customer decisions lifts results above
what can be achieved using either score in isolation.
Transaction analytics lift performance even further because they model an additional rich source
of data and provide an additional risk perspective. Authorizations of credit and debit account
transactions contain abundant detail—the what, when and how much of customer spending—
from which patterns indicative of risk can be drawn.
Some of these patterns are evident in Figure 1, which shows percentage changes in credit card
authorization amounts from a six-month period in 2007–2008 to the same period a year later as
the recession set in.
Sharper Risk »Distinctions in Real Time
1 “Global IT Spending Forecast for Risk Management: A Growth Opportunity for Business and Technology,” TowerGroup, July 20, 2009, # V60:02BE
Increasing Decision »Accuracy and Speed
www.fico.com page 3
Why Real-Time Risk Decisions Require Transaction Analytics
insights » insights »
In this chart, we see changing spending patterns for all consumers (blue bars), perhaps as a result
of economic stress. But accounts that were seriously delinquent by the 2008–2009 period showed
much more pronounced changes (orange bars) during this period. These customers have changed
their spending as the economic cycle has changed, including higher levels of cash advances, and
reduced spending in travel and home improvement.
It’s the combination of spending behaviors that creates a pattern indicative of rising or falling risk,
and sometimes it’s the absence of activity that is most indicative. For example, as consumers become
more risky, their spending shifts away from categories such as home improvement and retail. Single-
event transactional triggers (e.g., the third cash advance in a month) miss these clues to changing risk.
Transaction analytics—models that capture the most indicative characteristics and relationships in
transaction data—detect such spending patterns. As a result, they’re able to separate accounts that
appear the same in cycle-end activity summaries but actually have different levels of risk.
The best models are very sophisticated and very specific. They analyze characteristics that go
beyond whether the consumer bought groceries or gasoline to, for example, where the consumer
bought groceries and what time of day or night the consumer bought gasoline. They capture
additional dimensions of time and space, such as velocity of cash advances and changes in elapsed time for various types of purchases.
–30%
–50%
–10%
10%
30%
70%
50%
90%
Figure 1: Revealing risk in customer spending patterns Percent change in credit card authorization amounts per accountOctober 2008–March 2009 compared to October 2007–March 2008
All AccountsBad Accounts
PERC
ENTA
GE
CHA
NG
E IN
AU
THO
RIZA
TIO
N A
MO
UN
TSPE
R A
CCO
UN
T
All Accounts 6.21 13.27 4.23 2.19 1.90 6.67 4.01 6.06 2.93 0.65 4.27 1.62 2.74 20.48
Bad Accounts 3.82 65.70 12.99 10.87 1.04 5.29 6.35 1.48 11.80 15.97 13.21 11.69 8.71 14.89
Books & Education Cash Electronics Collectables Restaurant Entertainment Food Petrol Hobby Home
Improvement Liquor Retail Travel Convenience Check
www.fico.com page 4
Why Real-Time Risk Decisions Require Transaction Analytics
insights » insights »
Because, as shown in Figure 2, transaction patterns are customer-specific, analytic models
must also be able to differentiate between patterns that are unusual for one customer but not for another. Making an ATM withdrawal at 11:30 pm may be a departure from one customer’s
normal spending pattern, and it may indicate a changing risk profile, while it is completely
ordinary for another customer, who works on a night shift.
The most important point to understand about today’s advanced transaction analytics is that
the kinds of risk patterns shown in Figures 1 and 2 are detected as they develop. Indeed, the
best models pick up patterns indicative of changing risk within a handful of transactions.
Because a fresh risk score can be generated with every transaction, it reflects spending patterns
drawn from behavior taking place a day before or even minutes before the current transaction.
Overall Risk Level
Date Time Merchant Degree Of Risk
May 01 4:00 pm
May 09 6:30 pm
May 09 7:15 pm
May 19 5:30 pm
May 27 11:30 am
May 31 12:00 pm
June 01 9:00 am
June 19 6:00 pm
June 19 7:00 pm
June 21 4:30 pm
July 03 5:00 pm
July 08 6:00 pm
July 27 3:00 pm
July 31 12:00 pm
Specialty Retail
Specialty Electronics
High-End Retail
Specialty Grocery
Discount Retail
Petrol
High-End Hotel
Children’s Toys
Specialty Retail
Specialty Retail
Specialty Retail
Petrol
High-End Retail
Fast Food
Date Time Merchant Degree Of Risk
May 01 4:00 pm
May 09 11:30 pm
May 09 7:15 pm
May 19 10:30 am
May 27 11:30 am
May 31 12:00 pm
June 01 9:00 pm
June 19 9:00 pm
June 19 7:00 pm
June 21 4:30 pm
July 03 5:00 am
July 08 6:00 pm
July 27 3:00 pm
July 31 12:00 pm
Specialty Retail
ATM / Cash
Liquor Store
Casino
Discount Retail
Petrol
Casino
Discount Retail
Specialty Retail
Fast Food
Petrol
ATM / Cash
High-End Retail
Fast Food
Overall Risk Level
Figure 2: Customer spending patterns may reveal risk
Customer #1 Spending Pattern Customer #2 Spending Pattern
www.fico.com page 5
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
Figure 3 shows the potential advantage of real-time transaction scoring. Here we see one
account’s transaction scores, generated with each new transaction, as well as the behavior score
at the same point in time. The falling transaction score shows a significant change in risk, which
would otherwise not be picked up until cycle end when a new behavior score is generated.
Real-time transaction scoring, however, gives the bank the opportunity to stop transactions as
early as 11/15 as shown below—23 days before the end of the cycle.
FICO® Transaction Scores make such point-of-sale scoring and action practical and reliable.
Our patented profiling technique enables vast quantities of historical transaction data to be
brought to bear on the current transaction without the need to store and process a lengthy,
CPU-intensive data feed. The profiling technique, which captures individual customer spending
patterns, is dynamic. Account profiles are updated with every transaction. (For more information,
see the sidebar “FICO® Transaction Scores platform” on page 13.)
New FICO research demonstrates the added value of transaction analytics for reducing
losses. In the first two studies, combining a transaction score with a behavior score and FICO®
Score increases risk differentiation, enabling finer, more accurate segmentation in credit line
management and collections. In the third study, a transaction score used with a behavior score
accelerates creditor awareness of accounts showing signs of financial distress and enables timely
interventions to contain losses.
Research on Combined »Score Performance
FICO
® TRA
NSA
CTI
ON
SCO
REBEH
AVIOR SCO
RE
550
500
450
400
300
350
620
640
660
680
720
700
250
Figure 3: Timely alerts to changing risk
Transaction ScoreBehavior Score
11/10 11/12 11/14 11/16 11/18 11/20 11/22 11/24 11/26 11/28 11/30 12/2 12/4 12/6 12/8
www.fico.com page 6
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
Sharper credit line decrease decisions
In the first data study, comprising of more than 10 million North American credit card accounts,
FICO examined the predictive lift from combined use of behavior scores, the FICO® Score and
FICO® Transaction Scores in credit line decrease decisions. The results demonstrate significant
improvement in risk discrimination and, in particular, markedly greater risk purity of end strategy
segments across current and mildly delinquent accounts.
For simplicity and brevity, the charts below show the impact of combined scoring on just one
slice of the accounts: those with behavior scores in the lowest-scoring 10% (decile 1). This high-risk
segment represents about 6% of the entire account population.
As shown in Figure 4, the aggregate bad rate for the accounts in this behavior score decile is 13.53%,
where “bad” is defined as accounts with a maximum delinquency of 3+ cycles, bankrupt or charged-
off in the subsequent six months.
Adding the FICO® Score refines the view of risk, enabling more granular segmentation. The five
resulting segments, based on 20% FICO® Score divisions (quintiles), have a bad rate ranging from 5%
to nearly 27%.
Figure 4: Increasing segmentation granularity for credit line decrease by adding scores
1 2 3 4 5 Total
1
2
3
4
5
Total
Behavior score decile 1 (lowest-scoring 10%)
FICO® Score quintiles (20% divisions)
35.90%
28.49%
24.76%
20.54%
14.52%
26.99%
26.99%
24.02%
17.03%
14.28%
11.15%
7.48%
15.52%
15.52%
19.02%
13.19%
10.83%
8.30%
5.24%
11.50%
11.50%
14.98%
9.85%
8.10%
6.13%
3.74%
8.29%
8.29%
10.80%
6.17%
5.37%
3.73%
2.33%
4.97%
4.97%
22.95%
15.94%
13.04%
9.44%
5.33%
13.53%
13.53%
FICO
® Tr
ansa
ctio
n Sc
ores
qu
inti
les
(20%
div
isio
ns)
13.53%
Behavior score + FICO® ScoreBad rates: 4.97% to 26.99%
Behavior score aloneBad rate: 13.53%
Behavior score + FICO® Score + transaction scoreBad rates: 2.33% to 35.90%
+
+
www.fico.com page 7
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
Adding the transaction score further refines the view of risk, increasing segmentation granularity
even more. We break apart each of the FICO® Score quintiles into transaction score quintiles. This tri-
score approach expands the number and precision of risk-separated segments, with bad rates now
ranging from 2.33% bad rate to 35.90%.
This fine-grained segmentation enables the portfolio manager to assign credit line decreases with
much greater precision. Figure 5 shows a strategy diagram with segments receiving line decreases in
orange. Accounts receiving no decrease are in green.
In this diagram, we can see that:
When the behavior score and FICO® Score are used together, accounts falling in the lowest 60% of •
FICO® Scores (deciles 1–3) receive the decrease. These are represented by the three orange boxes
in the FICO® SCORE row of the diagram.
When the transaction score is used with the behavior score and FICO® Score, the greater precision •
in risk separation enables the strategy to be fine-tuned. With this added precision:
About 12% of accounts from the decrease population no longer receive the line decrease. •
These “swap outs” are represented by the green boxes at the bottom of the second and third
columns from the left (under the orange FICO® SCORE boxes labeled 15.52% and 11.50%).
About 8% of accounts not previously targeted for line decreases now receive them. These •
“swap ins” are represented by the two orange boxes at the top of the fourth and fifth columns
(under the blue FICO® SCORE boxes labeled 8.29% and 4.97%).
CURRENT ACCOUNTS
BEHAVIOR SCOREDECILE 1
BAD RATE: 13.53%
FICO® SCORE26.99%
TXN SCORE35.90%
TXN SCORE28.49%
TXN SCORE24.76%
TXN SCORE20.54%
TXN SCORE14.52%
TXN SCORE24.02%
TXN SCORE17.03%
TXN SCORE14.28%
TXN SCORE7.48%
TXN SCORE11.15%
TXN SCORE19.02%
TXN SCORE13.19%
TXN SCORE10.83%
TXN SCORE8.30%
TXN SCORE5.24%
TXN SCORE14.98%
TXN SCORE10.80%
TXN SCORE6.17%
TXN SCORE5.37%
TXN SCORE3.73%
TXN SCORE2.33%
TXN SCORE9.85%
TXN SCORE8.10%
TXN SCORE6.13%
TXN SCORE3.74%
FICO® SCORE15.52%
FICO® SCORE 11.50%
FICO® SCORE8.29%
FICO® SCORE4.97%
Increasing segmentation by adding FICO® Transaction Scores (TXN SCORE) to behavior score and FICO® Score
Credit line decrease
No credit line decrease
Figure 5: Strategy for credit line decrease
www.fico.com page 8
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
Better segmentation of accounts for credit limit decreases and other loss-mitigation treatments can
have a significant financial impact, as shown in Figure 6.
Sharper collections treatment decisions
FICO also performed a study on the impact of transaction scoring on the segmentation of delinquent
accounts, using the same 10 million North American credit card accounts as the previous study.
For simplicity and brevity, the charts below show the impact of combined scoring on just one slice
of the accounts: one-cycle delinquent accounts in the middle-scoring 10% (decile 5) of behavior
scores. The study demonstrated particular benefit in moderate score ranges such as this—those
traditional “gray areas” of risk, where portfolio managers may have the least confidence in applying
treatment to accounts.
Figure 6: Estimating the benefit of sharper credit line decrease decisions
Portfolio size 1 million accounts
Lift scenarios (% of accounts swapped-in for line decrease)
Monthly prevented bad $
Annual prevented bad $
Annual benefit per total account
Average bad rate of swap-in population
Average decrease amount per account
Utilization of incremental credit line
Monthly accounts subject to credit line decreases
Conservative: 10% Average: 15% Optimistic: 20%
$39,000
$468,000
$0.47
80%
13%
$750
5,000
$58,500
$702,000
$0.70
$78,000
$936,000
$0.94
Current accounts 900,000 (90%)
Sharper segmentation enables you to prevent losses by avoiding balance build on riskier accounts. The chart below shows representative results in three estimated lift scenarios. The actual benefit you gain will depend on how effective your current segmentation scheme is, as well as your specific policies around credit limit decreases.
It is important to note that these estimates show only one side of the benefit to be gained from more accurate credit limit decrease decisions. In addition to helping with loss mitigation, sharper segmentation also improves profitability. By identifying less risky accounts that can be removed from credit line decrease populations, it prevents unnecessary constraints on account usage and avoids causing dissatisfaction that can lead to the attrition of valuable customers.
www.fico.com page 9
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
As shown in Figure 7, the aggregate bad rate for the accounts in this behavior score decile is 23%,
where “bad” is defined as accounts with a maximum delinquency of 3+ cycles, bankrupt or charged-
off in the subsequent six months.
Adding the FICO® Score refines the view of risk, enabling more granular segmentation. The five
resulting segments, based on 20% FICO® Score divisions (quintiles), have a bad rate ranging from
13% to nearly 41%.
Adding the transaction score further refines the view of risk, increasing segmentation granularity. We
break apart each of the FICO® Score quintiles into transaction score quintiles. This tri-score approach
expands the number and precision of risk-separated segments, with bad rates now ranging from 5%
bad rate to 51%.
Figure 7: Increasing segmentation granularity in one-cycle delinquent accounts
1 2 3 4 5 Total
1
2
3
4
5
Total
Behavior score decile 5 (mid-range 10%) score decile 1 (lowest-scoring 10%)
FICO® Score quintiles (20% divisions)
51.15%
41.62%
39.16%
31.66%
24.85%
40.50%
40.50%
36.96%
28.85%
25.01%
20.03%
13.29%
25.77%
25.77%
30.51%
24.06%
21.54%
16.04%
10.44%
20.81%
20.81%
27.32%
19.60%
16.69%
13.64%
9.03%
17.23%
17.23%
23.07%
16.38%
13.03%
10.23%
5.03%
13.14%
13.14%
35.22%
26.58%
23.32%
17.51%
10.79%
23.32%
23.32%
FICO
® Tr
ansa
ctio
n Sc
ores
qu
inti
les
(20%
div
isio
ns)
23.32%
Behavior score + FICO® ScoreBad rates: 13.14% to 40.50%
Behavior score aloneBad rate: 23.32%
Behavior score + FICO® Score + transaction scoreBad rates: 5.03% to 51.15%
+
+
www.fico.com page 10
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
This sharper segmentation enables collections managers to make better decisions about where
to focus resources. Figure 8 shows various tiers of collections effort, with the highest one-cycle
collection priority in red and the lowest priority in blue.
The diagram and Distribution of Accounts table below it show that:
When the behavior score and FICO® Score are used together, the increased separation results in •
20% of accounts receiving high-priority collections treatment, 40% receiving moderate-priority
treatment and another 40% receiving low-priority treatment.
When the behavior score, FICO® Score and transaction score are all used, the greater precision in •
risk separation enables the strategy to be fine-tuned. Now only 8% of accounts receive high-
priority treatment—enabling managers to concentrate their most experienced and skillful
collectors where the risk is greatest. Another 8% of accounts are identified for low-priority
treatment, which may consist of allowing them to self-cure, and thereby avoid spending money
to annoy valuable customers. In between, accounts are also better separated by risk, enabling
managers to more finely target, test and evaluate appropriate treatments.
ONE-CYCLE DELINQUENT ACCOUNTS
BEHAVIOR SCOREDECILE 5
BAD RATE: 23.32%
FICO® SCORE40.50%
TXN SCORE51.15%
TXN SCORE41.62%
TXN SCORE39.16%
TXN SCORE31.66%
TXN SCORE24.85%
TXN SCORE36.96%
TXN SCORE28.85%
TXN SCORE25.01%
TXN SCORE13.29%
TXN SCORE20.03%
TXN SCORE30.51%
TXN SCORE24.06%
TXN SCORE21.54%
TXN SCORE16.04%
TXN SCORE10.44%
TXN SCORE27.32%
TXN SCORE23.07%
TXN SCORE16.38%
TXN SCORE13.03%
TXN SCORE10.23%
TXN SCORE5.03%
TXN SCORE19.80%
TXN SCORE16.69%
TXN SCORE13.63%
TXN SCORE9.03%
FICO® SCORE25.77%
FICO® SCORE 20.81%
FICO® SCORE17.23%
FICO® SCORE13.14%
Behavior score + FICO® Score
Behavior score + FICO® Score + transaction score
Collection priority
20% 0% 40% 40% 0%
8%36%32%16%8%
DISTRIBUTION OF ACCOUNTS:
Increasing segmentation by adding FICO® Transaction Scores (TXN SCORE) to behavior score and FICO® Score
Priority 1: 40%+ Bad RatePriority 2: 30-39%+ Bad RatePriority 3: 20-29%+ Bad RatePriority 4: 10-19%+ Bad RatePriority 5: <10%+ Bad Rate
Figure 8: Strategy for prioritization of collections treatments
www.fico.com page 11
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
Better segmentation of accounts for collections treatment can have a significant financial impact, as
shown in Figure 9.
Faster intervention to prevent losses
In a third study with a top UK card issuer, FICO sought to determine if the addition of transaction
scoring to behavior scoring would accelerate awareness of rising risk in distressed accounts and, if
so, how much loss mitigation value it would provide.
The study was conducted on a random sample of 850,000 accounts. The overall portfolio bad rate
(3+ cycles delinquent, charge-off or bankrupt) was 2.6%, compared to 2.7% bad rate average for all
the UK for the same period.
For the purposes of clear measurement in this study, we created a simplified scenario: The study
assumes that score cutoffs are the only trigger for intervention used, and that only one type of
intervention—terminating account utilization—is taken.
Figure 9: Estimating the benefit of sharper collections decisions
Portfolio size 1 million accounts
Delinquency % Balances affected Amount
With transaction scoring (75 basis point improvement in 1-to-3-cycle roll rate)
Baseline performance
1 cycle
1-cycle to 3-cycle roll
3-cycle-to-charge-off
1-cycle to 3-cycle roll
3-cycle-to-charge-off
1 cycle
15% of 1-cycle balances
85% of 3-cycle balances
7.5% of receivables
$19,687,500
$16,734,375
$131,250,000
7.5% of receivables
14.89% of 1-cycle balances
85% of 3-cycle balances
$131,250,000
$131,250,000
$16,608,867
Total receivables $1.75 billion
Sharper segmentation enables better collection prioritization, which results in lower roll rates for early-stage delinquencies. The chart below shows representative results from transaction scoring for a standard portfolio of 1 million accounts. By reducing the 1-to-3-cycle roll rate from 15% to 14.89% (compare the rows highlighted in green), transaction scoring reduces annual charge-offs by $1.5 million. The actual benefit you gain will depend on how effective your current collections segmentation scheme is, as well as your specific policies around collections treatments.
Results summary:
Annual reduction in charge-off
Annual benefit per delinquent account
Monthy reduction in charge-off
$1,506,094
$1.51
$125,508
www.fico.com page 12
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
Behavior scores and transaction scores were independently analyzed to determine four cutoff points
each, equivalent to cumulative bad rates of 5%, 10%, 15% and 20%. The intervention is triggered
whenever an account reaches either the behavior score cutoff or the transaction score cutoff within
these bad-rate categories.
The transaction scores triggered account intervention an average of five days ahead of the behavior
scores. Although that difference seems insignificant, the opportunity for loss mitigation was not.
When the transaction score triggered account intervention first (27% of all accounts and 25% of
the bad accounts), the changing behavior patterns they detected were indicative of substantially
rising risk. In those cases, the transaction score identified preventable balance build averaging $516
per bad account it triggered. When behavior scores triggered account intervention first (21% of all
accounts and 18% of the bad accounts), the preventable balance build was much less, only $28
per bad account. (Note: Our benefit analysis does not include the remaining 52% of the population
where both scores triggered account intervention at the same time.)
BALA
NCE
DIF
FERE
NCE
AT
TRIG
GER
Figure 10: Early alerts to significant risk
Behavior score triggered �rstTransaction score triggered �rst
SCORE CUTOFFS
$26 $26 $46$12
$405 $419
$558
$681
$0
$100
$200
$300
$400
$500
$600
$700
$800
20% 15% 10% 5%
www.fico.com page 13
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
On average, as shown in Figure 11, using transaction scores to trigger intervention prevented an
average of $197 of balance build per bad account, or an annual benefit of $2.56 per account in the
portfolio. For a portfolio of 1 million accounts, that equates to $2.6 million in balance build savings
from more timely decisions using transaction scores.
Figure 11: Added value of timeliness in risk scoring
20% 15% 10% 5% Average
Balance build prevented per bad account (annual)
Benefit per account (for all accounts in portfolio)
Annual benefit for portfolio of 1,000,000 accounts
Bad rate equivalents for score cutoffs
$163
$2.12
$152
$1.97
$237
$3.08
$2,557,000
$235
$3.06
$197
$2.56
FICO® Transaction Scores platform
The FICO® Transaction Scores deployment platform provides proven software for operationalizing transaction models. It incorporates the
same proprietary dynamic profiling technology used in FICO™ Falcon® Fraud Manager, relied on by 17 of the world’s top 20 credit card
issuers to protect transactions in more than 1.8 billion active accounts. This robust platform, which enables transaction data processing in
both real-time and batch mode, has enabled clients to handle volumes as high as 1.5 billion transactions a day. It is supported by a dedicated
team of software engineers.
Banks using FICO® Transaction Scores benefit from production-tested transaction variable libraries developed and refined by FICO analytic
modelers over 12 years of working with clients in the credit industry. These exclusive transaction characteristics have powerful predictive/
detection value, having been hand-crafted specifically to solve credit risk problems for bankcard portfolios.
FICO® Falcon Fraud Manager 6.0 can produce fraud scores and FICO® Transaction Scores from one installation of the application, sharing the
same transaction data feeds, and saving implementation costs and time. The FICO Connected Decisions Architecture also enables banks to
calculate FICO® Transaction Scores in real-time, then instantly make them available to FICO® TRIAD® Customer Manager for making real-time
decisions.
Why Real-Time Risk Decisions Require Transaction Analytics
insights »
For more information US toll-free International email web +1 888 342 6336 +44 (0) 207 940 8718 [email protected] www.fico.com
FICO, Falcon, TRIAD and “Make every decision count” are trademarks or registered trademarks of Fair Isaac Corporation in the United States and in other countries. Other product and company names herein may be trademarks of their respective owners. © 2010 Fair Isaac Corporation. All rights reserved.2623WP 01/10 PDF
The Insights white paper series provides briefings on best practices, research findings and product innovations from FICO. To subscribe, go to www.fico.com/insights.
For banks with transacting accounts, data from transactions is the most consistent and frequently
updated source of information on their customers. Under today’s economic and regulatory
pressures, it is essential for uncovering the additional insights that enable more accurate risk
predictions and sharper decisions on how to treat customers, from account management through
collections.
Most banks, of course, are currently interested in transaction analytics primarily for improving
control of risk. As the economy recovers, however, the opportunities that can be gained through
fine-grained separation of customers by other behavioral dimensions will garner more attention.
Precise decisions—driven by real-time detection of changes in not only risk, but customer needs
and lifestyle—will be essential for extending the right credit at the right time to maximize profit in
the recovering economy.
Learn more: Visit • www.fico.com/cards to learn about FICO’s overall program for Reengineering Card
Profitability, including using transaction analytics to reduce credit losses
Watch a • Tech Talk video interview on real-time risk decisions using transaction analytics
See related white papers at • www.fico.com/insights
Contact u• s to discuss how you can use transaction analytics to reduce credit losses
Better Results on »Both Sides of the Balance Sheet