Page 1
© 2014 Fair Isaac Corporation. Confidential. This presentation is provided for the recipient only and cannot be reproduced or shared without Fair Isaac Corporation’s express consent.
Collection Analytic Building Blocks
Mary Du PontPrincipal Consultant, Fair Isaac AdvisorsFICO
FICO™ Forum: Debt Manager™ 9 User GroupFairfax, VA | June 4–5, 1014
Page 2
Agenda
© 2014 Fair Isaac Corporation. Confidential.2
► The Building Blocks of Collection Analytics
► The Data and Tools Needed to Perform the Analysis
► Questions
Page 3
3 © 2014 Fair Isaac Corporation. Confidential.
Building Blocks of Collection Strategy
Use of RPC to
obtain promises
Most effective use of Calls to get RPC
Most effective use of Dialer and Collector hour to make calls
To increase promises and
reduce roll
To increase RPC
To increase calls
Page 4
4 © 2014 Fair Isaac Corporation. Confidential.
Dialer Effectiveness
Use of RPC to
obtain promises
Most effective use of Calls to get RPC
Most effective use of Dialer and Collector hour to make calls
Dialer Reports—Per dialer hour effectiveness- make the best use of the dialer
Call Handling Time—Per dialer hour effectiveness- Use the least time per call
RPC Rate by Time and Day—Call at the right time
Page 5
5 © 2014 Fair Isaac Corporation. Confidential.
Where to Call
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
RPC/SYSTEM HOUR
Bucket 1 | Mobile | Weekday Bucket 1 | Home | Weekday Bucket 1 | Work | Weekday
Bucket 1 | FreeDial | Weekday Bucket 1 | Mobile | Weekend Bucket 1 | Home | Weekend
Page 6
6 © 2014 Fair Isaac Corporation. Confidential.
Collector Effectiveness
► Focus on call handling time►Contact AND No Contact►Study by call disposition
►Set disposition level standards►Study by workgroup►Do not mix del levels
0.0
50.0
100.0
150.0
200.0
250.0
300.0
350.0
400.0
RPC Handling Time
man mel
Page 7
7 © 2014 Fair Isaac Corporation. Confidential.
Dialer/Collector Effectiveness
► Resource as denominator►System hour►Collector hour►Accounts
Dialer Statistics 1-30 days 31-60 days 61-90 days Total
resource= acct calls per acct
agent calls per acctdialer calls per acct
rpc per acctrpc no promise per acct
rpc promise per acctleft message dialer per acct
left message collector per acctleft message per acct
promise to pay per acctno answer per acct
resource= system hour calls per system hour
agent calls per system hour dailer calls per system hours
rpc per system hour promise to pay per system hour
no promise to pay per system hourdialer message left per system hour
collector message left per system hourno answer per system hour
resource= collector houragent calls per collector hour
rpc per collector hourrpc promise tp pay per collector hour
rpc no promise tp pay per collector hourcollector left message per collector hour
collector no message left per collector hour
Page 8
8 © 2014 Fair Isaac Corporation. Confidential.
Account Level Penetration Study—Know when to stop call and move to
different treatment
Monthly Penetration Rate Study—Cease calls when penetration no longer
increases RPC
RPC Prediction—Call accounts likely to be RPC
When to start calling
Call Effectiveness
Most effective use of Dialer and Collector hour to make calls
Use of RPC to
obtain promises
Most effective use of Calls to get RPC
Page 9
9 © 2014 Fair Isaac Corporation. Confidential.
► Strategy Questions►When to start calling
►collection entry strategy►When/where to call
►RPC rates►How often per day to call
►Penetration rates►When to stop calling and try
something else►When the odds of reaching the customer
on the next call falls below threshold
Managing Calling Resources
► Strategy Answer►Begin calls at 42 days delinquent►Best time to call is mobile
afternoons►2 attempts per day on weekdays►Stop predictive calling by day 93
when RPC rate drops below 2%
For Every Account!
Page 10
10 © 2014 Fair Isaac Corporation. Confidential.
When to Start Calling
► When the roll rate is above x
Day 1
Day 2
Day 3
Day 4
Day 5
Day 6
Day 7
Day 8
Day 9
Day 10
Day 11
Day 12
Day 13
Day 14
Day 15
Day 16
Day 17
Day 18
Day 19
Day 20
Day 21
Day 22
Day 23
Day 24
Day 25
Day 26
Day 27
Day 28
Day 29
Day 30
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Roll Rate by Risk Band
Page 11
11 © 2014 Fair Isaac Corporation. Confidential.
When and Where to Call
► RPC analysis to predict which accounts (not call patterns) have greater odds of RPC►Does risk band correlate to contact rate?►New characteristics uncovered that can be used in to further improve
segmentation►Days since last customer transaction or Worst cycle in the last 12 months►Count of months delinquent in last 12 months►Customer age—Under 25 population has 400 basis point diff from 25-35
Page 12
12 © 2014 Fair Isaac Corporation. Confidential.
0%
5%
10%
15%
20%
25%
30%
0
20000
40000
60000
80000
100000
120000
< 28 29-42 43-56 > 56
Acco
unt V
olum
e
Days since last customer activity
DistributionRPC Rate
When and Where to Call
Page 13
13 © 2014 Fair Isaac Corporation. Confidential.
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
0
11000
22000
33000
44000
55000
NoScore
201-250
251-300
301-350
351-400
401-450
451-500
501-550
551-600
>600
RPC
Rate
Acco
unt V
olum
e
Collections Score
Distribution
RPC Rate
When and Where to Call
Page 14
14 © 2014 Fair Isaac Corporation. Confidential.
Call Efficiency
► How often per day to call►RPC rates are highest in early days and early passes►Over 25% of all contact occurs on first pass on first day►Over 40% of all right party contact occurs on first pass on first day
RPC RATE Contact Distribution RPC Distribution
No. Day Pass 1 Pass 2 Pass 3 Pass 4 Pass 5 No. Day Pass 1 Pass 2 Pass 3 Pass 4 Pass 5 No. Day Pass 1 Pass 2 Pass 3 Pass 4 Pass 5Day1 41.3% 25.7% 15.5% 9.1% 6.0% Day1 28.2% 7.0% 2.4% 1.1% 0.6% Day1 47.5% 7.4% 1.5% 0.4% 0.1%Day2 32.4% 20.7% 11.6% 6.2% 5.5% Day2 14.0% 4.0% 1.4% 0.6% 0.3% Day2 18.5% 3.4% 0.7% 0.2% 0.1%Day3 25.0% 15.1% 9.2% 4.5% 3.1% Day3 7.8% 2.7% 1.1% 0.5% 0.2% Day3 7.9% 1.6% 0.4% 0.1% 0.0%Day4 18.9% 11.4% 6.1% 4.0% 2.8% Day4 4.6% 1.9% 0.8% 0.4% 0.2% Day4 3.6% 0.9% 0.2% 0.1% 0.0%Day5 15.0% 9.0% 4.6% 1.9% 3.8% Day5 3.0% 1.4% 0.6% 0.3% 0.2% Day5 1.8% 0.5% 0.1% 0.0% 0.0%Day6 11.0% 7.3% 3.0% 2.2% 0.8% Day6 2.0% 1.0% 0.5% 0.3% 0.2% Day6 0.9% 0.3% 0.1% 0.0% 0.0%Day7 9.7% 5.1% 2.7% 2.6% 0.5% Day7 1.4% 0.8% 0.4% 0.2% 0.1% Day7 0.5% 0.2% 0.0% 0.0% 0.0%Day8 7.1% 4.1% 3.6% 2.7% 2.1% Day8 1.0% 0.5% 0.3% 0.2% 0.1% Day8 0.3% 0.1% 0.0% 0.0% 0.0%Day9 5.0% 4.8% 2.2% 3.0% 2.3% Day9 0.7% 0.4% 0.2% 0.1% 0.1% Day9 0.1% 0.1% 0.0% 0.0% 0.0%Day10 2.8% 2.1% 1.4% 2.3% 0.3% Day10 1.6% 1.1% 0.7% 0.4% 0.2% Day10 0.2% 0.1% 0.0% 0.0% 0.0%
Page 15
15 © 2014 Fair Isaac Corporation. Confidential.
How Many Time to Call
► If there are 28 days to call, how many do you need to try before it is time to give up?
Sunday Monday Tuesday Wednesday Thursday Friday Saturday
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
✓✓✓✓
✓
Page 16
16 © 2014 Fair Isaac Corporation. Confidential.
Call Efficiency
► When to stop calling and try something else►Once 50% of week hours
have been called, there is 60% probability of “No Contact”
Page 17
17 © 2014 Fair Isaac Corporation. Confidential.
Call Efficiency
► Once 40% of week hours have been called, there is >60% probability of “No Contact”
Page 18
18 © 2014 Fair Isaac Corporation. Confidential.
RPC Effectiveness
RPC/Non RPC Roll Rate Analysis—Which RPC events change rolls and
by how much?
Most effective Use of Dialer and collector hour
Most effective Use of RPC
Most effective
use of RPC
Page 19
19 © 2014 Fair Isaac Corporation. Confidential.
Measuring Call Effectiveness
► She was going to pay anyway!
► This customer wont pay even if they come to the phone!
► This customer will pay if you can help him understand how/when
Page 20
20 © 2014 Fair Isaac Corporation. Confidential.
The Value of Contact
► Roll Rate discerns the value of RPC►Typically there are very high
and very low risk accounts for whom RPC has little value (change in behavior)
►Requires granular risk segmentation
Very High High Medium Low Very Low0%2%4%6%8%
10%12%14%16%18%20%
RPC versus No RPC LM versus No RPC LM versus RPC
90 Day Del RPC Roll Rate
LM Only Roll Rate
No Contact
Roll RateVery High 85% 87% 89%
High 80% 84% 91%Medium 70% 80% 88%
Low 50% 56% 62%Very Low 35% 36% 37%
Page 21
21 © 2014 Fair Isaac Corporation. Confidential.
► Scores!!►No analysis can be interpreted correctly without risk based segmentation
►Scores can be behavior scores, roll rate scores, or payment projection scores
► Data which captures collection and dialer activity►DM9 data►Roll event►Cycle based data is best especially in high volume bucket one
What Tools are Needed?
Page 22
22 © 2014 Fair Isaac Corporation. Confidential.
Summary
► Ask the right (simple) questions► One question at a time► In the right order► Use the data you have
Page 23
© 2014 Fair Isaac Corporation. Confidential. This presentation is provided for the recipient only and cannot be reproduced or shared without Fair Isaac Corporation’s express consent.
Thank YouMary Du [email protected]
FICO™ Forum: Debt Manager™ 9 User GroupFairfax, VA | June 4–5, 1014