Using incremental marketing
techniques to maximise
loyalty with minimum cost
Neil Skilling, Portrait [email protected]
Loyalty World 15th November 2010
A Typical Customer Lifecycle
New
BuildWin-Back
page 2
Cross & Up-
SellAnti-
Churn
Save
Typical Retention Strategies
●Pre-Emptive
●Build customer value into the relationship
●Ensure full-value understood, used and appreciated by customers
● Improve customer experience
●Proactive
●Target customers at risk ●Target customers at risk
●Target valuable customers at risk
●Reactive
●Product mismatch
●Save team and offer limits
●Win-back
page 3
Sophis
tication
Land Grab
Target on
value at risk
Target on
attrition risk
Significant Event
Increasing Sophistication in Proactive Retention
Time
Land Grab
Attrition Probability × Value
MAINSTREAM
CURRENT
PRACTICE
Attrition Probability
LAGGARDS
attrition risk
Untargeted
retention
Acquisition
Page 4
Value at Risk – Micro-Segmentation
Top5 Next 10 Next 20 Mid 30 Low40
High 1
High 2
High 3
Medium 1
Pro
pe
nsi
ty t
o l
ea
ve
Customer Profitability Segments= low / no investments
= medium investments
= high investments
Page 5
Medium 2
Medium 3
Medium 4
Low 1
Low 2
Low 3
Pro
pe
nsi
ty t
o l
ea
ve
SegmentNumber of
CustomersSaves
Save
Rate %
Value of
ResponseTotal Profit
Treated 100,000 75,000 75% 15,000,000£ 7,450,000£
Control 100,000 70,000 70% 14,000,000£ 13,880,000£
Uplift 5,000 5% 1,000,000£ 6,430,000-£
Measurement of Effectiveness
• Uplift = Treated “Response” – Control “Response”
Value = £200
Contact Cost = £0.50
Incentive = £100
• Uplift = Treated “Response” – Control “Response”
• Uplift can only be measured against a control group or background
response rate
• Campaign profitability is measured from Uplift only not total
responses
• Where did the Uplift come from?
• Can we target on Uplift only?
Page 6
McCoy’s Revelation
“It’s worse than that —
he’s dead, Jim”
– Dr Leonard H. McCoy
page 7
Yes
Fundamental Segmentation for Retention
SLEEPING DOGS
���
LOST CAUSES
�
CHURN IF
TREATED
Page 88
No Yes
No
PERSUADABLES
�
TREATED
CHURN IF NOT TREATED
SURE THINGS
��
When are negative effects likely?
dissatisfied / angry customers
risqué / offensive
communications
attrition risk
9
communications
intrusive contact mechanisms
forgotten standing charges
Sophis
tication
Land Grab
Target on
value at risk
Target on
attrition risk
Significant Event
Increasing Sophistication in Retention
Save Prob × Value
Target on
savable
value
Time
Land Grab
Attrition Probability × Value
MAINSTREAM
CURRENT
PRACTICE
Attrition Probability
LAGGARDS
attrition risk
Untargeted
retention
Acquisition
Page 10
STATE-OF-THE-ART
When Things go Well – Mobile Phone Contract Renewal
Yes
SLEEPING DOGS
LOST CAUSESCHURN IF
TREATED
Page 11
No Yes
No
PERSUADABLES
�
TREATED
CHURN IF NOT TREATED
SURE THINGS
When Things go Wrong – Mobile Telco
SegmentNumber of
CustomersSaves
Save
Rate %
Value of
ResponseTotal Profit
Treated 100,000 90,000 90% 18,000,000£ 8,950,000£
Control 100,000 91,000 91% 18,200,000£ 18,059,000£
Uplift 1,000- -1% 200,000-£ 9,109,000-£
Page 12
Value = £200
Contact Cost = £0.50
Incentive = £100
Avoiding Churn Stimulation – Credit Card Re-Pricing
Value = £50
Contact Cost = £0.50
SegmentNumber of
Customers
Improved
Performance
Improvement
Rate %
Value of
ResponseTotal Profit
Treated 100,000 49,520 49.5% 2,476,000£ 2,426,000£
Control 100,000 43,470 43.5% 2,173,500£ 2,173,500£
Uplift 6,050 6.1% 302,500£ 252,500£
page 13
Contact Cost = £0.50
Incentive = £0
PERSUADABLES
�
SLEEPING DOGS LOST CAUSES
SURE THINGS
• Customer Management
often involves difficult
conversations
• Knowing for whom this will
have a wholly beneficial
effect is a big win.
Customer Case Studies
Initiative ROI
Use Customer Insight as never used before: Leverage Data and Analytical Excellence to increase customer retention and profitability
Key breakthroughs with Portrait Analytics and Uplift Optimizer at Lloyds TSB Insurance
• Added over £10 million incremental profit per year in the last two years
• Increased the retention rate by 4.09%, 300% times better ROI
Business Goals
ROI
Initiative ROI
(£/per year)
Customer retention 8,000,000
Customer Cross-sell Cost Savings 400,000
Customer Loyalty Cost Savings 1,100,00
Customer Acquisition Growth 600,000
Customer Acquisition Cost Reduction 600,000
Online acquisition cost avoidance 100,000
Total ROI 10,800,000
300% times better ROI
• Increased the marketing response rate by 35%
47 k
108 k
164 k
193 k189 k184 k186 k
156 k
110 k
66 k
0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k
0%
1%
2%
3%
4%
5%
6%
0% 20% 40% 60% 80% 100%
0 k
50 k
100 k
150 k
200 k
250 k
Uplift Profit
Strong Positive Effect Marginal Effect Negative Effect
47 k
108 k
164 k
193 k189 k184 k186 k
156 k
110 k
66 k
0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k0 k
0%
1%
2%
3%
4%
5%
6%
0% 20% 40% 60% 80% 100%
0 k
50 k
100 k
150 k
200 k
250 k
Uplift Profit
Strong Positive Effect Marginal Effect Negative Effect
page 14
Customer Case Studies
Deliver next generation targeting for customer retention (churn) campaigns
Removing negative effects thus decreasing both churn and campaign costs
Campaign ROI increased over 11 times previous campaigns
Churn rate decreases by an additional 36% over traditional approach
• Traditional targeting reduced churn by 5% (from 23% to 18%)
• But significant negative effects observed
• Uplift model reduced these significantly
Business Goals
ROI
Results • Uplift model reduced these significantly
• Saving 40% of treatment costs
• Further reducing churn to 6.8%
Without Uplift
With Uplift
Treated
Volume
100%
Treated volume reduced by 40%
60%
Reduction in
Overall Churn
Without Uplift
With Uplift
5% 6.8%
Results improved
36%
19 %
15 %14 %
12 %
14 % 13 %
15 %14 %
10 %11 %
9 %8 %
11 %10 %
9 %
23 %
19 %
17 %
14 %
18 %
16 %
20 %18 %
11 %
13 %
11 %
9 %
13 %12 %
11 %
0 %
5 %
10 %
15 %
20 %
25 %
Introduction of
first prediction
model
Introduction of
refined prediction
model
months
an
nu
al d
efe
ctio
n r
ate
page 15
Lessons
●Embed value within the relationship and
avoid proactive retention
●Measure, Measure, Measure the
incremental effect and look for negative
effects
●Model on saveability not churn or attrition ●Model on saveability not churn or attrition
likelihood
●Avoid the sleeping dogs, the lost causes
and the sure things
page 16
Europe (Headquarters)
The Smith Centre, The Fairmile
Henley-on-Thames, Oxfordshire,
RG9 6AB, United Kingdom
T: +44 (0)1491 416600
F: +44 (0)1491 416601
Americas
125 Summer Street
16th Floor
Boston MA 02110, USA
T: +1 617 4575200
F: +1 617 4575299
Asia Pacific
Level 7
15-17 Young Street
Sydney NSW 2000
Australia
T: +61 2 9276 2728
F: +61 2 9276 2799
Edinburgh
39 Melville Street
Edinburgh
EH3 7JF, United Kingdom
T: +44 (0) 131 220 4491
F: +44 (0) 131 220 4492
ScandinaviaMaridalsveien 87, Bygg 1
0461 Oslo
Norway
T: +47 22 38 91 00
F: +47 23 40 94 99