2. 2
Altran Group
International Entrepreneur – Present in Over 20 Countries
Founded 1982
in Paris
17.200 Employees
Turnover 2009:
1,400 Mill. €
European Leader
in Innovation Consulting
International Coverage:
• Global Account Management
• Competence Networks
4. 4
Business Intelligence (BI)
Executive Dashboards
Financial Monitoring
Marketing Intelligence
Loyalty Programs
Data Mining & Risk Management
Data Governance / Data Quality
Operational Activities Monitoring
Technical BI Migrations
Partnerships:
Data Quality
4
8
Offers
BI Portfolio @ Altran CIS
5. 8
A question arises …
Are my Sunday breadrolls really the best?
„Our breadrolls are the best in town!“
Enquiring with the bakery:
6. 9
One question remains…
Are my Sunday breadrolls really the best?
I have to buy other breadrolls.
(no alternatives!)
Measurement and comparison are necessary:
Advantages
• Possibility to compare
• Learnings and Optimisation
potential
Disadvantages
• Costs
• Time & Effort for
analysis
7. 10
What is obvious with breadrolls seems less obvious with campaigns …
Are my campaigns really the best?
„Our campaigns are the best!“
Enquiring with the campaign manager:
8. 11
One question remains…
Are my campaigns really the best?
I have to run simultaneous campaigns.
(no alternatives!)
Measurement and comparison are necessary:
Advantages
• Possibility to compare
• Learnings and Optimisation
potential
Disadvantages
• Costs
• Time & Effort for
analysis
Campaigns for comparison
9. 12
One question remains …
Campaigns for comparison
Advantages
• Possibility to compare
• Learnings and Optimisation
potential
Disadvantages
• Costs
• Time & Effort for
analysis
About disadvantages:
10. 13
About disadvantages: Costs
3 possibilities for simultaneous campaigns
1. Advertise competitors products:
2. Advertise different product:
3. Virtual campaign:
Counterproductive
Ineffecient
Twice expensive
Ineffecient
Expensive
Irrelevant
Simple
Cost-minimising
Efficient
Virtual campaign: approx. 10% * of the selected customers will simply NOT be contacted
Contacted = Target group Not contacted = Control group
* Precise percentage depends on general framework.
11. 14
One question remains …
Campaigns for comparison
Disadvantages
• Costs
• Time & Effort for analysis
About Disadvantages
Advantages
• Possibility to compare
• Learnings and Optimisation
potential
12. 15
About disadvantages: Time & Effort for analysis (1/8) guarantee representativity
0%
5%
10%
15%
20%
25%
30%
35%
40%
less than
18
18-30 31-45 46-60 more
than 60
Target group
0%
5%
10%
15%
20%
25%
30%
35%
40%
less than
18
18-30 31-45 46-60 more
than 60
Control group
0%
5%
10%
15%
20%
25%
30%
35%
40%
less
than 18
18-30 31-45 46-60 more
than 60
Control group
Representative – mirrors distribution of the
target group
Not representative – considerably different
distribution than target group
Typical properties:
Geo-status, salary, …
Age distribution of target group
13. 16
Control group is representative
iff p(χ²) > 0,1
About disadvantages: Time & Effort for analysis (2/8) in all relevant criteria distributions
have to match
15. 18
•Conversion Rate
•Lift (Grundrauschen)
•Liftbeitrag als Ergebnis
•Return on Investment0
0.2
0.4
0.6
0.8
1
1.2
1.4
Target Group Control Group
Conversion Rate
(overall)
Conversion Rate
(focus product)
Overall conversion rate is significantly different, however
not for the product in focus.
Relationship to customer OK,
focus product campaign feeble.
About advantages: Time & Effort for analysis (4/8) Descriptive parameters
16. 19
•Conversion Rate
•Lift (Background Noise)
•Liftbeitrag als Ergebnis
•Return on Investment
Equally structured target group produces conversions
independent of campaigning!
Lift counts the conversions that
can be attributed to the
campaign (230 conversions).
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Target Group Control Group
Conversion Rate
(overall)
Conversion Rate
(focus product)
Lift = 0,23%-points
B
a
c
k
g
r
o
u
n
d
N
o
i
s
e
About advantages: Time & Effort for analysis (5/8) Descriptive parameters
17. 20
•Conversion Rate
•Lift (Background Noise)
•Lift Leverage
•Return on Investment
Equally structured target group contains earnings
independent of campaigning
(wrong: 1.240 * 900€ = 1,1 Mio. €).
Earnings due to campaigning are
230*900€ = 207.000€.
0
200
400
600
800
1000
1200
Target Group Control Group
Leverage
(900€ /conversion)
Conversion Rate
(overall)
Campaign profit
C
a
m
p
a
i
g
n
i
n
d
e
p
e
n
d
e
n
t
About advantages: Time & Effort for analysis (6/8) Descriptive parameters
18. 21
•Conversion Rate
•Lift (Background Noise)
•Lift Leverage
•Return on Investment (RoI)
RoI describes profit per spent Euro of the additional conversions due to
campaigning
How many Euro were earnt given any Euro spent?
RoI = -100%: total loss of investment
RoI = 0%: investment equals the profit due to campaigning
RoI of this campaign is
(207-130*)/130 = 59%,
i.e. per Euro spent
59 Cent earnt.
0
50
100
150
200
250
Earnings Costs
Earnings
vs. Costs
Conversion Rate
(whole)
Profit
Invest
* Costs of campaign with 100.000 people were assumed to be approx. 130 T€.
About advantages: Time & Effort for analysis (7/8) Descriptive parameters
19. 22
•Conversion Rate
• Characteristics of campaign
•Lift (Background Noise)
• Efficiency of campaign
•Lift Leverage
• Leverage of campaign
•Return on Investment (RoI)
• Performance of campaign
•Conversion Rate
•Lift (Background Noise)
•Lift Leverage
•Return on Investment (RoI)
About advantages: Time & Effort for analysis (8/8) Descriptive parameters
20. 23
One question remains …
Advantages
• Possibility to compare
• Learnings and Optimisation
potential
Disadvantages
• Costs
• Time & Effort for analysis
About advantages:
Campaigns for comparison
21. 24
About advantages: Possibility to compare
B
A
For second best offer company earns only 750€ (instead of 900€).
Also conversion rate in target group is clearly lower:
0
0.2
0.4
0.6
0.8
1
Target group Control group
Conversion Rate
(whole)
Conversion Rate
(focus product)
STILL THIS CAMPAIGNS PERFORMED A LOT
BETTER (COMPARED TO CONTROL GROUP):
ANY INVESTED EURO RETURNED 2,94€!
22. 25
Advantages
• Possibility to compare
• Learnings and Optimisation potential
• Time & Effort for analysis
Daisadvantages
• Costs
About advantages:
Campaigns for comparison
23. 26
About advantages: Learnings and optimisation potential: Traffic-lights
Due to quantification it becomes possible to map to traffic-lights and
reports which might even be legible for management.
PARAMETER
UNSATISFYING :
CHANGES REQUIRED!
PARAMETER NOT
SATISFYING :
CHECK PROPERLY!
PARAMETER
SATISFYING :
RE-RUN POSSIBLE!
24. 27
About advantages: Learnings and optimisation potential: Channel dependance and campaign
combinations
eMail
Outlet
Mail
Call Center
Invoice
Possible channels: Possible combinations
(follow-up campaigns):
… can and should be examined (quantified).
… can and should be compared and checked.
25. 28
About advantages: Correct Predictions and Forecasts
circulation
C
o
n
v
e
r
s
i
o
n
s
Doubled circulationSimple circulation
Conversions
counted
wrong
forecast
correct
forecast
Background Noise
Why is the number of conversions not doubled
with doubled circulation?
When ordering due to predicted conversion the correct forecast will be even less!
26. 29
Closed Loop as a basis of continuous optimisation in campaign performance
Combine Learnings and experience
Focus on right product and
campaign combinations
(learning matrix)
Process of planning Campaigns:
Selection of campaign participants
Roll-out of campaign
Analysis of any particular campaign
If need be: minor adjustments in
active process
Routine reporting of all campaignes
(possibility to compare)
If need be: specific analyses for
specific campaigns
Measure
and
Learn
27. 30
Presentation title / date / confidential / Issued
by ...
30
www.altran-cis.pt
Dr. Igor Schnakenburg
Senior Consultant
Email: igor.schnakenburg@altran.com
Phone: +49 (0)170 919 8272