Transcript Marketing Mix Modeling

Marketing Mix Models
© 2002 Marketing Management Analytics – www.mma.com
What is Marketing Mix
Modeling?
 An analytical approach which quantifies
the sales effect of marketing activity and
the financial return on that investment.
 The output is used to simulate the effects
of alternative marketing plans and
forecast sales into the future.
 All work, for all clients, is
completely custom.
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:2
Why do marketers do
Marketing Mix Analysis?
 To get a true empirical relationship between
marketing activity and sales.
 To conduct Return on Investment analysis.
 Benchmarking (are we the same / better /
worse than we were last year?).
 To find out how one business unit's marketing
interacts (if at all) with another’s (media “halo”).
 To learn upside potential and downside risk in
changing marketing spend.
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:3
Issues that can be addressed...

How much should we spend?



How should we spend it?


Which brands should receive support
and how should our budget be
allocated?
When should we spend it?



What is the recommended level of
spending?
How much is enough?
Before or after a price increase?
Immediately before a competitive
launch?
Where should it be spent?


1
2
3
4
National vs. Local
Cites, Rural
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:4
Marketing Mix Modeling
Marketing Mix Elements
Step 1
determine the contribution
to incremental volume from
each marketing mix
element
Step 2
overlay marketing
spending for each mix
element to evaluate ROI
Step 3
utilize MMA’s custom software for
optimization and simulation of
marketing spending and advertising
return
Marketing Spending
Custom Software
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:5
Mix Modeling Timeline
Project Steps & Timing
Kick-Off
Data
Collection
& Validation
4-5 Weeks
Model Specification
& Validation
4 Weeks
4 Weeks
Analysis
& Review
3 Weeks
1
3-5 Weeks
Final
Presentation
1
12-13 weeks
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:6
Data Collection
1
Sales
The capability to integrate several
disparate data sources
(Internal,
Syndicated)
Media
Nielsen
CMR
Agency
Promotion
Integrated
Database
Act Media
Catalina
Internal
Shipments
Web
Promotion
Financial
Data Analysis
System
2
The ability to identify what
variables should be included in
your model.
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:7
Marketing Mix Modeling
(How is it Done?)
© 2002 Marketing Management Analytics – www.mma.com
Modeling: What to measure?
Start with a Dependent Variable:
• Sales, or
• Awareness, or
• New Customers or
• Brand Interest
• Sales by Segment (i.e. Heavy vs. Light)
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:9
Modeling: What are the Drivers
Gather information on all variables that possibly
influence your dependent variable
Independent Variables:
• TV
• Promotions
• Competitive
• Print
• PR
• Economic
• Radio
• Coupons
• Environment/Weather
• Outdoor
• Sampling
• Industry Trends
• Internet
• Direct Mail
• Etc.
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:10
The models relate changes in sales to the
changes in marketing support and other
causal factors present during each week.
Sales
7
0
,
0
0
0
6
0
,
0
0
0
5
0
,
0
0
0
4
0
,
0
0
0
3
0
,
0
0
0
High
Activity
High
Activity
2
0
,
0
0
0
Low
Activity
1
0
,
0
0
0
0
1
5
9
1
3
1
7
2
1
2
5
2
9
3
3
3
7
4
1
4
5
4
9
5
3
5
7
6
1
6
5
6
9
7
3
7
7
Support
Week
TV GRPs
Radio GRPS
In-Store Promotions
1
10
13
214
100
75
22
142
-
33
60
30
41
50
5
49
40
-
58
89
-
67
284
40
20
70
65
-
Copyright © 2002 Marketing Management Analytics
78
25
www.mma.com
Page:11
Model Validation
- Signs
- Volume Trends
- Trade Contribution
Face Validity
Holdout Tests
Model
Stats
Other
Holdout Errors
&
Patterns
R-Square, Durbin
Watson, Avg % Error,
F-Stats, T-Stats, etc.
Other validations
against prior learnings
or known relationships
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:12
Basic Model Stats & Holdout
Validation
Analyzed Market Aggregate
Sales
700,000
600,000
500,000
400,000
Fit Period
Average
Error: 2.3%
300,000
200,000
Validation
Period
Average
Error: 2.8%
100,000
0
-100,000
5/2/93
8/1/93
10/31/93 1/30/94
5/1/94
7/31/94 10/30/94 1/29/95
4/30/95
104 Weeks Ending 4/95
Model
residuals
exhibit
nomuch
patterns
Models
The
model
need
should
to beshould
able
“explain”
to
predict
as
volume
of and
the
in the
significant
weekly
be understood.
near
weekly
future.
variance
Modelinmisses
holdout
sales should
as
validations
possible.
can
The help
RAverage
errors
and
Durbin
Watson
statistics
can
Squared
evaluate
statistic
a model’s
quantifies
predictiveness.
a models fit. help
Model Estimate Actual Sales Residuals
with these issues.
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:13
Marketing Mix Modeling
What Does It Answer?
© 2002 Marketing Management Analytics – www.mma.com
Incremental volume will be broken out
into the various marketing elements.
Volume Contribution - Example
Base
Promotions
Direct TV
Print
Total Volume 102.0 MM
Radio
Corp TV
Total Volume 110.0 MM
7.5%
50.0%
23.6%
9.7%
54.2%
19.5%
7.5%
Prior
Year
3.8%
7.5%
7.8%
3.9%
4.9%
Current Year
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:15
Comparing ROI efficiencies across vehicles
shows that all TV and Trade spending is
very efficient.
Marketing Efficiencies - Example
(Incremental $ Sales Per Marketing $)
$3.6
$3.3
$3.8
$4.2
$3.0
$2.6
TV Adv
Trade
SMF
Print
Radio
Total
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:16
Changes in yearly support, resulting contribution
to sales and efficiency will be reported.
Direct TV Advertising Performance - Summary
(Example)
Support
TRPs
% Contribution
Efficiency
% of Total Brand Volume
Cost per Incr. Vol
2.2%
1256
$3.41
$2.69
872
1.9%
Prior Year Current Year
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:17
Evaluation of Copy Effectiveness
TV Advertising and Cost Efficiency
Copy A vs Copy B
Volume Effectiveness
Cost Efficiency*
Volume Per MM Impressions
Total Copy A = 7,535
Cost Per Incremental Unit
$0.51
8,586
$0.42
7,987
7,458
All Other
Copy A
New Flavor
Copy B
Prior Year - Copy A
Current Year - Copy B
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:18
ROI efficiencies can also be compared across categories
by brand, department and marketing investment.
Profit per $1 Invested
Total Mix
1.98
Brand 1
Brand 2
1.20
Brand 4
1.24
Brand 6
Brand 7
Brand 8
Brand 9
0.85
1.10
0.66
2.10
1.17
1.24
In-store
1.30
0.94
0.90
1.14
1.20
0.48
0.68
0.80
1.34
1.17
3.02
0.45
0.65
1.50
0.61
0.68
2.54
0.54
1.30
0.65
Cons. Promo.
0.56
1.34
0.40
0.57
0.84
Print
2.34
0.93
Brand 3
Brand 5
TV Media
0.72
1.34
0.60
0.61
0.76
0.70
0.84
0.74
1.22
Copyright © 2002 Marketing Management Analytics
www.mma.com
Page:19