Market Mix Model

(MMM)
 Why 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.
Questions Marketing Mix Can Address
How much should we spend?
• What is the recommended level of spending?
• What is the marginal opportunity?

What should we spend on?

• Which brands should receive support?
• How should our budget be allocated?
• Which media and promotional vehicles?

How should it be spent?

• Which tactics? Dayparts? Commercial lengths?
• Which coupon face values? Which trade tactics?

When should we spend?

• Continuously? Flighted? In season?
• Before or after pricing/promotion events?

Where should we spend?

• Which markets?
Clients Reasons for Using MMA Results

Ways MMA Results Are Used

% of Responses
Allocate/Refine Marketing Spending
43%

Media Planning
23%

Due-to's/What If Scenarios/Forecasting 11%

Evaluate Trends
8%

Defend New Product Support

4%

Other 11%

Base: 2021 Survey Response

Sales are Driven by Many Factors Example
To drive sales, company X allocates its resources to three areas: TV, Radio, and Trade

Spending Support

($ in Millions)

$2 400 GRPs
Radio

Trade $40 40% ACV Sales

TV $15 1450 GRPs

Total $57
The models relate changes in support to the changes in
sales
Sales
7
0
,
00
0

6
0
,
00
0

5
0
,
00
0

4
0
,
00
0

3
0
,
00
0 High
Activity High
2
0
,
00
0
Activity
1
0
,
00
0
Low
Activity
0
1591
31
72
12
52
93
33
74
14
54
95
35
76
16
56
97
37
7
Support

Week 5 13 22 29 38 45 53 61 65 78
TV GRPs - 214 142 - 50 - 89 - 284 -
Radio GRPS - 100 - 60 - 40 - 22 - -
In-Store Promotions 10 75 - 30 5 - - - 25 25
The Model Isolates Factors Impacting Marketing
Incremental and Base Sales

Total Brand

Incremental Base
• TV Media • Price
• Print • Distribution
• Radio • Competition
• Trade • Long-term impact from marketing
• CP Trade
• Internet, etc.

CP
Key Factors under Consideration
Media Trade
Sales/Volume • Features
• TV
• Displays
– By market • Features & Displays
– By daypart • TPR
• Magazine Brand X Sales • Discounting
• Newspaper
• Radio
• Internet

External Factors
• Competition
Consumer • PR
• FSIs • Economy
• Sampling
Other • Weather
• Catalina • Distribution (by pack size, • Catastrophic events
• Act Media segment, form, etc.)
• Pricing
• Etc.
Simplified Model Form
Sales = Constant
+ bT * fT (TV advertising)
+ bR * fR (Radio advertising)
+ bTP * fTP(FSI circulation)
+ bD * fD (Distribution)
+ bP * fP (Base Price)
+…
+e

¨ Model Inputs
¨ Empirically determined transformations
¨ Model optimized coefficients
Data Transformation
Some marketing events impact sales within the period of execution (price discounts), while other
events impact sales for an extended period following the execution (TV, Print, etc).

120

100

80

GRPs 60

40

20

Weeks
How Do We Model Advertising?
Transforming delivered GRPs into a working "adstock", accounting for both advertising saturation and
advertising decay.

• A brand's adstock is a function of the GRPs aired for it.

GRPs
• Every week without advertising causes some adstock “leakage.”

• The impact of an additional GRP depends on the current level of

adstock.
Saturation
• A GRP aired today is known to have an impact on sales in future
Principles weeks. The magnitude of that impact varies based on GRP levels
in preceding weeks.

Adstock

Decay
Advertising also experiences diminishing returns. In other words, at higher levels of advertising, the
marginal impact of additional GRPs gets smaller .

Low Marginal
Opportunity

Contribution

Very High Marginal

Opportunity

GRPs: 10 100 500 1000

Reach: ~10% 65% 75% 80?
Estimated Sales/Volume Composition
The marketing mix model determines the weekly contribution by each factor

Base Holiday TV Adv Print Trade

Inventory

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77
 Determining Factors Driving Changes
Comparing sales by factor in various periods, we can determine the factors of volume change.

Inventory
Base Holiday TV Adv Print Trade

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77

Factor Period 1 Period 2 % Change vs. Year Ago

Total Sales 650,000 700,000 +7.6%
TV 195,000 200,000 +2.5%
Print 20,000 30,000 +50%
Metrics
Incremental Contribution
Percent of a brand's volume driven by a marketing activity (e.g. TV drives 5% of brand’s volume). Short-term
volume gains only.

Base Volume
Total volume excluding short-term gains from all marketing elements. (Nielsen’s defined baseline excludes Trade
incremental only, while baselines from mix models excludes all measured short-term volume from marketing.)

Subsidized Volume
Promoted sales volume that would have sold in the absence of marketing activity.

Due-Tos
Measures a change in contribution by marketing element.
Incremental, Base, Subsidized

Volume Decomposition

TV Higher
Coupon Volume
Subsidized
Risk
Trade
Incremental
Trade
Base
Highest TV Volume Lower
Volume Core Volume
Risk Franchise Risk

Coupon

Base
Volume
Sources of Volume Change
(Due-To) Definition
• A Due-To chart reflects drivers of change from year-to-year.
• Due-To’s look at the difference in volume driven by a given marketing element, relative to the
total change for the brand.
• In this example, the brand’s total volume increased +20% (vs. Year 1). This can be explained
by growth in Trade, causing a +10% increase on the total business. Brand and Print Media are
also driving a +5% increase each.

Year 1 Year 2 Difference between: Due-To

52 Weeks 52 Weeks
Year 1 and Year 2 Total % Change (vs. Year 1)
Ending 2/02 Ending 2/03

Total Volume 1,000 1,200 200 20%

Direct Mail 700 800 100 10%

Brand TV Media 150 200 50 5%

Print Media 150 200 50 5%

Source of Volume Change (Due-To)

% Change in Volume vs. Prior: -3.9%

1.2%

0.4% Items
0.2% Comp. C Comp. B per
CP
Print Price TV Price Store
Trade Comp.A Price
Price Decrease -0.2% -0.3%
-0.6%
-0.8%

-1.8%
-2.0%
Effectiveness
The actual amount of a brand's volume driven per unit of marketing activity (e.g., pounds per 1MM
television impressions.) Equivalizes for the effort amount.
TV
Incremental units or sales dollars per 100 HH GRPs or MM Impressions
Print & Radio
Incremental units or sales dollars per 100 Target GRPs or MM Impressions
FSI
Incremental units or sales dollars per 1 MM circulations, or 1 MM Circ. Dollars (circ.* face value)
Trade
Incremental units or sales dollars per:
• $10 MM ACV Promoted Store Weekly

Efficiency
Provides the financial link between spending and the incremental volume generated as a result.
Expressed as either incremental units generated per dollar spent (higher = better) or dollar spent per
incremental unit (lower = better).
TV Effectiveness & Efficiency

TV Effectiveness By TV Copy
Incremental Units. Per MM Impressions

Most Effective 3,334

2,039
1,191

Copy A Copy B Total

HH GRPs: 800 1,200

Financial Efficiency By Marketing Element

Cost per Incremental Unit
$4.54 Most Efficient
$3.82

$2.39 $2.63 $2.20

TV Print Radio Coupons Trade

Payback
Interpretation:
• Every $1 spent on TV advertising has resulted in 77 cents in incremental, short-term profit
• Longer-term benefits of advertising are not accounted for in payback.

Payback = Incremental Contribution (Eq. Units) X $ Profit Per Eq. Unit

Spending
From the model:

TV Pay back = 76.1 MM Units X $.07 / Units = $.77, or 77%

$6.8 Million
TV Halo
TV ad volume driven by another product’s advertising.

Half-Life
The number of weeks for a commercial to lose half of its advertising impact after airing.

Saturation – Marginal Opportunity Index

Measure of Media saturation. Effectiveness of next GRP based on existing media plan.
TV Advertising Halo
Brand A TV copy generates volume for Brand A Volume and Brand B (halo). And vice-
versa.

Brand A Volume Brand B Volume

GRPs: 1,804 2,953

3.5%

2.9%

1.1% 1.1%

Brand A Copy Brand B Copy Brand A Copy Brand B Copy

Direct Halo Halo Direct