Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
(2017) 4:18–27
DOI 10.1007/s40547-017-0070-2
RESEARCH ARTICLE
Abstract Industries invest in over billion dollars annually to Keywords Generic advertising . Brand advertising .
drive up the primary demand and to fence off competing new Inter-industry competition . Game theory
industries’ threat to their customer bases. Companies in addi-
tion to contributing to industry generic campaign also heavily
rely on brand advertising in order to capture a greater market
1 Introduction
share. Using a game-theoretic approach, we study generic and
brand advertising competition under such an inter-industry
More than a billion dollars is spent annually on generic adver-
competitive framework. We built an analytical model to study
tising aiming to drive up the primary demand for the entire
two competing industries each simultaneously making generic
product category (Armbruster and Nichols 2001).1 In addition
advertising decisions followed by firms within each industry
to the agricultural or commodity products such as beef,
simultaneously conducting brand advertising. We found that
cheese, and orange juice [7], more recent examples of such
the mere presence of a rival industry can act as an impetus for
industry campaign have also been found in large ticket items
an industry to invest in generic advertising. Model analyses
including the diamond jewelry. De Beers launched the
and numerical studies suggest that there is a clear interactive
category-driving campaign BFewer, better things,^ in 2008
nature between the two types of advertising decisions under
following the financial crisis. By 2015, the combined annual
inter-industry competitive framework. The generic advertising
budget with another generic campaign BForever Mark^ has
spending of an industry increases as the firms within that
equaled around half of De Beers’ annual promotion spending,
industry are more asymmetric. While a firm’s brand advertis-
$100 million [3].
ing spending increases as the generic advertising of its asso-
Recent advances in the social and technological environ-
ciated industry becomes more effective and that of the rival
ment have led to the appearance of new industries whose
industry becomes less effective. Extensions of the main model
products or services compete with traditional industries that
suggest that there is a first-mover advantage in generic adver-
satisfy similar needs of customers in the market. Examples
tising under inter-industry competition.
include, but are not limited to, electric cars versus cars with
internal combustion engines, solar panel versus the traditional
rooftop, organic versus conventional food, and the
* Yuanfang Lin commission-free house transaction versus the conventional
ylin1@conestogac.on.ca practice involving real-estate agents. Correspondingly, many
industries have been using generic ad campaign to counteract
Sandeep Krishnamurthy the threat from competing industries that are squeezing their
sandeep@u.washington.edu primary market base. Two most recent industry activities of
1
this kind include the Bsolar smear^ campaign funded by
School of Business & Hospitality, Conestoga College, 299 Doon Utility Arizona Public Service meant to turn Arizona
Valley Drive, Kitchener, ON N2G 4M4, Canada
2 1
School of Business, University of Washington, Bothell, 18115 One of the most famous examples is perhaps the BGot Milk?^ campaign by
Campus Way NE, Room UW1-233, Bothell, WA 98011-8246, USA the California Milk Advisory Board.
Cust. Need. and Solut. (2017) 4:18–27 19
ratepayers against net energy metering [30], and the National The remainder of the paper is organized as follows.
Association of Realtors BGet Realtor®^ campaign targeting Section 2 discusses relevant literatures and the contribution
the emerging generations of home buyers, sellers, owners, and of this paper. Section 3 presents the setup of the analytical
investors [32]. An industry’s generic campaign budget comes model. Section 4 derives the analytical results, presents the
from various sources with an important one being the contri- main model findings, and provides further discussions via
butions from companies within the same industry via certain numerical studies. Section 5 concludes with summary, mana-
mechanism (see, for example, [16, 17]). This is how the gerial recommendations, and a brief discussion of future re-
Canada Organic Trade Association launched the BThink search directions. All proofs are delegated to the Appendix.
Before You Eat, Think Canada Organic^ campaign
which involves the purchase of print and online ads
throughout the year from broad public media such as 2 Related Literature
CBC, The Globe and Mail, Bfoodie^ periodicals like
the Edible magazines, as well as social networks such This paper is mainly in the nexus of two streams of literature:
as Facebook and Twitter [28]. the inter-industry competition and the generic vs. brand ad-
Besides contributing to the associated industry’s category- vertising decision. We provide a brief review here before stat-
driving campaign, individual companies also heavily invest in ing the contribution of this paper.
advertising that promote own brands aiming to win over con-
sumers from other competing brands for a greater market 2.1 Inter-Industry Competition
share. Inside the newly surging organic milk industry,
Horizon Organic’s advertising budget totaled $9.3 million, The game-theoretic literature in marketing generally focuses
$9.5 million, and $10.2 million in the year 2008, 2009, and on competition within an industry on the four Ps-product (e.g.,
2010, respectively. While Silk spent $29.1 million to advertise [13]), price (e.g., [21]), distribution (e.g., [18]), and promotion
in major media in 2009 alone [23]. (e.g., [11]). The effects of one industry’s marketing activities
The main objective of this paper is to understand the rela- on other related industries have received lesser attention. Even
tionship between generic and brand advertising strategies when product-level competition among substitutes is studied,
vis-à-vis inter-industry competition. In particular, we ask the the focus is generally within the industry rather than across
following research questions: how should an industry respond industries (e.g., [21]). This might be due to the conventional
to a generic advertising campaign by a rival industry? Will the thought that there is no competition between firms in different
generic advertising by one industry or by its rival industry industries.
affect the brand advertising strategy of competing brands Within the economics literature, inter-industry dynamics
within the focal industry? Does the brand advertising intensity with respect to brand advertising has been of some interest
(and the resulting asymmetry between competing brands) following the Bgalbraithian hypothesis^ that brand advertising
within an industry affect the generic advertising decision at in one industry might shift demand across industries [12].
the industry level? Will an industry benefit from making ge- However, the general finding has been that Bif advertising is
neric advertising prior to its rivals? effective, its major effect is to alter market shares within an
To answer these questions, we construct a multi-stage industry rather than to increase the general demand for the
game-theoretic model with two competing industries each product^ ([31], p. 223). Interest in inter-industry dynamics is
consisting of two competing brands. In the first stage, each growing, however, as evidenced by recent interest in collusion
industry simultaneously decides on the level of generic adver- among differentiated industries (e.g., [29]) and generic adver-
tising and in the second stage, firms within each industry tising (e.g., [2]).
simultaneously make brand advertising decisions. Findings
from our analytical model and numerical studies suggest that 2.2 Generic versus Brand Advertising
the mere presence of a rival industry can act as an impetus for
an industry to invest in generic advertising. There is a clear The study of generic advertising in marketing has been in-
interactive nature between the two types of advertising deci- creasing in the last decade. Krishnamurthy [15] studies the
sions under inter-industry competitive framework. The gener- fundamental tradeoff between improving market share
ic advertising spending of an industry increases as the firms through brand advertising (i.e., going for a larger slice of the
within that industry are more asymmetric. While a firm’s pie) and expanding the size of the market through generic
brand advertising spending increases as the generic advertis- advertising (i.e., expanding the pie itself). A free-riding equi-
ing of its associated industry becomes more effective and that librium is found under the independent-contribution case
of the rival industry becomes less effective. There is also a where the dominant firm in the industry makes all contribu-
first-mover advantage for inter-industry competition via ge- tions to the generic advertising campaign while being indif-
neric advertising. ferent to the free riding of lesser firms. And the government-
20 Cust. Need. and Solut. (2017) 4:18–27
sponsored mechanism would lead to a higher generic adver- the high-speed railway (CRH) versus major airlines in China
tising budget with an accompanying increase in brand adver- serving the same city routes, or a solar versus conventional
tising. Krishnamurthy [16] studies a provision-point mecha- roof top installation. A duopoly setting is incorporated to
nism to fund generic advertising campaigns. To implement model intra-industry competition within each industry j.3
such mechanism, an industry association sets a target with That is, there are two (major) brands/firms, indexed by i (=1,
the understanding that the campaign will not be conducted if 2) within each industry.4 As in Bass et al. [2], we study the
target contributions are not achieved. This mechanism outper- case where advertising is the dominant marketing mix variable
forms the simple voluntary contribution mechanism theoreti- in market competition.5 Thus, demand for the product of firm i
cally and in experiments. Agricultural economists have further in industry j is given by the market size of this industry mul-
investigated the provision-point mechanism, and it is now tiplied by the specific firm’s relative market share. Firms can
viewed as a credible alternative to the voluntary contribution help enlarge the market size of the associated industry by
mechanism [10, 19, 20]. Bass et al. [2] analyze the equilibrium contributing to the generic advertising campaign at the indus-
generic advertising decisions for a dynamic duopoly. They try level and try to increase their relative market share through
confirm the intuition of Krishnamurthy [15] that Bwhen the own brand advertising.
asymmetry between the firms increases, there is a greater dif-
ference between their generic advertising contributions.^ (p.
3.2 Generic and Brand Advertising
562). However, they find that when a dynamic competitive
model is introduced, Bthe weaker firm’s investment in generic
The profit function for firm i of industry j can be written as:
advertising never goes to zero.^ Qi et al. [24, 25] apply chaos
theory to the model in Krishnamurthy [15] and study both πij ¼ M j G j ; G3− j Dij sij −Gij −Bij ð1Þ
symmetric and asymmetric systems. They find that there
could be Bcomplex bifurcating and chaotic behavior for the where sij is the per unit margin for firm i in industry j without
generic advertising efforts^ and derive the inter-relationships any advertising,
between brand and generic advertising expenditures. Gij is the firm i’s contribution to the associated industry j’s
Despite its obvious impact on competing industries, mar- generic advertising campaign (Gj = Gij + G3 − i , j),
keting studies on generic advertising has largely been based Dij is the market share of firm i in industry j,
on single-industry framework (e.g., [2, 15, 16, 19, 20]). The Bij is the brand advertising spending of firm i in industry j,
focus has either been on the interaction/coopetition between and
brand and generic advertising within an industry (e.g., [2, 5, Mj is the market size for industry j which is determined by
15, 22, 25]) or on funding mechanism for the generic adver- its initial (/existing) customer base (M0j) and potential expan-
tising campaign (e.g., [6, 16, 19, 20]).2 Roma and Perrone [27] sion gained from the generic advertising of its own (Gj) and
studies two firms’ simultaneous decisions in generic and brand counteracted by that of the competing industry (G3 − j).
advertising within an industry, and the associated price com- Mathematically,
petition with the study focus on the free-riding effects and the
M 0 j þ α jG j
coordination mechanisms of generic campaign. To our best Mj ¼ ; j ¼ 1; 2: ð2Þ
knowledge, this paper is the first to fill the gap in the literature M 0 j þ α j G j þ M 0;3− j þ α3− j G3− j
by presenting a general model to study the interaction between where αj > 0 is the parameter representing the effectiveness of
generic and brand advertising decisions under inter-industry industry j’s generic advertising campaign. The market size of
competition framework. industry j increases with his own investment in generic adver-
tising and decreases with the rival industry’s generic advertis-
ing. If neither industry has any generic advertising, market
size of industry j (Mj) will be determined by his initial cus-
3 Model Setup tomer bases in comparison to that of the competing industry.
The formulation in Eq. (2) has a long history in marketing
3.1 Inter-Industry Competitive Structure
research literature and has a good empirical validity (e.g., [4,
8]). In order to more closely model, the recently arising com-
We consider two competing industries, indexed by j (=1, 2)
petition between traditional and new product industries (e.g.,
providing substitutable products for a specific need of con-
sumers in the market. Examples include, but not limited to 3
We investigate in Section 4.4 the asymmetric case where there is an unequal
number of firms within each industry.
2 4
There are agricultural economic studies where the focus is on how policy A recent example could be iOS versus Android for smartphone operation
changes such as the introduction of a specific tax affects industries of substi- system.
5
tutable products (e.g., [14]) and how generic advertising changes when pro- For example, the CRH ticket price for the Beijing-Shanghai route is reported
duction is rationed in a competing industry (e.g., [9]). to be fairly close to the average discount airfare.
Cust. Need. and Solut. (2017) 4:18–27 21
conventional versus solar top roofing), we make some simpli- 4 Solution and Numerical Analysis
fication without loss of generalities. In particular, we set
M01 = ϕ > 0, while M02 = kϕ with 0 < k ≤ 1. The varying level 4.1 Equilibrium of Generic and Brand Advertising
of k would help capture different market scenarios of inter-
industry competition. For example, a larger k could represent Using backward induction, we start with the analysis of firms’
the situation where the newly emerging industry 2 has been individual profit maximization by making the brand advertis-
quickly establishing customer base thus posing a more serious ing decisions knowing the industry generic advertising level,
threat to the traditional industry 1.6 and then move up to the first stage solving for the industry’s
Within industry j, firm i’s market share (Dij) is determined generic advertising decision to maximize total industry profit.
by the interaction between the brand advertising of its own To reduce some computation burden while preserving the es-
and that of the competing brand in the same industry j sential components of our mathematical model, we make the
(Bij ∈ [0 , + ∞ ) in dollars). Mathematically, further simplification that sij = sj , ∀ i = 1 , 2. In other words,
the two firms within the same industry have equal per unit
Aij
Dij ¼ ð3Þ margin before advertising. Thus, any potential asymmetry be-
Aij þ A3−i; j tween the two firms with the industry will be due to the com-
where Aij is the attraction function. Consistent with the litera- parative effect of the two firms’ brand advertising.8 The fol-
ture (e.g., [15, 27]), we specify the attraction function ( Aij) as: lowing Proposition 1 states the equilibrium advertising strate-
gies of the two competing industries and individual firms
1=2
Aij ¼ θij Bij ð4Þ within each industry.
θ θ
where θij > 0 represents the effectiveness of firm i’s brand Proposition 1 Let Y j ¼ 1− ij 3−i; j 2 , λj = αjYjsj,∀i , j = 1 , 2,
advertising within industry j. According to Eq. (3), the brand ðθij þθ3−i; j Þ
kα j Y j s j λ2j λ3− j
advertising in the competing industry does not affect the mar- 1.) When 0 < ϕ < min ð1þk Þ2 ; , the equilibri-
ðλ j þλ3− j Þ
2
ket share of firms within the focal industry. This is a reason- um generic advertising level for industry 1 (with initial
able characterization according to the advertising literature λ21 λ2 −ϕðλ1 þλ2 Þ2
(e.g., [2, 15, 16]). market base ϕ) is G*1 ¼ α1 ðλ1 þλ2 Þ2
, while the equilib-
rium generic advertising level for industry 2 (with small-
λ1 λ22 −kϕðλ1 þλ2 Þ2
3.3 Game Structure and the Sequence of Moves er customer base kϕ ) is G*2 ¼ α2 ðλ1 þλ2 Þ2
. And the
equilibrium brand advertising level for firm i of industry
We study a two-stage game where at Stage 1, the two indus- λ j θij θ3−i; j s j
j is B*ij ¼ ;
2ðλ j þλ3− j Þðθij þθ3−i; j Þ
2
tries simultaneously decide their generic advertising spending
(Gj). After knowing each industry’s total generic advertising 2.) Otherwise, the equilibrium generic advertising level for
spending, the individual firms simultaneously decide own industry j (=1, 2) is G j ¼ 0, and the equilibrium brand
brand advertising expenditures (Bij) at Stage 2. Profits are then advertising level for each firm i of industry j is
θij θ3−i; j s j
realized as a result for each firm. We look for subgame-perfect Bij ¼ .
2ð1þk Þðθij þθ3−i; j Þ
2
Nash equilibrium of this game.
We want to emphasize that the primary research focus of
this paper is on the strategic implications of generic advertis- Proposition 1 presents a general market condition under
ing under inter-industry competitive framework. Therefore, which competing industries in equilibrium both invest posi-
we have assumed that at Stage 1, each industry can ensure tive amounts in generic advertising campaigns. Such condi-
that firms contribute to the industry generic campaign after tion is determined by three sets of parameters representing: (i)
the overall level is set by the industry or trade association. the two industries initial customer bases (ϕ , k); (ii) effect of
While we choose not to explicitly model the contribution generic advertising (αj, j = 1, 2), and (iii) effect of brand
mechanism here, existing literatures suggest that mechanisms advertising within an industry (θij, i, j = 1, 2). Note that the
such as provision point can help an industry reach a target effects of brand advertising within industry (θij) enters the
budget (e.g., [16, 27]).7 expression of equilibrium generic advertising (G*j ) in the form
θij θ3−i; j
of Y j ¼ 1− which is part of λ j defined in
ðθij þθ3−i; j Þ
6 2
Following the roof top example, this specification could imply that more and
more residences are adopting the solar panel rooftop as opposed to the tradi-
tional roofing material.
7
An industry can inform its members that no generic campaign will be con-
8
ducted at all if the total amount decided at Stage 1 is not met. The provision Derivations using the general margin parameter (sij) yields qualitatively con-
point literature shows that if an industry sets the provision point at the optimal sistent results, yet the mathematical expression is more complex which makes
level, it constitutes a credible threat leading to contribution. it challenging for readers to following the intuitions.
22 Cust. Need. and Solut. (2017) 4:18–27
ϕ = 0.2, k = 0.9, α2 = 5, s1 = 1, s2 = 1
G*1 G*2 B*11 B*12
Industry 1’s generic ad effect Brand ad effects within industry 1 Brand ad effects within industry 2
α1 = 2 θ11 = 0.9 θ21 = 0.1 θ12 = 0.9 θ22 = 0.1 0.0857 0.1497 0.0129 0.0324
α1 = 5 θ11 = 0.9 θ21 = 0.1 θ12 = 0.9 θ22 = 0.1 0.1875 0.1915 0.0225 0.0225
α1 = 10 θ11 = 0.9 θ21 = 0.1 θ12 = 0.9 θ22 = 0.1 0.1822 0.1662 0.0300 0.0150
α1 = 15 θ11 = 0.9 θ21 = 0.1 θ12 = 0.9 θ22 = 0.1 0.1573 0.1346 0.0338 0.0113
α1 = 2 θ11 = 0.5 θ21 = 0.5 θ12 = 0.9 θ22 = 0.1 0.0398 0.1337 0.0310 0.0338
α1 = 5 θ11 = 0.5 θ21 = 0.5 θ12 = 0.9 θ22 = 0.1 0.1458 0.1894 0.0565 0.0247
α1 = 10 θ11 = 0.5 θ21 = 0.5 θ12 = 0.9 θ22 = 0.1 0.1563 0.1779 0.0778 0.0170
α1 = 15 θ11 = 0.5 θ21 = 0.5 θ12 = 0.9 θ22 = 0.1 0.1405 0.1506 0.0890 0.0130
α1 = 2 θ11 = 0.9 θ21 = 0.1 θ12 = 0.5 θ22 = 0.5 0.1002 0.1290 0.0147 0.0842
α1 = 5 θ11 = 0.9 θ21 = 0.1 θ12 = 0.5 θ22 = 0.5 0.1854 0.1498 0.0247 0.0565
α1 = 10 θ11 = 0.9 θ21 = 0.1 θ12 = 0.5 θ22 = 0.5 0.1681 0.1190 0.0319 0.0365
α1 = 15 θ11 = 0.9 θ21 = 0.1 θ12 = 0.5 θ22 = 0.5 0.1405 0.0908 0.0353 0.0269
α1 = 2 θ11 = 0.5 θ21 = 0.5 θ12 = 0.5 θ22 = 0.5 0.0531 0.1171 0.0357 0.0893
α1 = 5 θ11 = 0.5 θ21 = 0.5 θ12 = 0.5 θ22 = 0.5 0.1475 0.1515 0.0625 0.0625
α1 = 10 θ11 = 0.5 θ21 = 0.5 θ12 = 0.5 θ22 = 0.5 0.1467 0.1307 0.0833 0.0417
α1 = 15 θ11 = 0.5 θ21 = 0.5 θ12 = 0.5 θ22 = 0.5 0.1273 0.1046 0.0938 0.0313
(Note: Numerical results for other scenarios of k ∈ (0 , 1] are available upon request, which suggest the similar pattern with regard to the impact of key
parameters on the equilibrium generic and brand advertising levels as discussed in Sections 4.2 and 4.3)
starting/declining stage normally characterized by relatively generic advertising and firm’s brand advertising which are
small customer base. confirmed through analytical derivations. We highlight the
main findings in the following two Corollaries.
Fig. 4 Impact of brand asymmetry on industry generic advertising. (Plot 5 Conclusion and Disscussions for Future Research
generated by assigning numerical values ϕ = 0.2 , α1 = 2 , α2 = 5 and
θ11 = 0.9 , θ12 = 0.9 , θ22 = 0.1. Graphic pattern persists under different
numerical parameter values) This paper adds to the advertising literature by introducing the
inter-industry competitive framework to study the interaction
between two characteristically different advertising strategies,
i.e., generic versus brand advertising. Firms within an industry
brand advertising decisions of all firms across the two indus- compete for market share using brand advertising and indus-
tries. Through backward induction, we derive the equilibrium tries compete for market size using generic advertising.
generic advertising levels for the two industries and compare Introducing this higher degree of complexity allows us to
them with those in the main model. Results of the comparison make the following theoretical contributions. Firstly, we iden-
are summarized in the following Proposition 2. tify specific market conditions under which the presence of a
rival industry triggers an industry to conduct generic advertis-
Proposition 2 With Yj defined in Proposition 1, ing. Secondly, we demonstrated and discussed strong connec-
tion between generic advertising for inter-industry competi-
1) The second mover (i.e., industry 2) in equilibrium spends tion and brand advertising for intra-industry competition. If
less on the industry generic campaign than in the main the generic advertising conducted by an industry is more ef-
model of simultaneous generic advertising decision; fective, greater brand advertising will result within that same
2) The first mover (i.e., industry 1) in equilibrium spends industry. If the generic advertising conducted by a rival indus-
more on the industry generic campaign than in the main try is more effective, firms of the focal industry will reduce
model of simultaneous decisions in generic advertising spending on the brand advertising and likely shift to contrib-
ifα1 > YY 21 ss21 α2 . uting to industry generic campaign. An industry will be more
likely to conduct generic advertising if the firms within that
industry differ considerably in their brand advertising effec-
8 tiveness which in turn suggests asymmetric distribution of
6 market shares. Finally, we find that the equilibrium generic
and track the degree of asymmetry within industry using generic and brand advertising with the other three P’s of mar-
criteria such as brand advertising effectiveness or market share keting mix (e.g., product differentiation, pricing [26], and
distribution. On the other hand, if managers are contributing to channel distribution) is another important and interesting area
a generic campaign that is effective in expanding market size for future research. Finally, a thoroughly designed lab study11
for the industry, they must fully expect the brand advertising could provide behavior verifications for the predictions of our
intensity to go up. And it could also be strategically beneficial analytical model as well as testing the contribution mechanism
for an industry or trade associations to engage in generic ad- that ensures individual firms’ contributions to the industry
vertising earlier than its rival industry. generic campaign under inter-industry competition.
Our model has a few limitations which also point to the
directions of future research. First, this is a static model in the
spirit of Krishnamurthy [15, 16]. We have investigated one
variation where industries make generic advertising decisions
sequentially. A fully dynamic version of this model is definite- Appendix
ly an interesting direction to pursue. Second, like the existing
studies on generic advertising (e.g., [2]), we have not included Proof of Proposition 1
price as a decision variable. This is a reasonable assumption
since our focus is on inter-industry competition, and price is Based on the model setup in Section 3, profit function for firm
unlikely to affect market size dynamics. The interaction of i of industry j can be written as:
1=2
!
M oj þ α j G j θij Bij
πij ¼ S j −Gij −Bij ; with i; j ¼ 1; 2: ð6Þ
M 0 j þ α j G j þ M 0;3− j þ α3− j G3− j 1=2
θij Bij þ θ3−i; j B3−i; j
1=2
Equilibrium brand advertising levels at the second stage of industry does not conduct any generic advertising. Suppose
the game are obtained from simultaneously solving the first- G2 = 0, straight profit maximization yields the optimal generic
∂πij
order conditions ∂Bij ¼ 0 and check the negative semi- advertising level for industry 1 as:
definiteness of the second-order Hessian matrix, which results
in pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
0 kϕα1 Y 1 s1 −ð1 þ k Þϕ
G1 ¼ ð9Þ
θij θ3−i; j M j s j α1
Bij ¼ 2 ð7Þ
2 θij þ θ3−i; j
and the above (9) is positive as long as 0 < ϕ < kα1 Y 1 s1
ð1þk Þ2
(con-
Substitute M01 = ϕ, M02 = kϕ and (7) back to (6), sum over dition ii). Similarly, we can derive the condition for G2 > 0 if
0