firm advertises a lot to its local customers (it actually chooses to reach all of them), it does
not advertise to the others who are certain to receive a competing ad. Prices are then
larger than if targeted advertising was infeasible. In contrast, for large advertising
costs, each firm advertises to both consumer segments but selects advertising intensities
below 1. Prices are then lower than they would be without targeted ads.
In this setting, if a firm drops its price, the set of consumers it wishes to target with a
higher advertising intensity is expanded. The jump in advertising intensity for the mar-
ginal consumers (those who would switch to the firm who has cut its price if perfectly
informed) contributes to make demand more price elastic. There is no such jump if the
firm advertises uniformly to all consumers.
56
The increased demand elasticity due to an
increase in advertising intensity with targeted advertising is reminiscent of the analysis of
Chamberlin (1933) on informative advertising that makes demand more elastic (see
Bagwell, 2007, p. 1709). However, this increased advertising intensity at the margin
in the case of targeting constitutes an additional marginal cost of increasing sales that
pushes prices upward. As a result of the two countervailing forces, prices are higher with
targeting if advertising costs are low,
57
but lower if advertising costs are high. If they are
sufficiently high, then targeted advertising reduces the profitability of the market.
In the analysis above, consumers benefit from targeting directly to the extent that they
are exposed to more ads pertaining to products they wish to purchase. As I just explained,
the indirect impact through price may go either way. Another potential benefit that is not
captured in the theoretical frameworks I have discussed so far is that consumers receive
less ads that are irrelevant. Provided that there is some nuisance cost associated with being
exposed to an ad, as is typically assumed in the media economics literature, such a reduc-
tion in overall ad quantity is beneficial to consumers.
Johnson (2013) proposes a frame-
work that incorporates this dimension as well as ad avoidance activity by consumers.
A parameter in his model measures the precision with which a firm is able to target
the consumers who are interested in its product. He finds that improved targeting has
an ambiguous impact on consumer welfare. Although it improves the relevance of
the ads they receive, it increases the firm’s incentives to advertise to more consumers,
which decreases welfare for those who get the least relevant ads.
In
Johnson (2013), advertisers are subject to a negative externality imparted by con-
sumer’s ad avoidance. But in practice they also exert some negative externality on each
other by sending out ads that compete for the consumer’s attention. Consumers are
exposed to many ads along with many other pieces of information to which they may
56
From first-order condition (4.14), a firm that drops its price actually finds it optimal to drop its uniform
advertising intensity.
57
For low enough ad costs, consumers are perfectly informed with uniform advertising and there is no
difference in the price elasticity of demand between the two regimes.
181
Advertising in Markets
not pay full attention. This creates information congestion, the topic that closes this
discussion of ad technology.
4.5.3 Information Congestion
Van Zandt (2004) and Anderson and De Palma (2009) propose some related models of
“information overload.” The general idea is that messages sent to the same recipient by
different sources may crowd each other out because the recipient can or wishes to only
pay limited attention to the information she receives. As a very direct illustration,
Anderson and De Palma (2009, p. 688) quote a figure measuring that 46% of the bulk
mail in the US remains unopened. Whether because of some exogenous inability to
process a large amount of information or because processing it is too costly, consumer
attention is scarce and advertisers compete for that attention.
Anderson and De Palma (2012) use an ingenious construction to analyze how this
competition for attention among firms selling in different markets interacts with price
competition among firms in the same market. They consider multiple markets for homo-
geneous products. Competition in each market is monopolistic competition, similar to
Butters (1977): (i) a continuum of firms send ads with a price offer; (ii) consumers are
identical with unit demand for each product and a consumer can buy from a firm only
if she has received an ad from that firm; (iii) each ad must bring in zero profit because of
free entry. By contrast with
Butters (1977), an ad reaches all consumers but it is noticed by
each consumer with some probability that is less than one so that each consumer only
notices a fraction of the ads sent.
58
Furthermore, because of information congestion,
the probability that a consumer notices a particular ad in one of the markets is decreasing
in the total number of ads addressed to the consumer from all firms in all markets. Markets
differ as to the social surplus generated by a sale. As in Butters, equilibrium in each market
is characterized by a price distribution: each ad trades off a higher probability of a sale
achieved by a lower price with the higher revenue per sale afforded by a higher price.
Because without congestion the advertising intensity would be socially optimal as is
the case in Butters (see
Section 4.2.4), here advertising is excessive: each firm sending
an ad does not take into account the added congestion it creates for all the other ads sent.
The model uncovers a rich pattern of externalities across sectors. If the number of
active markets increases, this increases information congestion, which in turn increases
prices in the pre-existing markets. The share of all those markets in total advertising goes
down and the absolute advertising level decreases for the weaker markets (in terms of
potential social surplus per sale). The impact of an increased consumer attention on over-
all advertising intensity is non-monotonic. If attention is low, few ads are sent because
they are very unlikely to make a sale. If attention is very high, competition is very intense
because each consumer gets to compare a lot of prices and hence advertising intensity is
58
In Butters, each ad reaches only one consumer so each consumer observes only a subset of all the ads sent.
182
Handbook of Media Economics
low again due to the limited profitability of a sale. It is only for intermediate levels of
attention that sending an ad is very profitable so that advertising is very intense. Regard-
ing the distribution of advertising activity across markets, it is very heterogeneous if con-
sumer attention is low, with strong sectors advertising the most. Improved consumer
attention intensifies competition in all sectors, which equalizes profit between stronger
and weaker markets. This results in a more uniform advertising intensity across markets.
4.5.4 Concluding Remarks
The theoretical literature on advertising reach typically assumes decreasing returns to
advertising expenditure, which ensures a well-behaved optimization problem for the
firms.
Butters (1977) has shown that this is consistent with an advertising technology that
reaches consumers randomly. There is also some empirical support for such a specifica-
tion of advertising costs. Yet there may be good reasons to think that returns to adver-
tising expenditures can be increasing, in particular when advertising is only effective if
consumers are reached by more than one ad message, or if a massive advertising reach
can be achieved through such mass media as television.
A proper analysis of advertising costs actually requires taking into account how pricing
decisions of media firms are determined. One important issue is whether there is some
form of price discrimination involving quantity discounts. Besides, as I already alluded to
in the concluding remarks of
Section 4.2, the pricing of ad space has potentially major
implications for the welfare analysis of advertising intensity. In other words, it is necessary
to know whether the price charged for ad space is above or below its social cost. For
instance, as explained by
Anderson and Gabszewicz (2006, p. 294), if media platforms
charge both sides (advertisers and the audience) then the price per ad unambiguously
exceeds the social cost of ads, which is the nuisance cost for the platform’s audience.
59
Then the social cost of advertising is actually lower than the private cost incurred by
advertisers, which pulls the market outcome in the direction of an insufficient provision
of ads.
But the welfare analysis of advertising must also account for the nature of the infor-
mation provided by the ad. Indeed, the ad cost may depend on which information is
provided. For instance, more ad space may be needed to convey the desired information
either because the consumer should be exposed repeatedly (as is the case for signaling the
high quality of an experience good) or because conveying a lot of information requires
more space. Second, the benefit of advertising additional pieces of information (if there is
any) may vary greatly depending on the nature of the information, as illustrated by the
analysis in
Section 4.3.
59
This is assuming that the entire potential audience is served, with single homes and keeping the media
content exogenous.
183
Advertising in Markets
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.