Marketing Spending Models - Innovation and Marketing in the Pharmaceutical Industry - page 560

physician meetings or postmarketing surveillance studies, which are presumably
subsumed among “other marketing expenditures,” however, are still rare.
The vast majority of models are descriptive (85 %) and based on aggregate data
(67 %). The frequent use of aggregate data is clearly driven by the greater access to
this data. Note that in countries such as Germany, it is even not allowed for com-
mercial vendors to collect and sell individual physician or patient data. Individual-
level data, however, are necessary to study important questions such as the flow of
information among a social network of physicians (Nair et al. 2010 ) or the alloca-
tion of resources across physicians (e.g., Montoya et al. 2010 ).
Table 19.2 reveals that several models (36 %) consider a possible interaction
among decision variables, e.g., marketing spending categories (e.g., Wittink 2002 ),
other mix variables such as price (e.g., Narayanan et al. 2004 ) and quality (e.g.,
Venkataraman and Stremersch 2007 ), and strategic decision variables such as time
to market (Fischer et al. 2005 ). In contrast, interactions among competitors and
across countries are each incorporated in only three studies out of 39 (8 %). Given
the large product variety in many categories and the global nature of the business,
these issues warrant greater attention in model building.

19.3.2 Generalizations About the Responsiveness

of Pharmaceutical Demand

Given the rich empirical research on marketing spending models for pharmaceuti-
cals, a natural question arises whether we can assert any generalizations about the
responsiveness of drug demand to marketing efforts. Three recent meta-analyses
approached this question. Kremer et al. ( 2008 ) summarize the findings of 58 studies
that report 781 elasticities of various pharmaceutical spend categories. Sridhar et al.
( 2014 , in this topic) use 373 elasticities from 48 pharmaceutical studies. The meta-
analysis by Albers et al. ( 2010 ) focuses on personal selling that includes other
industries, as well. All three studies provide a strong generalized result that detail-
ing is a very effective marketing element. Kremer et al. ( 2008 ) report an average
elasticity of 0.326, which appears to be a bit higher than the average elasticity of
0.210 in Sridhar et al. ( 2012 ). An elasticity of 0.210 means that sales rises by 21 %
if detailing expenditures increase by 100 %. As these authors note, the difference
may be due to the fact that they only consider current-period brand-level elasticities
whereas the finding in Kremer et al. ( 2008 ) also includes long-term and category-
level elasticities.
The study by Kremer et al. also provides insights into the demand responsiveness
with respect to other spend categories that are generally lower compared with detail-
ing. These elasticities are 0.123 for professional journal advertising, 0.073 for
DTCA, and 0.062 for other physician-oriented spending categories. Interestingly,
the responsiveness of detailing is, on average, 4.5 times higher than that for DTCA.
Expenditures in the USA, however, have been only 1.45 times larger, on average,
over the last years (see Fig. 19.2 again). An easy explanation for this striking