Understanding Sample Usage and Sampling as a Promotion Tool: State of Industry Practice and Current Research - Innovation and Marketing in the Pharmaceutical Industry - page 516

used in the ROI Analysis of Pharmaceutical Promotion (RAPP) by Neslin ( 2001 )
and Wittink ( 2002 ). Pooling multi-brand data increases sample size and expands the
range of hypotheses to be tested. This type of analysis can provide directional views
to a particular company, because it is based on cross-brand experience.
From the practice perspective, the advantages of aggregate sample modeling
include:

- Sample effectiveness can be compared with both personal promotion, such as

detailing and nonpersonal promotion, and other marketing instruments, such as
journal spending and direct to consumer spending (DTC).

- Easier to evaluate competitive sample drop effects.

- Can provide life cycle perspective because of historical and cross-brand

perspective.

- Dynamic effect and seasonality can be easily specifi ed and estimated.

- ROI assessment at brand level can be easily conducted.

Limitations of aggregate sample modeling are:

- Longer historical data is needed for modeling so the result can be useful for long-

term strategic planning but may not be accurate for short-term implementation.

- May be subjected to aggregation errors.

- Limited by sample size or data period.

- Multi-collinearity between detailing and sample drop.

- Not tactical for fi eld implementation and action, for example, at segment level.

17.5.2 Physician Level Modeling

Test-control study is a commonly used approach in the industry to evaluate impacts
of promotion such as detailing, new messaging, DTC, direct mail, and sampling at
individual physician level. Typically, the response measure is prescription shares by
a physician. The test and control groups are selected in such a way that across groups
the physicians are similar in prescribing patterns (in terms of volume) and external
factors. ANOVA or ANCOVA model is applied to test prescribing share or volume
difference between test (sample exposed) and control (no sample exposure) groups
while other factors are controlled through either covariates or data selection. The
same approach is also used for pre- and post-analysis, which requires longer duration
of the data. The ROI can be easily calculated based on test and control analysis.
Normally, this type of approach is used on secondary data, such as physician
prescription data or longitudinal patient data. A large sample size can ensure that a
suffi cient sample can be obtained with more factors controlled. Different companies
used different matching methods to form test and control groups that are compara-
ble. One of the main limitations of the test and control analysis is that it is diffi cult
to eliminate all confounding factors so that physicians in test and control groups are
exactly comparable except for differences in sample or other promotion stimuli.