Pharmaceuticals companies using celebrity direct-to-customer (DTC) campaigns need an analytics system of computing potential ROI based on the right set of variables.
The Work Ahead in Intelligent Automation: Coping with Complexity in a Post-Pa...
Pervasive Analytics: Measuring the Impact of a Celebrity Direct-to-Customer Campaign in Pharma
1. Pervasive Analytics:
Measuring the Impact of a Celebrity
Direct-to-Customer Campaign in Pharma
While celebrity endorsements in direct-to-customer campaigns
can enhance consumer appeal and increase brand recall amid the
cacophony of a crowded marketplace, such initiatives require life
sciences organizations to choose wisely and work methodically to
measure the true impact of these initiatives.
Executive Summary
In the highly competitive pharmaceuticals market
where competitors often develop and launch sim-
ilar products, brand marketing plays a critical role
in differentiating comparable offerings.
Unlike traditional promotional techniques such
as detailing and sampling that are directed pri-
marily at physicians, pharmaceuticals companies
are now focusing on more cost-effective promo-
tional channels to bring awareness and influence
end customers — patients, as well as physicians.
Today’s consumers are more actively involved in
managing their wellness, aggressively seeking
information about the drugs and therapies that
may affect them. Direct-to-customer (DTC) adver-
tising, therefore, has become the primary focus
area for many companies.
DTC advertising is defined as the effort made by
a company via mass communications to directly
persuade a consumer to choose and purchase
its product. DTC started with newspapers, and
later radio and television, and has progressed
to websites, videos, social networks, etc. For the
pharmaceuticals industry, where a consumer
needs a prescription from a physician to obtain
prescription drugs, DTC advertisements play
a significant role in enhancing the knowledge
patients have about their diseases and therapies.
The goal is to persuade patients to consult with
their physicians about the advertised drug, thus
influencing their prescribing patterns.
According to the U.S. Congressional Budget Office
(CBO), from 2006 to 2008 pharmaceuticals
manufacturers spent an average of $71 million
per drug for certain drugs on DTC advertising
during the first two years following these drugs’
FDA approval.1
Top-25 brands contribute to more
than 60% of total DTC pharmaceuticals spend-
ing in the U.S., according to the CBO. However, as
more blockbuster drugs lose patent protection,
and amid thin R&D pipelines, pharmaceuticals
companies are working extra hard to ensure the
ads have extended impact on consumer mind-
sets. Many companies are therefore relying on
celebrities who can immediately jumpstart sales.
• Cognizant 20-20 Insights
cognizant 20-20 insights | september 2015
2. However, involving a celebrity is not easy. Millions
of dollars are reportedly spent on celebrity cam-
paigns and a celebrity’s image directly impacts
the brand image. This white paper highlights key
elements of a celebrity campaign and outlines a
robust approach to measure the true impact of
these initiatives.
Celebrities in Direct-to-Customer
Campaigns
Joan Lunden offered the first celebrity endorse-
ment for the drug Claritin in 1998. In 2000, Wyeth
hired supermodel Lauren Hutton to hawk hor-
mone therapy for menopause. Paula Deen and
her sons, Bobby and Jamie Deen, teamed up
with Novo Nordisk on their diabetes drug, Victoza,
in 2012.
While there is some anecdotal evidence that
suggest celebrity advertising successfully gen-
erates a multifold impact on sales and enhances
message memorability, it doesn’t mean that
all celebrity ads are highly successful; there
may be negative effects that can erode brand
sales. Negative effects can occur as a result of
a decline in the celebrity’s professional image/
career, private life incidences or when the celebri-
ty endorses multiple competitive products, which
can undermine consumer confidence in the integ-
rity of the message.
World-champion golfer, Eldrick “Tiger” Woods, for
example, was once considered the most success-
ful celebrity endorser of all time, hawking brands
such as Accenture, Gatorade, Gillette, AT&T,
NetJets, EA Sports, Tag Heuer, Nike, etc. However,
negative publicity following disclosures about
Woods’ personal life damaged his reputation,
which resulted in a major loss of sponsorships.
Finding the right celebrity who speaks to the
core brand proposition is hard; ensuring that the
brand isn’t subsumed by the celebrity is exceed-
ingly critical. Companies must understand this
before deciding whether or which celebrity to
sign on to represent their brands.
Common questions that should be considered
include:
• What is the return on investment for this
celebrity endorsement?
• Which segment is most influenced by this
celebrity campaign?
• Which media type (TV, print, online, etc.) is
most impactful?
• What is the share of free media publicity vs.
paid media?
We have developed a robust, four-step approach
to help pharmaceuticals companies measure the
true impact of a celebrity campaign. Our method
is built on a highly qualitative
approach, and depends heav-
ily on the business judgments
of a very experienced group.
Measuring the True
Impact of Celebrity
Campaigns
When a popular celebrity is
associated with a brand, there
is a greater chance of free
publicity in various media
channels that is spawned by
the paid endorsement. Again, the nature of pub-
licity can be positive, neutral or at times negative.
It is essential to account for all sentiments and
arrive at a true impact assessment (see Figure 1,
next page).
Data Exploration and Validation
The success of any analytics initiative is a func-
tion of data quality as well as analytical maturity
and data consistency. Data collected from mul-
tiple sources such as historical sales, detailing,
sampling, celebrity campaigns, demographics,
financial and other market events is scrutinized
for missing records, outliers and redundancy
using techniques like Univariate, Histogram and
Box Plot. Appropriate treatment techniques are
followed to fix missing and outlier records or, in
some cases, exclude them when there is a low
concentration of data. Data consistency is also
ensured by a rigorous distribution and trend anal-
ysis on the multiple data sets. Multiple data sets
are then integrated and further divided into train-
ing and validation data sets.
This phase confirms that the data collected from
multiple sources is sufficient and consistent and
that the data trends are making proper business
sense.
Variables Creation and Segmentation
This phase focuses on the creation of multifold
independent variables that best explain the vari-
ation of the brand sales. Independent variables
are created for paid and free impressions for the
celebrity campaign along with other promotion-
al activity to capture the respective impact on
the brand sales. Both current and lag variables
cognizant 20-20 insights 2
Finding the right
celebrity who
speaks to the core
brand proposition
is hard; ensuring
that the brand
isn’t subsumed
by the celebrity is
exceedingly critical.
3. cognizant 20-20 insights 3
are created for these promotions and sales data
to capture the decay effect and the carry-over
effect. These exhaustive independent variables
will help in separating the true impact of the cam-
paign on the brand sales from other significant
promotions.
The customers in the analysis universe are seg-
mented based on factors such as geography,
prescribing behavior and demographic data, to
see how the celebrity campaign impacts seg-
ments differentially. This will also help in targeting
the right customer with the right media type such
as TV, print, social media, etc.
Model Design and Finalization
Various customer segments may behave dif-
ferently when exposed to a particular media
promotion. A time series regression model is
established at a segment level to access the true
picture of the celebrity promotion. Brand sales
are taken as the dependent variable; independent
variables comprise current and lag values of the
paid and the free media promotion along with
other significant promotions and market events.
Multiple iterations based on the significance of
the independent variables, using statistical out-
puts such as p-value and model goodness of fit,
help in building a robust model. Model effective-
ness is further checked with validation data sets
to ensure a full proof model.
ROI Evaluation
The return on investment (ROI) for a celebrity
DTC campaign is evaluated using the estimates
of the finalized model parameters. The segment
level model estimates corresponding to the
celebrity campaign variables are utilized to evalu-
ate the incremental sales impact for a particular
segment, which also is used to account for the
carry-over effect of the campaign.
This long-term carry-over impact is then used to
evaluate the ROI for different segments. Our mea-
surement equation is:
ROI = (Long-term carry-over impact /
Cost of the celebrity campaign)
#
2
• Data sanity and consisten-
cy for sales and promo-
tional data like detailing,
samples, celebrity
promotion, etc.
• Distribution, correlation
and trend analysis.
• Creation of training and
validation data sets.
Data Exploration &
Validation
• Current and lag variable
creation for sales data.
• Current and lag variable
creation for paid and free
publicity of the celebrity
campaign.
• Current and lag variables
for all other promotions.
• Segmentation based on
geographic and market
potential.
Model Design &
Finalization
• Multiple variable regres-
sion model execution
using sales as the
dependent variable.
• Refinement using
statistical significance of
independent variables.
• Model validated using
R square, adjusted R
square and validation
data sets.
ROI Evaluation
• Using regression
coefficient of the paid
and free celebrity
promotion variable to
get the direct incre-
mental impact.
• Coefficient of the lag
variable of the sales
data considered for the
carry-over impact.
• Return on investment is
calculated as: long-term
carry-over impact/cost
of the celebrity
campaign.
Variable Creation &
Segmentation
Measuring a Celebrity Endorsement Initiative
Figure 1
4. Looking Forward
Celebrity endorsements can give a brand a touch
of glamour, provide added consumer appeal
and increase brand recall in a crowded market.
Celebrity campaigns across industries have suc-
cessfully generated multifold returns on the
investment. According to a databank of case stud-
ies maintained by the Institute of Practitioners
in Advertising (IPA), Bob Hoskins’ campaign for
British Telecom delivered £297 million in incre-
mental revenue resulting in a ROI of 6:1, and
TESCO ads featuring actresses Prunella Scales
and Jane Horrocks provided incremental revenue
of £2.2 billion equating to a ROI of 2.25:1. In one
of our recently concluded analyses on a celebrity
campaign for a pharmaceuticals company, a ROI
of 8.3:1 was achieved (see sidebar above).
Analytics can help companies measure the suc-
cess of a celebrity campaign. Analytics also can
help companies segment customers based on
their level of responsiveness to the celebrity cam-
paign, which can be used for future targeting.
Analytics can also help companies identify the
right budget across media channels to help maxi-
mize the impact of celebrity campaigns.
cognizant 20-20 insights 4
Quick Take
Evaluating the Impact of a Celebrity DTC Campaign on Brand Sales
Client Situation
A European life sciences major engaged us to
assess the performance of a direct-to-celebrity
campaign in which it spent $7.5 million to reach
52,803 physicians. The details:
• The client hired a celebrity chef for its campaign
on diabetes.
• The campaign plan:
>> To focus on brand prescribing.
>> Execute via geography and media channels
(TV/print/online).
Challenges
• Separating the effect of the campaign on brand
sales from other significant promotions such as
detailing, samples and coupons.
• Removing other extraneous effects, such as
other market events and managed care wins.
Solution
• We built a regression model with the following
key parameters:
>> New prescriptions for the brand.
>> Detailing.
>> Samples/coupons.
>> Media impressions.
>> Other market events, etc.
• Provided impact by both paid and earned media
impressions.*
• Highlighted the most effective promotion
channel.
Benefits
• Paid media return:
>> 58% of total.
>> Return/impression: $0.51.
• Earned media return:
>> 42% of total.
>> Return/impression: $0.38.
>> Total return: $64.5 million.
* Earned media impressions: Free publicity result-
ing from celebrity connection.