EFFECTIVENESS OF COMBINING PRINT AND ONLINE

ADVERTISEMENTS

DISSERTATION FOR MA PART II EXAMINATION

SESSION 2014-2016

RESHMI ROY
ROLL – 92/JMC/150021

REGISTRATION NO -014-1221-0535-11

DEPARTMENT OF JOURNALISM AND MASS COMMUNICATION

EFFECTIVENESS OF COMBINING PRINT AND ONLINE
ADVERTISEMENTS
 ACKNOWLEDGEMENT

I take this opportunity to express my profound gratitude and deep regards to my guide Dr

Tapati Basu, Department of Journalism and Mass Communication, University of Calcutta for her

exemplary guidance, monitoring and constant encouragement throughout the course of this

thesis. Her blessings help and guidance time to time helped me to complete my work.

Last but not the least, I thank Almighty, my parents and friends for their constant

encouragement without which my work would not be possible.

 CONTENTS

TOPICS PAGE NO

1. INTRODUCTION 1

2. PURPOSE OF THE STUDY 4

3. CRITICAL APPRAISAL OF STUDY 5

4. OBJECTIVE OF STUDY 14

5. STATEMENT OF PROBLEM 15

6. SCOPE OF INVESTIGATION 16

7. HYPOTHESIS 1 (Hypothesis, method and sources) 17

8. HYPOTHESIS 2 (Hypothesis, method and sources) 30

9. CONTENT ANALYSIS 46

10. DATA COLLECTION AND DATA ANALYSIS 58

11. SURVEY SAMPLE QUESTIONS 61

12. INTERVIEWS 64

13. CONCLUSION 69

14. BIBLIOGRAPHY AND REFERENCES 70

 INTRODUCTION

To understand the phenomenon of this combination (online and print), first we need to study

and learn the both aspects individually and what are the merits of them and what are the

demerits of them. Here, we need to specially note and criticize the demerits that are the

disadvantages of using these two techniques individually so that the emergence of combination

of these two techniques took place.Cross-media marketing is a form of cross-promotion in

which promotional companies commit to surpassing traditional advertisement techniques and

decide to include extra appeals to the products they offer. The material can be communicated

by any mass media such as e-mails, letters, web pages, or other recruiting sources. Cross-media

advertising is a strategy used by business owners to market a business using various types of

media. Owners may use all of media types individually or combine several mediums to create a

cohesive marketing campaign. Cross-media advertising is one often overlooked by business

owners, especially Internet entrepreneurs, according to Web Pro News. Keep the target

consumer profile in mind when choosing media for a given business to reach maximum profits.

Marketers have increasingly made online marketing an integral part of their multi-channel

communications strategies. There are certainly advantages to using online marketing including

economies of scale, direct fulfilment, exceptional targeted advertising capabilities, and of

course the ability to track advertising. There has been growing interest in internet advertising,

documented for example by the fall 2002 Special Issue on Advertising and the New Media of

the Journal of Advertising. As part of this research, the effectiveness of internet advertising

such as banner exposures has been shown strikingly by in the IAB’s study conducted in

cooperation with Millward Brown, (IAB, 1997) and Briggs and Hollis (1997). In comparisons with

~1~
the effectiveness of print advertising, however, results have not been unequivocal. According to

Sundar et al. (1998) the print medium is superior to the online medium in memory measures,

ascribing the effect to the novelty of online advertisement, the delivery mechanism and users’

expectations. On the other hand Gallagher, Foster, and Parsons (2001) and Gallagher, Parsons,

and Foster (2001) use advertising hyper-linked to a related text and state that both media are

equally effective. While web advertisement is often used for brand building and has also proven

effective for that purpose (IAB 1997, Li, Daugherty, Biocca 2002; Gong, Maddox 2003, Dahlen,

Rasch and Rosengren 2003), a comparison with print advertising in this dimension is still

missing. Additionally, since online advertising is not novel any more, the question arises, how

banner exposures affect experienced web users, without the additional effect of hyper-links in

the content. Furthermore, both studies used few (two and four) pages or screens with content

and advertising, whereas a typical surfing or reading situation includes reading a number of

articles. Since print products and the internet differ, the former being a linear-sequential, the

latter a hierarchical medium (Bezjian-Avery, Calder, Iacobucci 1997), this might also account for

differences in advertising effectiveness not looked at yet. Magazine advertising sales people

have a big advantage when selling digital ads to their clients. When a national magazine

campaign is planned in 2014 and beyond, there is almost certainly a digital campaign coming

down the pike behind it. Digital planning comes later because the lead-time to get it into media

is far shorter; 90 days for monthly magazines compared to a week or even less for digital. When

your magazine wins the business, that win should give your sales person a head-start and an

advantage to sell the Internet campaign onto your site.

You and your team have an advantage for two reasons:

~2~
 Your non-print competitors may not even know that a campaign is being planned. But

your team can assume — until proven otherwise — that a digital plan is in the works.

 Research shows buying media from the same brand online and off is more effective.

Provided your sales team can communicate this research, there should be a bias toward

adding digital impressions to the print impressions directed to your audience.

~3~
 PURPOSE OF THE STUDY

Due to the growing importance of the internet, marketers increasingly are pursuing an

integrated multi-channel communication strategy to increase to increase advertising

effectiveness. Despite the high relevance of this topic, however, only scarce literature

comparing different media channels and investigating the effectiveness of cross media. Cross

media advertisements can be sound. To fill this gap, my experimental study addresses the

research question whether the combination of online and print advertising generates a better

result in advertising effectiveness than each two medium working separately. Advertising in a

newspaper has been a well established practice more than 100 years, whereas online

advertising activity started in 1994. Some 22 years later digital advertising is now an integral

part of the global advertising market. The purpose of cross media advertising is to reach a large

audience and customer base. If the target customer profile is broad and covers an age range

between 18 years of age and 65, the company may use various types of media to reach all

customers. While the elderly may read the newspaper or watch television to see advertising,

the young generation may use websites and applications on smartphone to learn about the

new products and services.

According to my research, I want to critically appraise the merits and demerits of using the

combination (cross media). The research is bending towards the merits aspect rather than the

demerits because combination of online and print media is much more effective than the both

two functioning individually.

~4~
 CRITICAL APPRAISAL OF STUDY

Cross media advertising is a strategy used by business owners to market a business using

various types of media. Owners mainly uses the two kind of medias, one of the print

advertisements and other the digital or online advertisements. Cross media advertising is one

often overlooked by business owners, especially internet entrepreneurs, according to Web Pro

News. Keep the target consumer profile in mind when choosing media for a given business to

reach maximum profits.

There are certain questions which makes the critical appraisal of the research more effective:

The questions and the justifications are as follows:

Is the Study's Research Question Relevant?

Definitely, the relevancy is up to the maximum as the topic satisfies the test and gives positives

results and thus it is far better than the both two advertising weaponed functioning

individually.

~5~
Does the Study Add Anything New?

My research work did not ended up to the effectiveness area but also shown a light to a very

effective medium which can be a tool of cross media effectiveness that is Cross-media

campaigns.

What is Cross media campaign?

With the new product introduction, it is always a challenge for a company to plan an effective

multimedia mix. Due to the increasing emphasis on integrated communication strategies and

the important role of the internet, however, it is of particularly interesting to combine online

and print advertising. It is the values of advertising effectiveness.

When analyzing the effectiveness of cross media campaign, scientific studies state the

combination of media channels shows higher effectiveness than the internet functioning solely.

How the productivity of advertisements increases if it’s combined?

The productivity certainly increases because it effects to a large number of customers, number

of response is more, number of feedback is more, highest return rate and personalized touch to

the customers.

What are the outcomes of effectiveness of advertisements measurements?

1. Total brand recall

~6~
 2. Unaided brand recall

3. Product category aided brand recall

4. Brand aided advertisement message recall

5. Recognition of displayed advertisements.

What is application of cross media and its effects?

In practice, the application of cross media can lead to a better result than single media

campaigns. Due to cost efficiency considerations, cross-media advertising may maximize

productivity of advertising expenditures.

What Type of Research Question Does the Study Pose?

The most fundamental task of critical appraisal is to identify the specific research question that

an article addresses, as this process will determine the optimal study design and have a major

bearing on the importance and relevance of the findings. A well-developed research question

usually identifies three components: the effectiveness, the barriers which occur frequently and

the outcomes of interest. In general, research questions fall into two distinct categories, below.

Questions about the Effectiveness: The effectiveness is above 65% positive than the other.

Thus the combined weapon of advertising is much more effective rather is it a high expensive

medium than the other.

Questions about the Frequency of Events: Frequency of events carries nothing but the

obstacles and hurdles to get the best results of the effectiveness. Certain barriers consist such

~7~
as money factor and labour factor.

What are the advantages and disadvantages of using combined online and print media

advertisements?

Advantages

1. Personalized touch to the customers

2. Multiple opportunities to reach the audience.

3. Guaranteed returns on ROI

4. Greater response rates and interaction

Disadvantages

1. Requirement of large scale of money/funding/large infrastructure/excellence in operations

with zero error.

2. Complexity in its nature

3. Time consuming

4. Labour intensive

Was the Study Design Appropriate for the Research Question?

Yes the study design was appropriate as a in the test of hypothesis a Austrian magazine was

~8~
taken for the research study which gave a positive result proving the effectiveness of cross

media advertisement.

Economic Evaluations

Economic-evaluation studies focus on cost-efficiency, or which effectiveness can provide the

greatest benefit for the least cost. Several types of economic-evaluation studies exist, including

cost-benefit, cost-effectiveness and cost-utility analyses, all of which differ in how they

measure health benefits. An important feature of critical appraisal of any cost analysis is an

assessment of how well the various cost and consequences of individual effectiveness have

been identified and measured. The CASP has developed a checklist to aid with the appraisal of

economic evaluation studies.

Do the Data Justify the Conclusions?

The next consideration is whether the conclusions that the authors present are reasonable on

the basis of the accumulated data. Yes the data has justified the conclusion rather there was

some frequent barriers. Not only it satisfied the results, it also gave rise to a new additional

phenomenon known as cross media campaign.

Are There any Conflicts of Interest?

Conflicts of interest occur when personal factors have the potential to influence professional

~9~
roles or responsibilities. Identification of a potential conflict of interest is not synonymous with

having an actual conflict of interest or poor research practice. Potential conflicts of interest are

extremely common, and the most important questions are whether they have been recognized

and how they have been dealt with. A main mechanism for dealing with potential conflicts of

interest is open disclosure. In the process of critically appraising a research article, one

important step is to check for a declaration about the source of funding for the study and, if a

potential conflict of interest had been identified for a statement about how this conflict was

managed. For example, the researchers might state specifically that the sponsoring agency had

no input into the research protocol, data analysis or interpretation of the findings. Many

journals now routinely require authors to declare any potential financial or other conflicts of

interest when an article is submitted. Though there are conflicts of interest lying the research

the investigation proved and gave a satisfying result of proving the combined medium to be the

better it its own way. Conflicts of interest which took place in this research are few incorrect

data findings and later it was corrected. No other potential threats were discovered.

~ 10 ~
Does the Study Test a Stated Hypothesis?

Yes the study was tested hypotheses, and the 3 tests were taken in the research

1. Online vs Print

2. Print vs Combined

3. Online vs Combined

In the last 2 test which was our primary test, it gave results that the combine method is slightly

more effective than the other two functioning solely. The hypotheses tested was null thus the

combined medium is effective than the online and print respectively.

Did the Study Methods Address the Key Potential Sources of Bias?

No key potential sources of bias were detected in the investigation. Potential sources of bias

are nothing but random errors which occurred in the research which may inclined the

statement to be a biased statement proving both the medium(combined and print & online) are

equally effective in its own, but as in this research we have proved that the combined is

effective with no potential sources of bias. The method which proved that the key potential

sources of bias are not detected are the Critical Appraisal Skills Programme (CASP) includes

such tools and the program coordinators have developed separate checklists for the appraisal

of systematic reviews, RCTs, cohort studies, case-control studies, diagnostic test studies.

The CASP which are taken are as follows

~ 11 ~
Systematic Reviews and Meta-analyses

A meticulous, standardized protocol is used in a systematic review to identify, critically appraise

and synthesize all the relevant studies on a particular topic. Some systematic reviews may then

proceed to a meta-analysis, in which the results from individual studies are combined statistically

to produce a single pooled result. Although planning to undertake a systematic review or a meta-

analysis prospectively is possible, the majority of these types of article are retrospective and a

risk of bias exists, which arises from the selection of studies and the quality of these primary

sources. Publication bias, which results from the selective publication of studies with positive

findings, is of particular concern, as it distorts overall perceptions of the findings on a particular

topic.

The QUORUM (Quality of Reporting of Meta-Analyses) statement provides a comprehensive

framework for assessments of the quality of reporting in meta-analyses and systematic reviews.

Randomized Controlled Trials

In an RCT, the random allocation of participants should ensure that advertisement mediums

groups are equivalent in terms of both known and unknown confounding factors; any differences

in outcomes between groups can, therefore, be ascribed to the effectiveness. Study design alone,

however, will not guard against bias if crucial aspects of the study protocol are suboptimal. The

potential for selective enrolment of patients into the study can be one an important source of bias

if the group to which individuals will be allocated is known or can be guessed. Centralized

~ 12 ~
methods of randomization, for example a computer-generated allocation, are preferable to less

concealed methods, such as use of colour-coded forms or pseudo-random sequences based on

medical record numbers or days of the week. Failure to conceal the allocation sequence has been

shown to result in a greater distortion of the results than lack of double-blinding -- another major

source of bias in RCTs.

~ 13 ~
 OBJECTIVE OF STUDY

1. To increase the effectiveness of the cross media advertising.

2. Reaching larger audience and customer base

3. To use the principle of united we stand, understanding and analyzing the individual qualities

of online and print advertisements and executing them in an integrated combined form, so that

we can get the maximum results and effectiveness.

~ 14 ~
 STATEMENT OF PROBLEM

The barriers which occurred during the investigation are manipulating data and records which

could give negative hypothesis results thus making my investigative title less effective. Other

potential threats which could make the statement bias are the de-merits of the cross-media

advertising technique which are being discussed. The online and print if combined can act well

if the operations functioned is having less number of errors and should be free from bias.

~ 15 ~
 SCOPE OF INVESTIGATION

The investigation involves:

1. The concept of combining online and print advertising and how it is effective than this

medium acting separately.

2. Combining of online and print advertising is an integrated multi-channel communication

strategy to increase advertising effectiveness.

3. Increased amount of viewers, greater response and higher returns are achieved due to

the emergence of cross media.

4. Cross media campaigns are effective and can lead a new product to a bloom period.

5. Cross media advertising need a large infrastructure.

6. Case studies that have an advantageous effect by the using the theory of cross media

advertising.

7. Application of cross media and the measurement of it effectiveness.

8. The challenges being faced if the both advertising techniques being combined into one.

9. The research work did not ended up to the effectiveness area but also shown a light to a

very effective medium which can be a tool of cross media effectiveness that is Cross

media campaigns.

10. Conflicts of interest which took place in this research are few incorrect data findings

and later it was corrected. No other potential threats were discovered.

~ 16 ~
 HYPOTHESIS 1

Characteristics of the internet as a marketing medium have been discussed in Novak and

Hofmann (1996). Although news magazines are similar to their internet pendants in that they

are both dominated by text and pictures, important differences with regard to advertising

effectiveness obtain. These include (a) attitude towards advertisement, (b) complexity and (c)

the carrier material. Advertising effectiveness depends on users’ receptiveness towards an ad

and on their attitude towards advertising. For print advertisement Metha (2000) found out,

that „respondents with more favorable attitudes towards advertising recalled a higher number

of advertisements the day after exposure“. In the internet, it is frequent that advertising is used

with higher levels of forced exposure than only static banners (Cho, Lee, Tharp 2001), who’s

level would compare to print advertising.

Since higher intrusiveness leads to ad avoidance and irritation (Edwards, Li, and Lee 2002), a

less favorable attitude among consumers vis-à-vis internet advertising can be supposed.

Assuming, that the overall attitude towards internet advertising is less favorable than towards

print advertising, lower ad memory can be expected. The internet with its hierarchical structure

is a more complex medium than print, being linear sequential. By clicking through websites and

choosing hyperlinks, the internet user has more control about what he actually sees, as

compared to a more passive reader of a news magazine (Bezjian- Avery, Calder, Iacobucci

1997). A print reader will more likely be exposed to an ad, even if it is only by skimming through

a magazine. On the contrary, an internet user directly clicks to an article of interest and will

easier skip undesired information, resulting in less advertising exposure and thus less effective

~ 17 ~
advertising. Moreover, the more active role in the internet requires deciding and thus

concentration, whereas a printed magazine can be skimmed through without concentrating on

navigation, allowing higher receptiveness. While news magazines are printed on paper, content

in the internet can only be read at screens. This is resulting in one of the fundamental

differences between print and internet: a screen will not be grasped and physically manipulated

as is the case with a magazine consisting of paper. Therefore, the hepatic component of

consuming content is different. Printed text can be touched and magazine pages turned,

whereas the screen impression is controlled at distance and indirectly, through clicking on

mouse, keyboard, touchpad, tracker balls, etc. The spatial plane of mouse movements is turned

from back / forth to down / up. Although laptop computers or personal digital assistants with

touch screens would enable a more flexible use of the internet, the predominant access

medium to the internet is still a desktop computer (European Commission and EOS Gallup

Europe 2002). But even touching a screen, consisting of glass or plastic, would result in a

different sensual experience than touching paper. In addition, reading a printed magazine is

linked with different odors and sounds than reading at a computer screen. While this difference

might not necessarily result in a less intense experience of content consumption, and thus a

worse advertising effectiveness in the internet, the combination of print and online advertising

can be assumed to be more effective than only exposure to either of the two media.

Moreover, as Sundar et al. (1998) have speculated, a computer screens “with its thick boxed

boundaries, *might+ limit readers’ attention to the center of the screen.” This could lead to a

reduced perception of ads placed at the border of pages. The carrier material can additionally

influence advertising effectiveness, as flickering displays and unfavorable color characteristics

~ 18 ~
with screens (e.g. radiated light spectrum differs from natural white light, contrast, or

brightness) might be tiring and thus negatively impacting concentration. Sundar et al (1998)

report a number of studies, in which subjects recalled “about the same amount of news

information from newspaper and computer”. However, this might not prevail for advertisement

information.

The discussed arguments lead to the following research hypotheses:

H1: Two exposures to a static advertisement lead

a) To a better recall

b) To a more favorable brand attitude and

c) To a higher purchase intention,

If the advertisement is displayed in print than if displayed in the internet

H2 Exposure to a static advertisement leads

a) To a better memory

b) To a more favorable brand attitude and

c) To a higher purchase intention,

If the advertisement is displayed in both the print medium and in the internet than if displayed only in

the internet.

H3 Two exposures to a static advertisement lead

a) To a better recall

b) To a more favorable brand attitude and

c) To a higher purchase intention,

~ 19 ~
If the advertisement is displayed in both the print medium and in the internet than if displayed only in

print.

RESEARCH DESIGN

The above hypotheses were tested in an experimental design, which allowed comparing the

effectiveness of one advertisement presented in print, web, or both. The participants (N=134)

were exposed twice to a current German news magazine containing twenty double pages with

articles and advertisements, either in the print format or the online equivalent, with the

respective advertisement either in print, online, combined, or not at all. After subjects read

through the magazine, their ad memory, brand attitude and purchase intention was measured

by an online questionnaire.

Participants

134 undergraduate business students took part in the experiment. The subjects were randomly

assigned to four groups, print, internet, combined, and control. No significant differences have

been found with regards to previous knowledge of the brand, attitude towards advertising,

internet experience, and usage frequency of the magazine, sex, and personality. Twelve

participants were excluded as not experienced internet users (participants' own estimated

internet experience: rather inexperienced or not experienced; average time of internet usage

per week: never; regular internet usage for: less than three months). According to Dahlen more

experienced internet users „*...+ are less inclined to react to unexpected stimuli .Experience

~ 20 ~
with print was assumed. Although earlier studies in this field mentioned the importance of a

comparable media experience (e.g. Gallagher, Forster, and Parsons 2001), it has not yet actually

been controlled for. Participants have been recruited in the university computer lab, with the

incentive of a one in twelve chance to win prizes worth a total of 120 Euros. The purpose of the

research was not disclosed.

Advertising Stimulus

A current issue of the weekly German news magazine "Der Spiegel" was used. It seemed

especially suitable, as an equivalent internet version is existing, publishing a similar range of

topics and a number of identical articles. Additionally, it is a well-known magazine that students

will know, but do not generally read. That was important in order to decrease the likelihood

that participants had already been exposed to the ad. Usage frequency during the last six

months among the participants was on average between "less than five times" and "never".

The advertisement was for the history magazine "GEO Epoche". GEO Epoche is a relatively new

monothematic magazine from the GEO family of brands with two issues per year. The GEO

brand is well known, but not GEO Epoche. The brand was chosen as it was thought to be

around the center of the FCB grid, with the two axes "involvement" and "feeling". A nice side

effect was that the advertisement highlighted the topic of the magazine rather than the brand,

which clearly distinguished it from other advertisements of the GEO brand family. An actual

advertisement and magazines were chosen in order to insure realistic quality of the stimulus

material and increase generalizability. The print magazine was prepared such that twenty

double pages could be read, including the table of contents. The advertisement was placed in

~ 21 ~
the second half of the available pages, between different other articles and advertisements. It

appeared next to an article about Jeff Koons, in the culture section. The advertisement had the

size of a full page, which is the standard advertisement format in that magazine.

For the preparation of the internet stimulus, the spiegel.de webpage was downloaded to a local

network drive. Most of the articles in the print version were available online, which allowed

creating a similar advertising context, with the ad placed also next to the equivalent article

about Jeff Koons in the same section. Links to magazine sections that have been taken out, e.g.

Travel, were made invisible, in order to keep users from spending their time clicking at different

disabled hyperlinks. Hyperlinks leading away from the site were disabled. Users could access

the article and the ad either with one click from the homepage, but had to scroll down, or with

two clicks via the culture section. The advertisement was a standard skyscraper banner

displaying the same text and layout as the printed version. It was not actually used as an

advertisement, but has been produced professionally in order to look similarly real.

Standard advertisement formats were used in order to increase applicability to practical

decisions, even if that means that the size of the advertisements differed. However, the

skyscraper format was used as the one that is the most comparable internet pendant to a single

magazine page. Skyscrapers are the largest banners and they are situated vertically next to the

text, whereas e.g. standard banners are horizontally on top of the text. Interstitials or pop-ups

were not used since these would interrupt users, which is not the case with standard print ads.

Animation and hyperlinks, although readily available for internet advertising, were omitted in

order to isolate the effect of medium.

~ 22 ~
Dependent variables

Internet advertising effectiveness has been measured using both web metrics and traditional

measures. A good overview about web metrics is provided by Bhat, Bevans, Sengupta (2002).

Traditional measures are based on recall (see, Alba, Hutchinson, and Lynch 1991), brand

attitude, and purchase intention.

Recall was operational zed through seven different questions, three recalls and four recognition

questions. Recall questions were open ended (e.g., "Various products or groups of products

have been advertised. Which of these do you remember?") or closed (e.g., "Do you remember

having seen one of the following brands?"). Recognition items asked for details of the

advertisement, e.g. "Which slogan was used to advertise for the issue?" Similar to Sundar et al.

(1998) correct recall and recognition answers were added up to for the memory variable.

Brand attitude was operationalized through a thirteen question Likert-type skale (nice/ugly;

comfortable/awkward; gentle/rough; attractive/repellent; tasteful/tasteless; exciting/boring;

colorful/colorless; fascinating/bland; convincing/not convincing; honest/dishonest;

novel/ordinary; fresh/frumpy; lively/liveless) asking participants to evaluate the GEO brand.

Four items have been excluded after the experiment as they proved not suitable to the GEO

brand (nice/ugly; attractive/repellent; honest/dishonest; and fresh/frumpy). Answers ranging

from one (positive) to five (negative) were summed up to create the variable “brand attitude”,

with 9 the best and 45 the worst value.

~ 23 ~
Purchase intention was measured through the question: "Assuming you are confronted with

the decision of buying GEO Epoche. How likely would that be?" with a five point scale

(certain/uncertain).

Procedure

Participants were randomly assigned to four groups, print, internet, combined, and control.

Groups contained between 30 and 37 subjects. Each participant was exposed to a medium

twice, i.e. two times to the printed magazine, two times to the internet magazine or a

combination of both. Half of the combined and control groups received the print version first,

half of it second. The medium included the GEO Epoche advertisement for the print, internet

and combined groups, the control group received media without the ad.

The participants were asked to perform two tasks, one for each exposure. In the first task

participants were asked to concentrate on the layout and on the proportion between articles

and advertising. The second task was comprised of skimming through the magazine with a

special focus on the culture section. The time limit for each task was five minutes. The tasks

were designed to increase the probability of being exposed to the relevant ad as well as to

simulate a usage situation somewhere in the middle between serious and playful (Rodgers and

Thorson, 2001).

After the two tasks, the stimulus material was removed and the participants were directed to

an online-questionnaire. The first questions were not related to the stimulus material

(questions about interests, usage of the magazine "Der Spiegel" and internet experience), in

order to simulate retroactive interference and long-term effects of the ad. The online-mode of

~ 24 ~
the questionnaire prevented participants from changing recall answers after they learned about

the focal brand.

Stimulus material, tasks and questionnaire were pre-tested by a small random sample resulting

in minor adjustments of the tasks and the questionnaire.

Control variables

Variables potentially influencing the results were previous knowledge of the brand, attitude

towards advertising (operationalized through the factor value of the agreement to four

statements,

e.g., "Advertising is informing"), internet experience (operationalized through asking for

participants' own estimated internet experience, their average time internet usage per week,

and the time they have been usage the internet regularly), usage frequency of the magazine,

sex, and personality (operationalized through the factor values of "attitude towards security"

and "willingness to perform" derived from ten related items and through items checking the

interest for six related dimensions (e.g. interest in history).

Results

Hypothesis testing

In order to test the hypotheses, one-way ANOVAs were conducted for the three different

samples and three dependent variables (see table 1 to 3). Hypothesis one predicted higher

advertising effectiveness for print advertising than for internet advertising. In accordance with

Sundar et al. (1998), the sample showed significant (F=2.813, p <.1) differences with regards to

short-time recall. Effects on brand attitude and purchase intention were supported, even if a

~ 25 ~
tendency (F=2.438) is visible with brand attitude. For the second hypothesis, claiming that

combined print and internet advertisement is more effective than only internet advertisement,

support was found with regards to brand attitude (F= 3.661; p<.1). However, better recall and

higher purchase consideration were supported. Between combined print and online advertising

and only print advertising (hypothesis three), significant differences were found.

Interestingly, brand attitude was worse for the print condition than for the internet condition,

with a difference of nearly two points or 7.7% (26.97 versus 25.04). It was also worse for the

combined condition than for the internet only condition, with a difference of 2.87 or 11.5%

(27.91 versus 25.04). This finding has two implications: (a) Together with the fact that the

control group had a significantly more positive brand attitude (18.45 vs. 26.52; F=55.77, p<.01)

than the other groups, it gives rise to the suspicion that the advertisement was actually not well

~ 26 ~
suited to increase the brand attitude towards the GEO brand. (b) The difference between the

combined condition and the internet condition is a little larger than the difference between the

print condition only and the internet condition. The combined condition, although not

significantly, also has a higher average than the print condition. That indicates that a stronger

effect for the media combination than for either of the single conditions might well be the case.

The reason for a lacking effect on purchase intention under all conditions seems to be a too

little sample size, as participants recalling the ad (aided recall) among all groups showed a little

higher purchase intention than did participants not recalling the ad (t=-1.964; p<.05).

Discussion

The most interesting result is that a combination of internet and print advertisement did show

significant differences from only print advertisement regarding recall and brand attitude. Since

internet and print address different senses, we had expected to lead to better results. However,

judging from the difference between internet and print alone, the combination can also be

seen as the combination of a highly effective medium. From that line of argument, the

~ 27 ~
combination should perform well than the highly effective medium alone. This argument

though seems to be the case, either: Since there are significant differences between print and

combined, the finding suggest that one print exposure can be replaced by one internet

exposure without reducing effectiveness, although internet advertisement alone is far less

effective than print advertisement. This is especially important since thousand contact prices in

the internet are much lower than in print. This finding is additionally interesting, since the

design of the study has not drawn upon all possibilities of internet advertising. Since animated

and clicked-at banners can be expected to be more effective than static banners, cross-media

campaigns are definitely a considerable alternative to pure print campaigns, potentially

rendering advantages to integrated media groups, and worth looking at in greater detail. Lower

advertising effectiveness for the online condition than for print is not surprising, as online

advertising was among all participants considered more disturbing than print advertising (1.98

versus 2.94 on a five point scale). The internet magazine is hierarchical, as well, requiring

concentration on navigation and reducing the likelihood of being exposed to the ad.

Additionally, although modern computers and screens were used, negative impacts of the

carrier material cannot be excluded. However, it is possible, that smaller size of the skyscraper

ad is responsible for that effect. The fact that only the memory measure proved significant

might be due to the fact that the stimulus material and the tasks forced the participants

towards the ad in every condition, with the effect that few significant differences have been

found. Thus the main statement is that the combined method is better than this two (print &

online) by differences notable. The method may have few problems and barriers, still it is

effective it its own way.

~ 28 ~
 Limitations and further research

Although usual, student samples do pose problems to external validity. Additionally, modern

computer facilities were used with the website hosted at a network drive. Since the majority of

internet users are still accessing the internet by telephone lines, this might pose a problem

regarding generalizability. Only one product category was tested, the results might be different

for other products. Significance levels of 10% do not suffice strict scientific standards.

In this study, print advertisements were compared to static internet banners. While this allows

isolating effects to the medium itself, it does not compare the two media in use of their full

potential. Further research would be required to compare animated and hyperlinked

advertisements to print, as well as the combined use of print and online with different motives,

e.g. referring to one another.

~ 29 ~
 HYPOTHESIS 2

A Bayesian network modeling approach for cross media analysis

Existing methods for the semantic analysis of multimedia, although effective for single-medium

scenarios, are inherently flawed in cases where knowledge is spread over different media types.

In this work we implement a cross media analysis scheme that takes advantage of both print

and online information for detecting high-level concepts. The novel aspect of this scheme

is the definition and use of a conceptual space where information originating from

heterogeneous media types can be meaningfully combined and facilitate analysis decisions.

More specifically, our contribution is on proposing a modeling approach for Bayesian Networks

that defines this conceptual space and allows evidence originating from the domain knowledge,

the application context and different content modalities to support or disproof a certain

hypothesis. The obtained results have shown that the proposed modeling approach exploits

the complementary effect of evidence extracted across different media and delivers

performance improvements compared to the single-medium cases. Moreover, by comparing

the performance of the proposed approach with an approach using Support Vector Machines

(SVM), we have verified that in a cross media setting the use of generative rather than

discriminative models are more suited, mainly due to their ability to smoothly incorporate

explicit knowledge and learn from a few examples.

The automatic extraction of semantic metadata from multimedia content has been recognized

as a particularly valuable task for various applications of digital content consumption. Current

literature has made considerable progress in this direction especially for single-medium

~ 30 ~
scenarios. However, the methods proposed in the literature do not apply in cases where

information is spread over different media types and unless considered simultaneously, its

contribution cannot be fully exploited by the analysis process. Motivated by this, cross media

analysis seeks to enhance semantic metadata extraction by exploiting information across

media. Practically, the aim of such methods is to combine the evidence extracted from different

media

types and accumulate their effect in favor or against a certain hypothesis. These pieces of

evidence can belong to different levels of granularity and used differently by the analysis

mechanism. For instance, we can consider cross media analysis to be a general fusion problem

that is carried out at different levels of abstraction, namely result-level [1], [2], [3], extraction-

level [4], [5], [6] and feature-level [7], [8], [9].

In the result-level approach, information from each data source is initially extracted separately

and, still separately, transformed into conceptual information. Though result-level approaches

are closer to human cognition and more suited for exploiting explicit knowledge (i.e.,

knowledge that is explicitly provided by experts in the form or rules, ontology’s or other formal

languages for knowledge representation), their major drawback is that each extractor has to

produce its own internal evidence in order to extract the conceptual information. In the

extractor-level approach the conceptual information is not extracted separately from each

modality but instead, the analysis mechanism takes into account evidence from other

modalities. Information coming from one medium may assist the information extraction

module of another medium, using as input the output of another extractor. However, in

contrast to the result-level approaches where knowledge is incorporated into the conceptual

~ 31 ~
space, in this case it can only be exploited as part of a task specific mechanism. The feature

level approach consists in using all low-level features that can be extracted from each medium

within the same analysis process. Initially, low-level features (e.g., text-tokens, named entities

or image descriptors) are extracted separately from each modality and integrated into a

common, concatenated representation. Subsequently, the common representation is used as

input for the analysis process (i.e., classification, indexing, etc). Feature-level analysis aims at

exploiting the joint existence of low-level features into the same resource, but it is rather

difficult to incorporate explicit knowledge in this case. Our work is motivated by the need to

boost the efficiency of cross media analysis using the knowledge explicitly provided by domain

experts (i.e., domain knowledge). This was the reason for developing a method that operates

on the result-level of abstraction and allows domain knowledge to become part of the

inference process. Our method combines the soft evidence (soft in the sense that a confidence

degree is attached to every piece of evidence) collected from different media types, to support

or disproof a certain hypothesis made about the semantic content of the analyzed resource.

Soft evidence are obtained by applying single-medium analyzers on the low level features of

the different media types. Subsequently, these pieces of evidence are used to drive a

probabilistic inference process that takes place in a Bayesian Network (BN). The structure and

parameters of the BN are constructed by incorporating domain knowledge (expressed using

ontologies) and application context (captured by conditional probabilities). We use the soft

evidence to update the observable variables of the BN and verify or reject the examined

hypothesis based on the posteriori probability of the remaining variables. Fig. 1 demonstrates

the functional relations between the components of the proposed cross media analysis scheme.

~ 32 ~
The novelty introduced by the proposed method is that it manages to integrate into a common

inference framework three types of information, a) information obtained from the analysis of

heterogeneous content (i.e., the output of single-medium analyzers supplied as soft evidence),

b) information about the domain that is provided explicitly, and c) contextual information

that is learned from sample data. Our contribution is on proposing a modeling approach for the

BN that results in a conceptual space of likelihood estimates. In this space the evidence

originating from the domain knowledge, the application context and the different content

modalities can be meaningfully combined and facilitate semantic metadata extraction.

We show using content from a real world application taken from the car manufacturing

industry as well as from the TRECVID2010 competition, that performing cross media analysis

using the proposed method leads to significant improvements compared to the cases where

single-medium analyzers act separately. We also prove experimentally that, in a cross media

setting, the generative models outperform the discriminative ones in fusing the extracted

evidence, mainly due to their ability in efficiently handling prior knowledge

~ 33 ~
and learning from a few examples.

2. Modeling the Bayesian Network

In this section we describe the main contribution of our work, which is a generic approach for

modeling BNs that can be used to define a conceptual space suitable for combining

heterogeneous types of information. The types of information that are handled by this

approach are: a) conceptual information shared amongst most individuals that determines the

logical relations between concepts, such as sub-class, union, intersection, disjoint, etc (i.e.,

domain knowledge), b) information that qualitatively evaluates the co-existence of concepts,

encoding for example how likely is for one concept to be present

4 when another concept is verified (i.e., application context), and c) information extracted from

content analysis that encodes the support received from the analyzed low-level features in

favor of a specific concept (single-medium evidence). Our approach relies on probabilities and

probabilistic inference to define the common conceptual space.

More specifically, the explicitly provided domain knowledge is used to determine the structure

of the BN and in this way enforce the logic rules of the domain during inference. The

application context is approximated by the co-occurrence frequency between domain concepts,

information that can be extracted using a sample of the population that is being modeled. The

application context is encoded into the Conditional Probability Tables (CPTs) of the BN nodes,

which influence the inference process when belief propagation takes place. However, the most

critical point is how to incorporate the information received from content analysis. In order to

~ 34 ~
do this, we treat the outcome of single-medium analyzers as soft evidence that are used to

instantiate the nodes of a BN operating on a conceptual true-false space.

The reason for selecting these states (i.e., true, false) to be the only possible states of all

network nodes, was to establish a “lingua franca” between the heterogeneous types of

information and facilitate the incorporation of domain knowledge in decision making. By

adopting the proposed modeling approach the constructed BN does not operate on the low-

level features of the content, which would constitute a typical application of the BN theory.

Instead, it operates on the space determined by the probability estimates (that we call

conceptual true-false space), obtained through the application of machine learning techniques

on the low-level features (as described later in Section 3.2). In the following, we provide details

on how the proposed modeling approach can be used to determine a BN for analyzing

compound documents, but can be seamlessly applied to analyze any other multi-modal

resource, or handle an arbitrary number of modalities.

Let us consider a set of compound documents D where each document is composed of its visual

and textual part:

Di = [Ti, Vi] (1)

Let also ti, vi be the features extracted from Ti, Vi respectively. We consider the single-medium

analyzers to be the functions fcj (·) and gcj (·) that outputs the probability of a given concept cj

being valid for a document either based on its textual or visual low-level features, respectively:

fcj (Ti) = P(cj = true|ti), for the textual part of Di

gcj (Vi) = P(cj = true|vi), for the visual part of Di (2)

~ 35 ~
Thus, if we have a single-medium analyzer that is trained to detect all domain concepts 8cj 2 C,

it produces |C| probabilities when applied on a document Di. In order to construct a BN that

operates on a conceptual true-false, for every concept cj we create a discrete random variable

with two states rz = {true, false}. Then, we link these nodes based on their logical relations (as

explained in Section 3.3.1) and learn the CPTs by applying the Expectation Maximization

algorithm on sample data (as detailed in Section 3.3.1). We consider the output of the single-

medium analyzer to

formulate a new feature space y, determined from the probability estimates. We refer to this

new feature space as conceptual true-false space. By applying the Bayes rule in feature space y

we have for each concept cj :

Pcj (rz) represents our prior knowledge about cj and in the conceptual true-false space we

accept that Pcj (rz = true) is equal to the frequency of appearance of cj in the domain (i.e., how

often appears in the training set). Respectively, we accept that Pcj (rz = false) = 1 − Pcj (rz =

true). Pcj (y) is a scale factor that guarantees that the posterior probabilities sum to one and

equals:

Pcj (y|rz) is the likelihood (or class conditional probability) of rz with respect to y. Pcj (rz|y) is

the posterior probability of rz after considering the analysis outcome and taking into

consideration prior knowledge. In order to facilitate the analysis process we need to calculate

the posterior probabilities for each independent piece of conceptual information (i.e., 8cj 2 C),

~ 36 ~
so we need to know Pcj (rz = true|y). It is clear from eqs. (3) and (4) that in order to calculate

this value, what we are missing is Pcj (y|rz = true) and Pcj (y|rz = false). Recalling that fcj (·) and

gcj (·) provides us with

a probability expressing how much support cj receives from the textual or visual low-level

features of the document respectively, we incorporate the content analysis outcome into the

decision process by instantiating the nodes of the BN as follows:

**** Here textual evidence refers to print and visual evidence refers to online

Thus, during the analysis process we inject, as explained above, the output of single-medium

analyzers into the BN and perform probabilistic inference by propagating evidence beliefs.

Eventually, the resulting posterior probability for the “true” state of the node corresponding to

the concept that we want to detect, is considered to be the confidence degree for this concept.

3. Cross media analysis scheme

In order to verify the benefits of the proposed modeling approach we use it to design a cross

media analysis scheme that detects high-level concepts. High-level concept detection is usually

the output of knowledge-related tasks and typically requires the synergy of information

scattered in different places. The more the available information, the more easily is for the

~ 37 ~
knowledge worker to infer the presence of a high-level concept. Independently of whether

these pieces of information act cumulatively or complementary, they have an impact (i.e.,

positive or negative) on the confidence of the fact that a certain high-level concept is valid for

the analyzed resource. In order to model this process we rely on the approach presented in

Section 2 and implement a generative classifier based on BNs. The role of this classifier is to i)

fuse the information extracted from different media types on the grounds of knowledge and

context, ii) produce a confidence degree about the validity of a high-level concept in the

analyzed resource, and iii) make a decision by applying a fixed threshold on this confidence

degree. Since cross media analysis is mostly about simultaneously evaluating the appropriate

evidence extracted across different media types, an important issue for making the

aforementioned framework suitable for such purposes is the strategy by which evidence (and

as a consequence their source modalities) are considered to be co-related. In the following

subsections we elaborate on the components that are used to implement the cross media

analysis scheme for compound documents, which are: a) a dismantling mechanism and a

modality synchronization strategy for handling the compound media resources, b) the single-

medium analysis techniques for extracting evidence using low level features, and c) the

techniques used to construct and perform inference on a BN that is modeled.

4. Experimental Study

The goal of our experimental study was to evaluate our proposed methodology in three

different aspects: i) how much improvement is achieved by the employment of the proposed

cross-media analysis scheme compared to single medium solutions, ii) whether the choice of a

generative over a discriminative model is more suited for fusing evidence coming from

~ 38 ~
heterogeneous sources, and iii) whether the additional cost of engineering an ontology for

expressing domain knowledge, actually pays off in terms of efficiency when compared with less

costly approaches like using a simplified BN or learning its structure from data using the K2

algorithm [33]. Finally, we have verified the efficiency of our framework to more general

application, by using the proposed modeling approach.

4.1. Tested

The domain selected for performing our experimental study concerns forecasting the launch of

competitors’ models, as defined in cooperation with Centro Ricerche Fiat (CRF)1. The goal of a

competitor analysis departments to constantly monitor the existent competitors’ products,

understand market trends and try to anticipate customer needs. The information needed to

achieve that, is scattered throughout the Internet (i.e., blogs and forums), and covered by a

long tail of international and national automotive magazines. In a typical scenario the main role

is played by the person responsible for data acquisition that has the responsibility of daily

inspecting a number of resources such as WWW pages, car exhibitions, car magazines, etc, that

are likely to publish material of potential interest. The collected information is subsequently

used in the set-up stage of new vehicles (i.e., the development stage where a first assessment

of the future vehicle’s features is carried out).This process is of great value to many companies

because it contributes to keeping new products design up to date. One of the tasks defined by

the experts was to be able to automatically evaluate a document with respect to its interest for

the car components ergonomic design. The fact that most of the collected documents use both

visual and textual descriptions motivated the construction of a cross media classifier

~ 39 ~
recognizing compound resources that are valid for the high-level concept car components

ergonomic design.

** This was the advertisement that was posted in the online website of the automobile

company FIAT. The advertisement gave details about the car features which could possibly

satisfy the customers demand. If we look closely in the advertisement it is a print advertisement

which was posted in online. Here, the advertisement is a cross media mechanism that FIAT

chose to advertise it around the world and it gained a huge amount of response. Thus we could

say the combination of online and print advertisement is much more effective.

4.3. Single vs Cross media analysis

~ 40 ~
In the case of visual-only analysis, the general knowledge about the specific domain was

expressed by the ontology depicted in Fig. 6(a). This ontology associates five visual concepts,

namely air ducts, steering wheels, gear levers, car pedals and interior with the high-level

concept car components ergonomic design. The trained BN used for this setting is depicted in

Fig. 6(b).Five detectors trained to identify the five concepts of the domain ontology were

implemented using the method of Section 3.2.1. These detectors were trained using an

independent dataset of 3230 images depicting car interiors that was strongly annotated at

region-detail. Each of these detectors was attached to the corresponding BN node of Fig. 6(b)

and was used to trigger the process of probabilistic inference. By applying these five detectors

on every image contained in a document page and using their output to instantiate the network

nodes, we are able to decide about the existence of the high-level concept car components

ergonomic design in a document page, based solely on the information depicted on the images

of this page. The obtained results are depicted in Fig 11.

The figures are as follows:

~ 41 ~
~ 42 ~
~ 43 ~
Thus it is proved by the testing that the cross media is far more efficient from the individuals.

The results showing that when performing cross media analysis at the result-level, the

generative models are more suited for incorporating explicit knowledge and outperform the

discriminative models that luck a straightforward way to benefit from such knowledge. Our

plans for future work include the use of the proposed modeling approach for combining

information from more media types (i.e., images, text, sound, and sensor data).

~ 44 ~
~ 45 ~
 CONTENT ANALYSIS

Content analysis is a method for summarizing any form of content by counting various aspects

of the content. This enables a more objective evaluation than comparing content based on the

impressions of a listener. Content analysis is a technique for systematically describing written,

spoken or visual communication. It provides a quantitative description and as well as

qualititative description.

Cross-media marketing is a form of cross-promotion in which promotional companies commit

to surpassing traditional advertisement techniques and decide to include extra appeals to the

products they offer. The material can be communicated by any mass media such as e-mails,

letters, web pages, or other recruiting sources. Cross-media advertising is a strategy used by

business owners to market a business using various types of media. Owners may use all of

media types individually or combine several mediums to create a cohesive marketing campaign.

Hybrid Media is More Powerful Media

Magazine advertising sales people have a big advantage when selling digital ads to their clients.

When a national magazine campaign is planned in 2014 and beyond, there is almost certainly a

digital campaign coming down the pike behind it. Digital planning comes later because the lead-

time to get it into media is far shorter; 90 days for monthly magazines compared to a week or

even less for digital. When your magazine wins the business, that win should give your sales

person a head-start and an advantage to sell the Internet campaign onto your site.

You and your team have an advantage for two reasons:

 Your non-print competitors may not even know that a campaign is being planned. But

~ 46 ~
 your team can assume — until proven otherwise — that a digital plan is in the works.

 Research shows buying media from the same brand online and off is more effective.

Provided your sales team can communicate this research, there should be a bias toward

adding digital impressions to the print impressions directed to your audience.

Combining Online Media and Print Marketing

Just a few years ago people were shouting "Print is dead" from the rafters. These

premature prognosticators couldn't have been further from the truth as online and print

marketing are melding into cohesive multi-channel campaigns that prove to be more

effective than either on its own. Even the largest corporations that were founded online

have stepped into the print marketing arena to drive traffic to their sites. The following

are a few ways companies are combining online media and printing to further their

business success.

TESTING

The discussed arguments lead to the following research hypotheses:

H1: Two exposures to a static advertisement lead

a) To a better recall

b) To a more favorable brand attitude and

c) To a higher purchase intention,

If the advertisement is displayed in print than if displayed in the internet

H2 Exposure to a static advertisement leads

~ 47 ~
 a) To a better memory

b) To a more favorable brand attitude and

c) To a higher purchase intention,

H3 Two exposures to a static advertisement lead

a) To a better recall

b) To a more favorable brand attitude and

c) To a higher purchase intention

Modeling the Bayesian Network

In this section we describe the main contribution of our work, which is a generic approach for

modeling BNs that can be used to define a conceptual space suitable for combining

heterogeneous types of information. The types of information that are handled by this

approach are: a) conceptual information shared amongst most individuals that determines the

logical relations between concepts, such as sub-class, union, intersection, disjoint, etc (i.e.,

domain knowledge), b) information that qualitatively evaluates the co-existence of concepts,

encoding for example how likely is for one concept to be present

4 when another concept is verified (i.e., application context), and c) information extracted from

content analysis that encodes the support received from the analyzed low-level features in

favor of a specific concept (single-medium evidence). Our approach relies on probabilities and

probabilistic inference to define the common conceptual space.

~ 48 ~
OUTCOME

Cross media advertising is a strategy used by business owners to market a business using

various types of media. Owners mainly uses the two kind of medias, one the print

advertisements and other the digital or online advertisements. Cross-media advertising is one

often overlooked by business owners, especially internet entrepreneurs, according to WEB PRO

NEWS. Keep the target consumer profile in mind when choosing media for a given business to

reach maximum profits. So, if we combine the two medium of advertisements (online and

print) these are the possible outcomes which are briefly mentioned below both in

advantageous and disadvantageous forms.

Advantages

1. Personalized touch to the customers

2. Multiple opportunities to reach the audience.

3. Guaranteed returns on ROI

4. Greater response rates and interaction

Disadvantages

1. Requirement of large scale of money/funding/large infrastructure/excellence in operations

with zero error.

2. Complexity in its nature

3. Time consuming

4. Labour intensive

~ 49 ~
Effects of print and internet on advertising effectiveness

Characteristics of the internet as a marketing medium have been discussed in Novak and

Hofmann (1996). Although news magazines are similar to their internet pendants in that they

are both dominated by text and pictures, important differences with regard to advertising

effectiveness obtain. These include (a) attitude towards advertisement, (b) complexity and (c)

the carrier material. Advertising effectiveness depends on users’ receptiveness towards an ad

and on their attitude towards advertising. For print advertisement Metha (2000) found out,

that respondents with more favorable attitudes towards advertising recalled a higher number

of advertisements the day after exposure“. In the internet, it is frequent that advertising is used

with higher levels of forced exposure than only static banners (Cho, Lee, Tharp 2001), who’s

level would compare to print advertising. Since higher intrusiveness leads to ad avoidance and

irritation (Edwards, Li, and Lee 2002), a less favorable attitude among consumers vis-à-vis

internet advertising can be supposed. Assuming, that the overall attitude towards internet

advertising is less favorable than towards print advertising, lower ad memory can be expected.

The internet with its hierarchical structure is a more complex medium than print, being

linearsequential. By clicking through websites and choosing hyperlinks, the internet user has

more control about what he actually sees, as compared to a more passive reader of a news

magazine (Bezjian- Avery, Calder, Iacobucci 1997). A print reader will more likely be exposed to

an ad, even if it is only by skimming through a magazine. On the contrary, an internet user

directly clicks to an article of interest and will easier skip undesired information, resulting in less

advertising exposure and thus less effective advertising. Moreover, the more active role in the

internet requires deciding and thus concentration, whereas a printed magazine can be skimmed

~ 50 ~
through without concentrating on navigation, allowing higher receptiveness. While news

magazines are printed on paper, content in the internet can only be read at screens. This is

resulting in one of the fundamental differences between print and internet: a screen will not

be grasped and physically manipulated as is the case with a magazine consisting of paper.

Therefore, the hectic component of consuming content is different. Printed text can be touched

and magazine pages turned, whereas the screen impression is controlled at distance and

indirectly, through clicking on mouse, keyboard, touchpad, trackerballs, etc. The spatial plane of

mouse movements is turned from back / forth to down / up. Although laptop computers or

personal digital assistants with touch screens would enable a more flexible use of the internet,

the predominant access medium to the internet is still a desktop computer (European

Commission and EOS Gallup Europe 2002). But even touching a screen, consisting of glass or

plastic, would result in a different sensual experience than touching paper. In addition, reading

a printed magazine is linked with different odors and sounds than reading at a computer

screen. While this difference might not necessarily result in a less intense experience of content

consumption, and thus a worse advertising effectiveness in the internet, the combination of

print and online advadvertising can be assumed to be more effective than only exposure to

either of the two media. Moreover, as Sundar et al. (1998) have speculated, a computer

screens “with its thick boxed boundaries, *might+ limit readers’ attention to the center of the

screen.” This could lead to a reduced perception of ads placed at the border of pages.

The carrier material can additionally influence advertising effectiveness, as flickering displays

and unfavorable color characteristics with screens (e.g. radiated light spectrum differs from

natural white light, contrast, or brightness) might be tiring and thus negatively impacting

~ 51 ~
concentration. Sundar et al (1998) report a number of studies, in which subjects recalled “about

the same amount of news information from newspaper and computer”. However, this might

not prevail for advertisement information.

We have proposed a modeling approach for the BN that determines a conceptual space. This

space allows machine learning techniques and probabilistic inference frameworks to be

effectively combined for the purpose of semantic multimedia analysis. We have used the

proposed conceptual space to combine evidence originating from different multimedia types

and perform cross media analysis of compound documents and video shots. Our experiments

have verified that there are cases where the information contained in a multi-modal resource

can only be extracted if evidence are considered across media. Moreover, it has been proven

that information coming from the domain knowledge is particularly useful, especially when

dealing with heterogeneous types of content. Interesting were the results showing that when

performing cross media analysis at the result-level, the generative models are more suited for

incorporating explicit knowledge and outperform the discriminative models that luck a

straightforward way to benefit from such knowledge. One important requirement of the

presented scheme is that it needs a deep modeling of the analysis context (in terms of

engineering the domain ontology and producing cross media annotations), which makes the

approach appropriate for cases where this effort is justified by the added value in the

application. Our plans for future work include the use of the proposed modeling approach for

combining information from more media types (i.e., images, text, sound, sensor data).

~ 52 ~
~ 53 ~
~ 54 ~
Our work is motivated by the need to boost the efficiency of cross media analysis using the

knowledge explicitly provided by domain experts (i.e., domain knowledge). This was the reason

for developing a method that operate son the result-level of abstraction and allows domain

knowledge to become part of the inference process. Our method combines the soft

evidence(soft in the sense that a confidence degree is attached to every piece of evidence)

collected from different media types, to support or disproof a certain hypothesis made

aboutthe semantic content of the analyzed resource. Soft evidence is obtained by applying

single-medium analyzers on the low level features of the different media types. Subsequently,

these pieces of evidence are used to drive a probabilistic inference process that takes place

in a Bayesian Network (BN). The structure and parameters of the BN are constructed by

incorporating domain knowledge (expressed using ontology’s)and application context (captured

by conditional probabilities). We use the soft evidence to update the observable variables of

the BN and verify or reject the examined hypothesis based on the posteriori probability of the

remaining variables. The novelty introduced by the proposed method is that it manages to

integrate into a common inference framework three types of information, a) information

obtained from the analysis of heterogeneous content (i.e., the output of single-medium

analyzers supplied as soft evidence), b) information about the domain that is provided

explicitly, and c) contextual information that is learned from sample data. Our contribution is

on proposing a modeling approach for the BN that results in a conceptual space of likelihood

estimates. In this space the evidence originating from the domain knowledge, the application

context and the different content modalities can be meaningfully combined and facilitate

semantic metadata extraction.

~ 55 ~
Cross media advertising is a strategy used by business owners to market a business using

various types of media. Owners mainly uses the two kind of medias, one of the print

advertisements and other the digital or online advertisements. Cross media advertising is one

often overlooked by business owners, especially internet entrepreneurs, according to Web Pro

News. Keep the target consumer profile in mind when choosing media for a given business to

reach maximum profits.

There are certain questions which makes the critical appraisal of the research more effective:

 Is the Study's Research Question Relevant?

 Does the Study Add Anything New?

 What is Cross media campaign?

 How the productivity of advertisements increases if it’s combined?

 What are the outcomes of effectiveness of advertisements measurements?

 What is application of cross media and its effects?

 What Type of Research Question Does the Study Pose?

 What are the advantages and disadvantages of using combined online and print media

advertisements?

~ 56 ~
 Was the Study Design Appropriate for the Research Question?

 Economic Evaluations

 Do the Data Justify the Conclusions?

 Are There any Conflicts of Interest?

 Does the Study Test a Stated Hypothesis?

 Did the Study Methods Address the Key Potential Sources of Bias?

 Systematic Reviews and Meta-analyses

 The QUORUM (Quality of Reporting of Meta-Analyses)

 Randomized Controlled Trials

~ 57 ~
 DATA COLLECTION AND DATA ANALYSIS

Case Study 1:

Rate of admission increases in Ideal Public School, Howrah: In the year 2014, Ideal Public School

gave advertisements only in the newspapers and catalogues for their admission purposes, but

last year 2015 they published both their admission brochures through media channels, social

media interfaces by combining print and online advertisements and their rate of admission

increased by 25.4% more.

Print Advertisement
Catalogues
Hand Bills
Newspapers

Fig 1: It shows the advertisement techniques used by the school in 2014 and they have a
normal admission rate. 5.3% increased in their admission.

Combined (Online and

Print Advertisement)

Newspaper
Handbills
Official Website
Facebook
Radio
Social media

Fig 2: It shows the advertisements techniques used by the school in 2015 and they have an
increased rate in their admission by 25.4%.

~ 58 ~
Case Study 2:

More buyers being moved to e-commerce (Flipkart). Flipkart is one of the recognized e-

commerce in India. In the early period of Flipkart, it used to give advertisements only in online

and digital medium and the only customers they got was teenagers and the people who used

gadgets on daily basis. But for last 2 to 3 years Flipkart also uses hoarding, newspapers,

handbills, catalogues, banners and print media as its advertisement weapon in today’s perfect

competitive market. Thus not like as the earlier stage the digital generation is attracted but also

the age of 40 to 80 years moved into the world of e-commerce. Flipkart’s rate of selling has

been increased by 46.3%.

CUSTOMER HAVING KNOWLEDGE ABOUT FLIPKART

Age 15-30

Age 30-45

Age 45-60

Above Age 60 (Unknown

about Flipkart)

Fig 3: Here we see that when the Flipkart was at its start level, it used only online
advertisements and thus the people who are not accustomed of using gadgets and internet
on daily basis are unaware of the e-commerce website and application. As in the pie chart
you can see the violet pie (age above 60) is divided among the other three.

~ 59 ~
 CUSTOMER HAVING KNOWLEDGE ABOUT FLIPKART

Age 15-30
Age 30-45
Age 45-60
Above Age 60

Fig 4: It shows that when the combined technique was used all the customers of various age
was added into the customer base of Flipkart, and the violet pie (age above 60) is also a
customer in the e-commerce website.

~ 60 ~
 SURVEY SAMPLE QUESTIONS

The questions asked to get a better survey results.

SAMPLE 1

Name:

Age :

Occupation:

Q1) Which is the best and effective advertising medium according to you?

A) Radio & Television

B) Print

C) Online

D) Combined (Online & print)

Q2) Which among the e-commerce do you prefer in terms of branding and advertisement?

A) Flipkart

B) Snapdeal

C) Amazon

D) Myntra

~ 61 ~
Q3) Why do you prefer combined (online and print) advertisements?

A) Easily accessible

B) Connects socially

C) Being update mostly

D) Reach greater customer base

Q4) Is cross platform advertising current practices and issues of the future?

A) Yes

B) No

Q5) If you had any chance of starting an advertisement with low cost budget, what method do

you choose?

A) Online

B) Print

C) Both

D) Radio and Television

~ 62 ~
Q6) Is it necessary to always to give advertisements if brand matters?

A) Yes

B) No

Q7) Do you believe in statistics or hypothesis before making an advertising move?

A) Yes

B) No

Q8) Do you believe that endorsing products with celebrities, promotes the advertisement more?

A) Yes… Justify

B) No… Justify

Q9) In your daily life, in which medium you get connected with the advertisements mostly?

A) Radio

B) Television

C) Print

D) Online

Signature:

Date:

~ 63 ~
 INTERVIEWS

1. SACHIN BANSAL

(CO-FOUNDER OF FLIPKART)

Q) Which one is the effective medium according to you? Is it the online and print functioning

solely or both performing together?

S. Bansal: Frankly, as I am totally inclined to the world of internet and digital media, my clean

opinion goes to the both performing combined nature. If you notice the front end of the flipkart

website or application, you will clearly notice that it is the application of both the online and

print advertisement. The banner which you see in the roads is the same which is being posted

in the website or application, the fact that is here in the online you can zoom out and zoom in

the image and go through the detailing of it. And as per my data, in the year 2015 I gained a lot

of profit by using this technique of cross media, thus making myself to the point of break even.

In the year 2015 what I used is that I gave advertisements on different TV channels, Facebook

and YouTube, but I didn’t stop my techniques here, we began to spread different catalogues

and posted banner and print advertisements. So, I wish that I have justified myself that why I

am inclined towards the combined technique rather than the both acting solely. In the near

future we are thinking of more futuristic cross media mechanisms so that we can do

hypodermic precision to the customer demands and satisfy them well. Hopefully in the 2016

Diwali big billion offer we can make huge move by the technique.

~ 64 ~
 INTERVIEWS

2. CHANDRASEKHAR GHOSH

(FOUNDER, MANANGING DIRECTOR AND CEO OF BANDHAN BANK LTD)

Q) What was the marketing strategy taken by Bandhan Bank in the last year to get such high
profits?

C S Ghosh: Bandhan got license from RBI in the year 2014 to be bankable within 2015, so we all

started making preparations and in the year 2015, month of august Bandhan Bank was

transformed into bank. But in this competitive arena, it was too difficult to gain customers, so

we focused our strengths to advertisements and marketing. If you could remember, when the

day the bank was inaugurated, the program was telecasted to 35 channels and in Kolkata,

around 30,000 cut outs, banners, prints were hoarded. May be I am a traditional fellow, but my

mindsets are into cross media techniques where I used my modern advertisement skills by

using both online and print advertisements to be work side by side. If you follow the page of

Bandhan Bank in Facebook, you will get it clear. Moreover, I contacted YouTube for the same.

Bandhan Bank spent 1/3 of its expenses to cross media advertisements and now in the

upcoming year we are thinking of something new so that more people will be getting attracted.

By using the cross media techniques few advantages we got that is

1. Personalized touch to the customers

2. Multiple opportunities to reach the audience.

And among the two benefits, number 2 is very much essential for us because maximum of our

branches are opened in rural and semi-urban areas.

~ 65 ~
 INTERVIEWS

3. BINOD KUMAR HOMAGAI

(FOUNDER OF WOW MOMO PVT LTD)

Q) Before launching WOW MOMO, did you calculated the advertising effectiveness?

B.K Homagai: Of course, without a statistics you just cannot move into making a new born

business, as I was a commerce graduate, I did not have a sufficient knowledge about the

measurement of advertising effectiveness, and how to calculate the same. So I took a training

of 3 months to get a view that what methods should be used after making the company

started. It was very important for me to select a medium for advertisements, but after few

studying and research I selected cross media technique (that is the combination of online and

print advertisements) because at the scratch level it was very important for my company to

reach a larger audience and customer base. Moreover, after few months we have opened a

portal in the internet which focused only on advertising. I have used social media for the

purpose of advertisements, spread different handbills, send texts to random numbers about

different offers and invited them for certain discounts, launched different new attracted

handbills on the road and as you know today is the world of Google and Internet so I managed

to make some online-print advertisements. In the last year during the Durga puja WOW MOMO

gained a profit of 20 crores just because of the use of the advertisements as the advertisements

helped WOW MOMO to reach a larger audience.

~ 66 ~
 INTERVIEWS

4. ANJAN CHATTERJEE

(FOUNDER OF HOUSE SPECIALITIES, MAINLAND CHINA, MACHAAN, OH CALCUTTA!!!)

Q) Do you believe in cross media advertising?

A.Chatterjee : Yes I do believe very much in cross media advertising because Cross-media

marketing is a form of cross-promotion in which promotional companies commit to surpassing

traditional advertisement techniques and decide to include extra appeals to the products they

offer. The material can be communicated by any mass media such as e-mails, letters, web

pages, or other recruiting sources. Marketers have increasingly made online marketing an

integral part of their multi-channel communications strategies. There are certainly advantages

to using online marketing including economies of scale, direct fulfilment, exceptional targeted

advertising capabilities, and of course the ability to track advertising.

Q) Do you have tested hypothesis before making cross media mechanism?

A.Chatterjee: Of course, you just cannot do a new thing just dreaming, you have to test results

out of it. I have completed more than 10-20 hypotheses before making my move and every

time the results were null thus making me more inclined towards the cross media mechanism.

Today ZOMATO, the food application is a conceptual example of print and online advertisement

working together, and the idea was peculated by me and my associates by the results of

different hypotheses testing. Not only hypotheses, I have completed several content analysis of

different articles and journals, moreover here data collection & analysis is an important feature.

~ 67 ~
 INTERVIEWS

SRIJIT MUKHERJEE

(INDIAN FILM DIRECTOR)

Q) From when have using cross media as one of your weapon during promotions of movies

before releasing?

S .Mukherjee: Honestly, speaking being a director you always don’t get an opportunity to select

the advertisement techniques of your film, but I am lucky that I am every time being called for

advices that which advertisement techniques should be selected during the promotional

activities of the unreleased film and what should be the techniques after releasing it? It totally

depends on the budget of the film, I am always favored to cross media that is combination of

online and print because it is very much effective. If the budget is low you should move only to

print advertisements and if you have sufficient budget to deal with go with cross media

mechanism of advertisements.

Q) What are the disadvantages of cross media advertising?

1. Requirement of large scale of money/funding/large infrastructure/excellence in operations

with zero error.

2. Complexity in its nature

3. Time consuming

4. Labour intensive

~ 68 ~
 CONCLUSION

The presented experimental study complements findings from the few published research

papers on the cross media comparison of advertising effectiveness by investigating the

differences between online advertising, print advertising and the combination of these two

channels.

Yet the cross media hypotheses – stating that cross media campaigns ought to be more

effective than print only or online only.

Although it cannot be strictly inferred that combining newspapers and online advertising

positively effects advertising effectiveness, the results supported the observed trend in

business to carry out integrated cross media campaign. With this study it was intended to

expand the knowledge about internet and cross-media advertising effectiveness. These topics

are of interest both for scholars and practitioners, as the internet's possibilities are not

unequivocally considered and are still developing. This renders business opportunities.

Immediate managerial implications come from the fact that cross media advertisement is at

least similar effective than traditional print advertisement

~ 69 ~
 BIBLIOGRAPHY AND REFERENCES

1. Effectiveness of combining online and print advertisements is the whole better than the
individual parts? (Article by Lea M Wakolbinger and Michaela Denk)

2. New media interactive advertising vs. traditional advertising Journal of Advertising Research
(Book by Bezjian-Avery, Alexa, Bobby Calder, and Dawn Iacobucci)

3. Banner advertisements through a new lens (Book by Dahlén, Micael)

4. Effective advertising: Understanding when, how and why advertising works (Book by J G
Tells)

5. Consumer behavior (Book by Schiffmann/Kumar)

6. A Bayesian network modeling approach for cross media

Analysis - (Book by Christina Lakka)

7. Cross Media Promotion (Book by Jonathan Hardy)

8. Advertising and the mind of the consumer (Book by Max Sutherland)

~ 70 ~