You are on page 1of 13

This article was downloaded by: [Universite De Paris 1]

On: 19 May 2013, At: 13:42


Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Journal of Advertising
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/ujoa20

The Influence of Home Page Complexity on Consumer


Attention, Attitudes, and Purchase Intent
a

Gary L. Geissler Ph.D. , George M. Zinkhan Ph.D. & Richard T. Watson Ph.D.
a

College of Business Administration, University of Arkansas

University of Georgia
Published online: 04 Mar 2013.

To cite this article: Gary L. Geissler Ph.D. , George M. Zinkhan Ph.D. & Richard T. Watson Ph.D. (2006): The Influence of
Home Page Complexity on Consumer Attention, Attitudes, and Purchase Intent, Journal of Advertising, 35:2, 69-80
To link to this article: http://dx.doi.org/10.1080/00913367.2006.10639232

PLEASE SCROLL DOWN FOR ARTICLE


Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to
anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should
be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,
proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in
connection with or arising out of the use of this material.

Summer 2006 69

THE INFLUENCE OF HOME PAGE COMPLEXITY ON


CONSUMER ATTENTION, ATTITUDES, AND PURCHASE INTENT
Gary L. Geissler, George M. Zinkhan, and Richard T. Watson

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

ABSTRACT: The Web is an increasingly important component of promotional strategy for many organizations. To date,
most advertising research has focused on banner ads. In the present study, home pages as promotional vehicles are evaluated.
Specifically, we examine the influence of perceived home page complexity on communication effectiveness, as measured by
attention to the home page, attitude toward the home page, attitude toward the company, and purchase intent. Qualitative
research is used to identify important factors that appear to influence perceptions of home page complexity, including
home page length, number of graphics, number of links, amount of text, and use of animation. A pretest experiment
further aided in the selection of appropriate factors and factor levels to manipulate in a subsequent, main experiment. The
findings in the main experiment suggest that consumers respond more favorably toward home pages that fall in a moderate
range of perceived complexity.

The Web is an emerging advertising medium (e.g., Brackett


and Carr 2001; Berthon, Pitt, and Watson 1996; Yoon and
Kim 2001), and organizations are including the Web as part
of their advertising and promotional mixes (Nour and Fadlalla
2000). U.S. Internet advertising expenditures are projected
to reach $15 billion by 2008, surpassing magazine advertising spending ( JupiterResearch 2004). Internet advertising
estimates are comprised of measurable spending (i.e., paid
messages on Web sites, search engines, or other interactive
media, such as e-mail). In light of rapidly evolving media
and associated advertising opportunities, scholars have recently
reexamined the relevance of traditional definitions of advertising and have proposed a more encompassing definition
(Richards and Curran 2002). Researchers have also begun
investigating product Web sites as a form of advertising (e.g.,
Hairong, Daugherty, and Biocca 2002). All Web sites are not
necessarily ads; Web sites may be created for a variety of purposes. Nonetheless, many Web sites are used to inform, persuade, or remind current and potential customers about a
company and its offerings.
The home page creates an initial impression of the
companys Web site and, in some cases, of the company itself
(Singh and Dalal 1999). As with advertising placed in other
Gary L. Geissler (Ph.D., University of Georgia) is an associate professor of marketing and advertising in the College of Business Administration at the University of Arkansas at Little Rock.
George M. Zinkhan (Ph.D., University of Michigan) is a CocaCola Company Professor of Marketing at the University of Georgia.
Richard T. Watson (Ph.D., University of Minnesota) is a J. Rex
Fuqua Distinguished Chair for Internet Strategy and a professor of
management information systems at the University of Georgia.

media, the first look at a Web site is crucial in gaining and


holding consumers attention and interest (Geissler, Zinkhan,
and Watson 2001). First impressions are also critical in establishing on-line relationships. As part of an integrated marketing communications (IMC) approach, home pages are an
increasingly important relationship marketing tool (Watson,
Zinkhan, and Pitt 2000).
Duncan and Moriarty (1998) proposed a communicationbased model of relationship marketing to illustrate how marketing theory has evolved, particularly in light of advances in
new media and an associated increase in interactivity. Communication, not persuasion, forms the foundation for building interactive, on-line relationships. By emphasizing two-way
communication, sharing information, gathering feedback, listening, and responding to current and prospective customers
and other shareholders become key activities (Duncan and
Moriarty 1998). All of these activities can be pursued via a
Web site. Researchers have recently suggested, Web advertisers must make decisions about the effectiveness of their
Web site as a communications medium or of a Web-based ad
campaign (Bhat, Bevans, and Sengupta 2002, p. 1).
In the present study, we focus on the information-sharing
aspect of the communication-based marketing model, which
suggests that sharing information can strengthen relationships. Specifically, we examine the amount and type of information that should be included on a home page to facilitate
communication effectiveness. Undoubtedly, the various elements included on a home page play a pivotal role in either
enticing users into the site or driving them away. Advertisers
must strike a balance between providing too much information (and too many graphics) upfront that may overwhelm,
confuse, or frustrate users and not providing enough of either
Journal of Advertising, vol. 35, no. 2 (Summer 2006), pp. 6980.
2006 American Academy of Advertising. All rights reserved.
ISSN 0091-3367 / 2006 $9.50 + 0.00.

70 The Journal of Advertising

FIGURE 1
Predicted Relationship Between Home Page
Complexity and Communication Effectiveness

Communication
Effectiveness

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

Home Page Complexity


to stimulate sufficient user interest. A key research question
is, How complex (or simple) should a home page be to facilitate effective communication? Our objectives are to (1) determine which elements of home page design (e.g., animation,
page length) are the most influential in terms of triggering
perceptions of home page complexity; (2) conduct a rigorous
test of Berlynes theory of stimulus complexity; (3) examine
the relationships between home page complexity perceptions
and user attention, attitudes, and purchase intent; and (4)
systematically develop a general set of guidelines concerning
elements that Web designers can use to influence both perceived complexity and communication effectiveness.
STIMULUS COMPLEXITY:
A THEORETICAL FRAMEWORK
Defining stimulus complexity or, more specifically, advertising complexity is a difficult task. Borrowing from the psychology literature, we feel that the work of Berlyne (1960) is
most relevant in addressing this key issue. Thus, Berlynes
theory of stimulus complexity is adopted as a theoretical framework. Berlyne acknowledged that complexity is, without any
doubt, the most impalpable of four elusive concepts (the others are novelty, uncertainty, and conflict). Yet complexity
can be roughly defined as the amount of variety or diversity
in a stimulus pattern (Berlyne 1960, p. 38).
More specifically, stimulus complexity depends on several
properties, including
Other things being equal, complexity increases with
the number of distinguishable elements
If the number of elements is held constant, complexity
increases with the dissimilarity between elements
Complexity varies inversely with the degree to which
several elements are responded to as a unit (Berlyne
1960, pp. 3839)

Various research studies have supported Berlynes theory (e.g.,


Day 1967; Morrison and Dainoff 1972).
The theory predicts an inverse, curvilinear relationship
between home page complexity and measures of communication effectiveness (see Figure 1). That is, for optimal effectiveness to be achieved, a home page must surpass a certain
minimal level of complexity but not be too complex. Thus,
home page complexity influences how people feel about an ad
and the advertised company (i.e., their attitude toward each),
as well as their purchase intent.
Two recent studies that also use Berlynes theoretical framework (Bruner and Kumar 2000; Stevenson, Bruner, and Kumar
2000) have indicated that Web page complexity may significantly influence consumer attitudes and intentions, even when
the Web page serves merely as a background for a digitized
commercial. In the prior experiments, complexity was manipulated at three levels. The experimental treatments included three different Web pages on which a state lottery
commercial, filling only a portion of the screen, was run. The
lowest level of complexity was simply a black screen. The
next treatment had a light blue background and the phrase
Play the lottery. The highest level of complexity was a Web
page that also had a light blue background and animation in
the form of two blinking phrases (Buy now and Play the
lottery), as well as a sports car racing across the screen at
different intervals. The two lower complexity backgrounds
were not seen as significantly differenteach positively influenced various dependent variables, including attitude toward the Web site, attitude toward the ad, attitude toward
the brand, and purchase intention. The most complex treatment had a negative effect on the dependent variables. The
research also indicated that attitude toward the Web site has
a significant positive influence on the advertising hierarchy
of effects, especially attitude toward the ad (Bruner and Kumar
2000; Stevenson, Bruner, and Kumar 2000).
In several ways, these findings indicate a need to examine
further the role of Web page complexity in influencing consumer attitudes and intentions. Berlynes theory suggests that
home pages in a moderate range of complexity will be most
effective. We examine the influence of perceived complexity
on communication effectiveness for a diverse set of Web home
pages. We believe that the background is just one element of
perception formation and that we will gain increased knowledge by studying perceptions of the home page.
The present studys central hypothesis is based on Berlynes
theory of stimulus complexity:
H1: Consumers will respond more favorably toward home pages
that are in a range of moderate complexity than to home pages
in either a less complex or more complex range.
The specific hypotheses we tested via an experimental design

Summer 2006 71

concerned the relationship between perceived home page complexity and communication effectiveness as follows:
H2: Home pages that are perceived in a range of moderate
complexity will achieve higher levels of communication
effectiveness than home pages in either a less complex or more
complex range on:
a. attention to the home page (ATThp)

TABLE 1
Web Usage Classification
Light users: Accessed the Web 1 to 3 times during the past 30
days.
Medium users: Accessed the Web 4 to 14 times during the past 30
days (i.e., at least once per week).
Heavy users: Accessed the Web 15 or more times during the past
30 days (i.e., at least every other day).
Triers: Have accessed the Web before, but not during the past 30
days.

b. attitude toward the home page (Ahp)


c. attitude toward the company (Ac)
d. purchase intent (PI)

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

PRELIMINARY RESEARCH
Overview
Although published guidelines pertaining to Web home page
design abound (e.g., Bucy et al. 1999; Hoque and Lohse 1999;
Kassaye 1997; Nielsen 1999; Woodbeck 1995), they often
rely on opinion and personal observation or experience and
not on empirical research (DAngelo and Little 1998). We
used a systematic approach to provide a basis for deciding
which factors and levels contribute to perceived home page
complexity. Extensive qualitative research and a pretest experiment were conducted. The qualitative research included
4 focus group discussions with Web users and triers, and
20 in-depth telephone interviews and 10 personal interviews
with Web site designers. This research helped identify potentially important factors influencing perceived home page
complexity, which were tested in a fractional-factorial, pretest experiment.
Qualitative Research
The focus groups relevant to the present investigation included
three groups of Web users (light, medium, and heavy users)
and one group of Web triers that had to meet Web usage
screening qualifications (see Table 1).
In addition, participants had to be age 18 or older. A mix
of students and nonstudents and a range of ages were recruited.
Approximately half of the participants were male and approximately half were female. Participants were not to have taken
part in another market research study during the prior three
months.
An experienced moderator conducted the discussions, which
were subsequently transcribed for analysis. Emergent themes
were identified and developed after reviews of each transcript.
Preliminary findings were compared across groups to identify common themes and key differences.

Thirty in-depth interviews were conducted among Web


site designers representing virtually every region of the United
States. Twenty interviews were conducted by telephone that
took from 25 to 45 minutes to complete. The remaining 10
interviews were conducted in-person at the designers work
locations and lasted approximately an hour or longer. The
primary screening qualifications were that the respondents
were involved in Web site design and in client discussions of
the needs and objectives for a Web site. Respondents design
experience varied greatly, from 3 sites to over 900 (median = 6
sites). Most of the respondents (16) revealed that they owned
or presided over the firm. The remaining respondents held
various positions, including designers (4), account executives
(3), project managers (3), graphics artists (2), a programmer,
and a salesperson. Responses were reviewed several times, and
common themes were identified and developed.
The discussions concerning home page complexity proceeded logically from a series of questions, including: How
would you describe the ideal home page? What should and
should not be included? What elements determine the complexity of a home page? Web designers, as well as Web users
and triers, all seemed to agree on what one designer referred
to as the home page essentials, which should communicate
who you are, what you offer, and whats inside (the Web site).
Company contact information or a link to a separate contact
page is also considered necessary. Home pages are perceived
as incomplete when designers fail to include all of the essential information up front, and complexity appears to increase
when additional information is included.
The primary elements that designers use to communicate
the essential information (or more or less than the needed or
desired information) include home page length, graphics,
links, text, and animation. It is no coincidence that these same
elements are thought to largely influence perceptions of home
page complexity as well. Both information and aesthetic aspects of a home page are important, and visual complexity is
part of any evaluation of home page complexity. Much online behavior is goal-directed, however, and users actively process information. The qualitative findings strongly indicate
that users are more interested in home pages that provide
essential information than in those that emphasize aesthetics.

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

72 The Journal of Advertising

Many users seemed to prefer relatively simple home pages


compared with more elaborate, aesthetic designs.
Consistent with Berlynes theory of stimulus complexity,
each of the five elements is evaluated both individually and
together in terms of the perceived complexity. The names of
the links indicate what detailed information is contained
within the Web site. Sharing information about who we are
and what we offer via text and various home page lengths is
self-evident. Graphics and animation also communicate information about the company and its brands. For example,
the pretest and main experiments included some treatments
showing a photograph of the business. This provides information about who we are that is useful for locating and
visiting the store and for forming an image of the store, even
without visiting in-person. All of the home page elements
combine to provide essential information and to form a central message, which should be consistent with other messages
communicated by the firm.
The qualitative research identified important factors that
appear to affect perceptions of home page complexity, including home page length, number of graphics, number of links,
amount of text, and use of animation. Although this list of
influential factors is not exhaustive, the findings indicate that
these five factors may be more important than others, such as
the use of color. Preliminary research also aided in the selection of appropriate factor levels to manipulate in the subsequent experiments, as described in the following section.
Pretest Experiment
A 251 between-subjects, fractional-factorial design with a total
of 16 cells was used in the pretest experiment. Each of the
five factors identified in the qualitative research was manipulated in the pretest experiment at two levels: amount of home
page text (less than 25 words or more than 100 words), number of home page graphics (1 graphic or 7 graphics), number
of home page links (1 link or 9 links), length of the home
page (1 page or 3 pages), and use (or nonuse) of animation.
The sample included 169 undergraduate business students.
A fairly even split of male (53%) and female (47%) participants was recruited. Virtually all (99%) of the participants
had used the Web before, averaging just over 12 times during the prior 30 days. Participants were randomly assigned to
treatments, and were asked to review and evaluate a home
page (preset on each computer screen) for a company that sells
music CDs, cassettes, and other audio recordings. Each participant evaluated 1 of 16 different versions of the home page.
Prior to the pretest, the stimulus materials were evaluated by
experienced Web users to ensure realism.
While it is feasible to measure complexity either subjectively based on perceptions or objectively (e.g., number of
bytes per computer screen) to test Berlynes theory, we chose

to use a perceived complexity measure in the pretest and in


the main experiment. Ultimately, advertising and design face
challenges from the reality of consumer perception of home
page complexity. Although bytes per screen and perceived
complexity per home page version were significantly correlated in both the pretest (r = .70, p < .001) and the main experiment (r = .69, p < .001), such a correlation may not be
evident in other situations. That is, perceptions of home page
complexity will likely vary by different groups of target consumers, while the objective measure may not vary in the same
way.
As a starting point, we reviewed a multi-item, seven-point
scale (Cronbachs = .67) used to evaluate the complexity of
print ads (Leigh 1984), since the qualitative research indicated that consumers felt that Web advertising was most similar to print advertising. Based on the focus group findings,
we omitted some items due to their similarity to others or to
their lack of relevance to home page complexity, and added
other items. The modified scale, which used endpoints labeled as does not describe at all and describes very well,
consisted of 12 items: varied, complex, familiar, dense, surprising, patterned, interactive, common, frustrating, crowded,
overwhelming, and navigable. The revised scale had a
Cronbachs of .56. Items were omitted based on a factor
analysis, improving the reliability estimate to .75. The final
complexity scale includes the following four items: complex,
dense, interactive, and varied. Exploratory factor analysis extracted only one principal component, thus demonstrating
that the scale is unidimensional. Also, confirmatory factor
analysis (CFA) using the AMOS Structural Equations Modeling program was then performed. CFA indicated that the scale
items are an appropriate measure of complexity relative to
the other multi-item scales. The final scale items appear to fit
well with Berlynes definition of stimulus complexity. That
is, they seem to relate to both the number of distinguishable
elements (e.g., dense) and the dissimilarity between elements
(e.g., varied).
Results from an analysis of variance (ANOVA) (F = 7.48,
df = 15, p < .01) and interaction plots suggested that number of graphics, number of links, and home page length were
the key drivers of home page complexity. That is, these three
home page design elements (including their interaction terms)
are the most influential (among those pretested) in terms of
perceived complexity. The lack of significance of the text variable may have been due to an experimental design constraint
(i.e., no more than a paragraph of text could be included in
the one-screen manipulation). A larger block of text may have
had more influence on perceived complexity. The text variable was excluded from the main experiment, however, because the design constraint was still present. By omitting
amount of text and use of animation from further examination, more levels of the remaining factors could be manipu-

Summer 2006 73
TABLE 2
Web Usage: Experimental Subjects Versus U. S. Population
Experimental sample

U. S. population

97%

71%

6%

5%

28%

38%

66%

56%

Total Internet/Web users


Internet/Web Usage Categories:*
Relatively inexperienced/
light users
Moderately experienced/
medium users
Very experienced/
heavy users

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

Source: UCLA Internet Report: Surveying the Digital Future, Year Three, UCLA Center for Communication Policy (January 2003), available at
(www.digitalcenter.org/pdf/InternetReportYearThree.pdf).
* U.S. population Web usage categories are defined as follows:
Relatively inexperienced users: Less than one year on-line
Moderately experienced users: One year to less than four years on-line
Very experienced users: Four or more years on-line

lated in the main experiment in an attempt to obtain a greater


range of complexity via a reasonable number of home page
versions.
MAIN EXPERIMENT
Selection of Participants
Samples of college students have been used by other researchers studying Internet advertising, primarily because students
generally tend to be experienced and frequent Web users (e.g.,
Brackett and Carr 2001; Gallagher, Foster, and Parsons 2001;
Stevenson, Bruner, and Kumar 2000). As adoption of the Web
continues, student samples are more likely to include experienced users than are samples of the general population. Differences in Web experience between student samples and the
general population anticipate the direction in which Web
perceptions are moving. While the proportion of people who
use the Internet at home (59.3%) and at work (51.2% of employed) continues to increase, a greater and rapidly increasing
proportion of students (73.7%) use the Internet at school.
Overall, the proportion of Americans who use the Internet (at
home, work, or school) increased from 67% in 2000 to 71%
in 2002 (UCLA Internet Report 2003). Thus, university student samples should continue to yield meaningful results in
the context of understanding Web user behavior into the foreseeable future (Gallagher, Parsons, and Foster 2001).
The present study included primarily Web users (97%)
who exhibited a broad range of Web usage (see Table 2). Participants reported accessing the Web from 0 to 225 times
during the past 30 days. The average number of times participants had accessed the Web during the past 30 days was

26; the median was 20 times. Although the proportion of


Web users in our sample is greater than that in the U.S. population (71%), we feel that it is important to include a large
proportion of Web users when evaluating the specific design
elements and communication effectiveness of home pages.
Also, Web usage among Americans (at home, work, or school)
will likely become even more pervasive in the future. For instance, it is predicted that Web penetration will eventually
reach levels similar to the high penetration of telephones (96%)
and televisions (98%) among U.S. households (Nielsen Media Research 2004; U.S. Census Bureau 2003).
Web usage can be defined and measured in numerous ways,
including number of times the Web is accessed during a specified time period (as in this study), number of years on-line,
and average number of hours on-line during a specified time
period. Research has shown that the measures are related. For
example, users time on-line increases with each additional
year of Internet experience (UCLA Internet Report 2003).
Although Web usage is measured differently, the Web user
categories in our sample (see Table 1 for Web usage classification) are roughly comparable to those among the general population of Americans.
For certain products, college students represent a key target market segment, and thus provide an appropriate sample.
We sampled undergraduate business students at a major southeastern university and asked them to evaluate different versions of a home page for a company that sells music cassettes,
CDs, and vinyl records. Company familiarity was not a confounding factorless than 1% of the participants claimed to
have heard of the company. Along with books and computers,
music is one of the most popular products among consumers
who make purchases on the Internet (Kwak, Fox, and Zinkhan

74 The Journal of Advertising


TABLE 3
Range of Manipulated Dimensions
Number
of links

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

1
5
9
13

Home page
length (number
of screens)
1
2
3

ing, and after the experiment. No time limitations were set


on any part of the process, which took from 10 to 15 minutes
to complete.

Number of
graphics
1
4
7

2002). A total of 360 students completed the experiment.


Just over half (52%) of the sample was female. Although
Internet usage continues to increase among all age and gender groups, our sample dimensions are similar to the age
and gender specifications for the highest Internet usage
groupsage 35 years and under (with especially high use among
those 24 and under), with slightly more men than women
(UCLA Internet Report 2003). A drawing for cash prizes (one
$100 prize and three $50 prizes) was held as an incentive for
participation.
Procedures
A total of 36 cells were created in a 3 3 4 between-subjects, full-factorial design. Two factorsnumber of home page
screens and number of graphics (other than the company
logo)were each manipulated at three levels. The third factornumber of linkswas manipulated at four levels (see
Table 3).
The stimulus materials were designed using Netscape
Navigator Composer, based on the saved home page of an
operational company. The home page design was manipulated
to generate the desired versions for testing. The various versions were assessed by a group of Web experts to ensure that
they were appropriate. Needed adjustments were completed
before testing. (The stimulus home pages can be viewed at
www.richardtwatson.com/experiment/.)
Subjects were given an oral introduction to the study when
they arrived at the computer laboratory. They were asked to
refrain from talking to each other while completing the experiment. Written instructions informed the participants that
they should take a few minutes to review the home page. They
reviewed and evaluated only 1 version of a home page from
the set of 36. Subjects were instructed to scroll up and down
the page and to refer to the home page when evaluating its
complexity. Nonworking links were used, since the focus was
on evaluating the home page itself. Subjects were then instructed to turn off their computer screens and to rate the
home page on several dependent variable measures. Classificatory information was also gathered.
Subjects were encouraged to ask questions prior to, dur-

Measures
The survey instrument primarily contained previously developed scales, modified when needed based on the pretest, to
measure perceived complexity (as described above) and several key dependent variables: attention to the home page
(ATThp), attitude toward the home page (Ahp), attitude toward the company (Ac), and purchase intent (PI). ATThp is a
self-report measure of the degree to which a person perceives
that a home page gained and held attention. Ahp is based on
an attitude toward the ad scale. Given the relative lack of Ac
scales in the marketing literature, the Ac scale was adapted
from an attitude toward the brand (Ab) scale, which measures
similar items. PI is a scale developed to measure purchase
intent of the products offered on the home page. The dependent variable scales are unidimensional, as evidenced by a series of factor analyses. All of the scales yielded acceptable
reliabilities, as reported in the Appendix showing the scale
origins and items.
Findings
Analysis of covariance (ANCOVA) revealed that Web usage
was not a significant covariate (F ratio = 3.20, p < .05). In
other words, the results do not vary across the levels of reported Web usage. This is possibly because students represent a relatively experienced and similar group of Web users.
The effects of specific, home page design elements on perceived complexity were determined by ANOVA (see Table
4). This analysis confirmed that the complexity manipulation worked. The overall model was significant (F = 3.54,
df = 35, p < .001). That is, number of graphics, number of
links, and home page size influence perceived complexity.
Although all of the main effects were significant, there was
also a significant interaction.
Perceived complexity tends to increase as the number of
links increases. Examining the mean complexity scores for
number of links in isolation was appropriate, since the main
effect for the number of links was significant and there was
no significant interaction with other variables. The interaction between the number of graphics and the length of the
home page significantly influenced ( p < .01) the perceived
complexity of the home page. Most notably, the one-screen
manipulation was seen as more complex than the two- or threescreen manipulations at each graphics level. Specifically, the
home page was viewed as most complex when it had only one
screen and four or seven graphics. It was moderately complex
when it had only one screen and one graphic, or when it had

Summer 2006 75

FIGURE 2
Proposed Model of Home Page
Complexity and Communication Effectiveness

TABLE 4
ANOVA Results
F value
Main effects
Links
Graphics
HPlength

Home Page Design


Elements:

6.17**
12.09**
19.42**

Two-way interactions
Link graphics
Links HPlength
Graphics HPlength

 Number of graphics
 Number of links
 Home page length

1.35
1.30
3.76*

Three-way interaction
Links graphics HPlength

.96

Notes: ANOVA = analysis of variance; HPlength = home page length.

Perceptions of
Home Page Complexity

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

** p < .01.
** p < .001.

seven graphics spread over two screens. The home page was
perceived as least complex when it had one or four graphics
spread over two screens or when it had three screens, regardless of the number of graphics.
We were interested in perceived differences between home
pages, so this was the unit of analysis. We expected to find
that a home pages complexity score in the moderate range
would be associated with greater communication effectiveness. In other words, an inverse (i.e., negative) curvilinear relationship between perceived complexity and communication
effectiveness was predicted. A linear regression model was not
significant in any of the tests. The appropriate regression model
is a quadratic, where the average ATThp, Ahp, Ac, and PI score
for each home page are dependent variables and the average
perceived complexity score constitutes as the independent variable. The results are summarized in Table 5.
Overall, the preponderance of statistical evidence supports the
central hypothesis (H1). Consumers respond more favorably toward home pages perceived in a range of moderate complexity
than those in either a less complex or more complex range.
A proposed model of home page complexity and communication effectiveness is presented in Figure 2. The model shows
the home page elements (i.e., number of graphics, number of
links, and home page length) found to have the greatest influence (among those tested) on perceptions of home page complexity. Perceived complexity, in turn, influences all of the
measures of communication effectiveness (i.e., ATThp, Ahp, Ac,
and PI) in a curvilinear manner, as predicted by Berlyne.
DISCUSSION
The level of home page complexity should be a key consideration for Web site designers and advertisers. Our study sup-

Measures of Communication
Effectiveness:





Attention to the home page


Attitude toward the home page
Attitude toward the company
Company purchase intent

ports the notion that home page complexity influences consumer attention, attitudes, and purchase intent. By examining a wide range of home page complexity, we demonstrate
that an optimal zone of moderate complexity exists, which
facilitates communication effectiveness. To reach maximum
effectiveness, home pages must go beyond some minimal range
of complexity, but avoid exceeding an optimal moderate range.
While the term zone implies that a multitude of combinations of home page elements can be used to achieve moderate
home page complexity, the solution is to present enough information and graphics, but not too much of either. Certainly,
some important information and attractive graphics on the
home page have a positive influence on consumer attitudes
and purchase intent. Too much information and too many
graphics have a negative effect, however, as consumers may
feel overwhelmed or lost, which may cause them to lose focus
and interest rapidly.
A critical issue when determining an appropriate level of
home page complexity is whether the home page is the ad or
simply the background for an ad. Berlynes theory suggests

76 The Journal of Advertising


TABLE 5
Quadratic Regression Model Results
Hypothesis
H2a
H2b
H2c
H2d

Dependent
variable

Quadratic
model significance

Variance
explained

Quadratic term
significance

Supported

Average ATThp score per page


Average Ahp score per page
Average Ac score per page
Average PI score per page

p < .001
p < .001
p < .005
p < .05

.335
.525
.271
.209

p < .001
p < .001
p < .005
p < .05

Yes
Yes
Yes
Yes

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

Note: ATThp = attention to the home page; Ahp = attitude toward the home page; Ac = attitude toward the company; PI = purchase intention.

that relatively complex stimuli draw more attention when


they are presented in the context of simpler stimuli. When
the home page is just a background for the ad, the page should
be quite simple and nondistracting, as previous research has
indicated (Bruner and Kumar 2000; Stevenson, Bruner, and
Kumar 2000). When the home page itself is the ad, the present
study indicates that moderate complexity is most desirable.
The appropriate combinations of elements used to create a
moderately complex page will likely depend on several factors, including the company type and the associated competitive advertising environment. Still, we can offer advertisers
some general guidelines for home page design, based on the
combined findings from the present study and from our pretest and qualitative research. Our work indicates that consumers appreciate the provision of numerous, working links
that help them navigate a Web site, and organizations should
consider having multiple links on the home page. Although
the complexity of a home page increases as the number of
links increases, a reasonable number of links alone should not
cause the home page to exceed a moderate range of complexity. For example, we included up to 13 links in our experiment, without exceeding moderate complexity (relative to
mean scores associated with the home page length and graphics effects). At some point, an excessive number of links will
begin to crowd the screen and overwhelm the user.
Also, advertisers and designers should be careful concerning the manner in which they combine the length of the home
page with the number of graphics (other than the company
logo). If a one-screen home page is designed, it is likely that
only one graphic other than the logo is needed to achieve
moderate complexity. If additional graphics are used, consideration should be given to spreading them over a two-screen
home page. The findings indicate that as the length of the
home page increases beyond two screens, it becomes more
difficult to surpass a minimal range of complexity.
While these general guidelines provide a useful starting
point, Web advertisers should certainly consider testing several versions of a given home page to ensure that it is designed to operate within a zone of moderate complexity.
Complexity is best viewed as a relative measure. In other words,

factors such as product category and target consumer segment


will somewhat influence what consumers perceive to be a low,
moderate, or high complexity range. Nevertheless, in future
studies, researchers could consider measuring both objective
complexity (e.g., bytes per computer screen) and perceived
complexity among different groups of target consumers to
better understand how the two measures are related and when
they differ significantly.
LIMITATIONS AND CONCLUSIONS
Given the relatively high level of Web experience among college students and sample size requirements, we used undergraduate business students. Student samples have yielded
meaningful results in previous research examining Web usage behavior (Gallagher, Parsons, and Foster 2001). We selected a product for which students represent a key target
market segment, and thus the experimental task should have
been salient. A companys Web strategy, including the form
and content of the home page, should be directly influenced
by the target market (Nour and Fadlalla 2000). Perceptions
of home page complexity may vary, according to the target
audience. Students are not the only target market for music
products, however, and the subjects may have demographic
and Web usage characteristics that differ from other consumers. For example, it is likely that younger, heavy Web users
would not be as impressed and overwhelmed by animated
graphics as older, light Web users. Future studies could include other product categories.
Although we conducted extensive pretesting to identify
important factors influencing perceptions of home page complexity, the list of treatment variables used in this study is
limited. Other factors that may affect perceived complexity
of a home page were omitted from the study. For example,
the download time of the graphics was not a factor, since the
home pages were cached in advance. In any event, download
time should play a diminished role in the near future, as bandwidth increases. Also, other researchers have included color
manipulations in an attempt to influence perceptions of Web
page complexity. However, color had no significant influence

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

Summer 2006 77

on ad effectiveness (Bruner and Kumar 2000; Stevenson,


Bruner, and Kumar 2000). Also, the results of our qualitative
research indicated that color was not an important factor in
triggering perceived Web page complexity.
Screen size influences the degree of scrolling required to
view all the information on a home page. In this study, screen
size was constant for all participants and its influence on complexity was therefore not discernible. Computer screens are
only one form of visual display, however. Screen designers for
personal digital assistants (e.g., Palm), mobile phones, car
navigation systems, portable music players (e.g., iPod), and
other information appliances, working with a smaller display
area, also face the problem of home page complexity. Thus, a
fruitful avenue for future research is to assess the generalizability of our findings to devices of quite different dimensions, used in different modes (e.g., walking around instead
of sitting down), and with different scrolling mechanisms (e.g.,
joystick versus mouse).
The power of the experiment to detect true differences is
most evident in the main effects. When individual cells and
associated, specific combinations of design elements are examined, there is considerable loss of power due to the relatively small sample size. However, examining the effects and
interactions across several cells (effectively combining individual cell samples) provides a large enough sample to evaluate broader issues, such as whether it is better to have more or
fewer graphics and links (versus an exact number or combination) relative to various home screen lengths.
Although the relationships between the dependent variables were not a focus of this study, preliminary regression
analyses provide support for a communication hierarchy of
effects. Future research could include a more rigorous test,
using structural equation modeling, of the relationships between the dependent variables within the context of the proposed model.
Our study makes several contributions. First, the qualitative research and experimental pretest helps to identify specific home page design features that are associated with
perceived complexity. We find that the five most influential
elements of home page complexity are number of graphics,
number of links, length of the home page, use of animation,
and amount of text.
Second, previous studies suggest that simple Web page
backgrounds generate the most favorable consumer response.
Intuitively, it makes sense that relatively simple backgrounds
would be viewed more positively, since such backgrounds are
less obtrusive and less likely to compete directly with core
messages. However, this finding does not hold when the Web
page itself is the advertisement, rather than just a background
for a commercial.
Third, previous research examined only three levels of complexitysimplest, more complex, and most complex. The

researchers stated that the relationship between background


complexity and attitude toward the Web site was viewed as
positive at low levels of complexity but . . . [became] negative at high levels. The relationship that is found in a study
depends upon the actual complexity levels that are measured
(Bruner and Kumar 2000, p. 39). Indeed, we contend that
numerous levels of complexity are needed to provide a more
rigorous test of the relationship. An exact application of
Berlynes theory tests both a linear and a curvilinear model.
Via testing a wide range of Web home pages, our study provides better direction for Web site designers and advertisers.
In summary, this study contributes to the growing body of
research investigating the influence that various elements of
Web advertising have on communication effectiveness. Specifically, the findings demonstrate an inverse, curvilinear relationship between home page complexity and attention to the
home page, attitude toward the home page, attitude toward
the company, and purchase intent. By examining a wide range
of home page complexity, we also provide support for Berlynes
theory of stimulus complexity, as it applies to home pages.
Overall, the old adage everything in moderation is a useful
guideline to consider for the design of effective home pages.
REFERENCES
Berlyne, Daniel E. (1960), Conflict, Arousal, and Curiosity, New
York: McGraw-Hill.
Berthon, Pierre, Leyland F. Pitt, and Richard T. Watson (1996),
The World Wide Web as an Advertising Medium: Towards an Understanding of Conversion Efficiency, Journal
of Advertising Research, 36 (1), 4354.
Bhat, Subodh, Michael Bevans, and Sanjit Sengupta (2002),
Measuring Users Web Activity to Evaluate and Enhance
Advertising Effectiveness, Journal of Advertising, 31 (3),
97106.
Brackett, Lana K., and Benjamin N. Carr, Jr. (2001), Cyberspace
Advertising Vs. Other Media: Consumer Vs. Mature Student Attitudes, Journal of Advertising Research, 41 (3), 23
32.
Bruner, Gordon C., and Anand Kumar (2000), Web Commercials and Advertising Hierarchy-of-Effects, Journal of Advertising Research, 40 (1), 3542.
Bucy, Erik P., Annie Lang, Robert F. Potter, and Maria Elizabeth
Grabe (1999), Formal Features of Cyberspace: Relationships Between Web Page Complexity and Site Traffic, Journal of the American Society for Information Science, 50 (13),
12461256.
DAngelo, John, and Sherry K. Little (1998), Successful Web
Pages: What Are They and Do They Exist? Information
Technology and Libraries, 17 (2), 7181.
Day, H. (1967), The Evaluation of Suggestive Complexity, Pleasantness, and Interestingness for a Series of Random Polygons Varying in Complexity, Perception and Psychophysics, 2
(7), 281286.
Duncan, Tom, and Sandra E. Moriarty (1998), A Communica-

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

78 The Journal of Advertising

tion-Based Marketing Model for Managing Relationships,


Journal of Marketing, 62 (2), 113.
Gallagher, Katherine, K. Dale Foster, and Jeffrey Parsons (2001),
The Medium Is Not the Message: Advertising Effectiveness and Content Evaluation in Print and on the Web,
Journal of Advertising Research, 41 (4), 5770.
, Jeffrey Parsons, and K. Dale Foster (2001), A Tale of
Two Studies: Replicating Advertising Effectiveness and
Content Evaluation in Print and on the Web, Journal of
Advertising Research, 41 (4), 7183.
Geissler, Gary L., George M. Zinkhan, and Richard T. Watson
(2001), Web Home Page Complexity and Communication Effectiveness, Journal of the Association for Information
Systems, 2 (2), 146.
Grossbart, Sandford, Darrel D. Muehling, and Norman Kangun
(1986), Verbal and Visual References to Competition in
Comparative Advertising, Journal of Advertising, 15 (1),
1023.
Hairong, Li, Terry Daugherty, and Frank Biocca (2002), Impact of 3-D Advertising on Product Knowledge, Brand
Attitude, and Purchase Intention: The Mediating Role of
Presence, Journal of Advertising, 31 (3), 4357.
Hoque, Abeer Y., and Gerald L. Lohse (1999), An Information
Search Cost Perspective for Designing Interfaces for Electronic Commerce, Journal of Marketing Research, 36 (3),
387394.
JupiterResearch (2004), Marketing and Branding Forecast: Online
Advertising and E-mail Marketing Through 2008, available at www.jupiterresearch.com (accessed July 16, 2005).
Kassaye, W. Wossen (1997), Global Advertising and the World
Wide Web, Business Horizons, 40 (3), 3342.
Kwak, Hyokjin, Richard J. Fox, and George M. Zinkhan (2002),
What Products Can Be Successfully Promoted and Sold via
the Internet? Journal of Advertising Research, 42 (1), 2338.
Leigh, James H. (1984), Recall and Recognition Performance
for Umbrella Print Advertisements, Journal of Advertising,
13 (4), 518, 30.
MacKenzie, Scott B., Richard J. Lutz, and George R. Belch
(1986), The Role of Attitude Toward the Ad as a Mediator of Advertising Effectiveness: A Test of Competing Explanations, Journal of Marketing Research, 23 (May),
130143.

Madden, Thomas J. (1982), Humor in Advertising: Applications of a Hierarchy of Effects Paradigm, Ph.D. diss., University of Massachusetts.
Mitchell, Andrew A., and Jerry C. Olson (1981), Are Product
Attribute Beliefs the Only Mediator of Advertising Effects
on Brand Attitude? Journal of Marketing Research, 18 (3),
318322.
Morrison, Bruce J., and Marvin J. Dainoff (1972), Advertisement Complexity and Looking Time, Journal of Marketing
Research, 9 (4), 396400.
Nielsen, Jakob (1999), User Interface Directions for the Web,
Communications of the ACM, 42 (1), 6572.
Nielsen Media Research-NTI (2004), TV Basics: Television
Households, available at www.nielsenmedia.com (accessed
July 24, 2005).
Nour, Mohamed A., and Adam Fadlalla (2000), A Framework
for Web Marketing Strategies, Information Systems Management, 17 (2), 4150.
Richards, Jef I., and Catharine M. Curran (2002), Oracles on
Advertising: Searching for a Definition, Journal of Advertising, 31 (2), 6377.
Singh, Surendra N., and Nikunj P. Dalal (1999), Web Home
Pages as Advertisements, Communications of the ACM, 42
(8), 9199.
Stevenson, Julie S., Gordon C. Bruner, and Anand Kumar (2000),
Webpage Background and Viewer Attitudes, Journal of
Advertising Research, 40 (1), 2934.
UCLA Internet Report: Surveying the Digital Future, Year
Three (2003), UCLA Center for Communication Policy
(January), available at www.digitalcenter.org/pdf/
InternetReportYearThree.pdf.
U.S. Census Bureau (2003), Current Population Survey, available at www.census.gov (accessed March 4, 2004).
Watson, Richard T., George M. Zinkhan, and Leyland F. Pitt
(2000), Integrated Internet Marketing, Communications of
the ACM, 43 (6), 97102.
Woodbeck, Dean (1995), Making the Web Work for You,
Currents, 21 (10), 4045.
Yoon, Sung-Joon, and Joo-Ho Kim (2001), Is the Internet
More Effective Than Traditional Media? Factors Affecting
the Choice of Media, Journal of Advertising Research, 41 (4),
5360.

Summer 2006 79

APPENDIX
Multi-item Scales Used in the Main Experiment
Complexity, Cronbachs = .75 (adapted from Leigh 1984)
Does not
describe
at all

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

Describes
very well
2

1. Complex
2. Dense
3. Interactive
4. Has much variety

Attention to the home page (ATThp) (seven-point semantic differential scale), = .83 (adapted from Madden 1982)
1. At gaining my attention, the Web home page was: not effective/effective.
2. At holding my attention, the Web home page was: not effective/effective.
3. During my review of the Web home page, I was: not alert/alert;
unemotional/emotional; not observant/observant; aimless/motivated.

Attitude toward the home page (Ahp) (seven-point semantic differential scale), = .94 (adapted from MacKenzie,
Lutz, and Belch 1986; Mitchell and Olson 1981)
1.
2.
3.
4.
5.
6.
7.
8.

Uninformative/Informative
Unfavorable/Favorable
Incomplete/Complete
Unlikable/Likable
Poorly structured/Well-structured
Dull/Exciting
Not entertaining/Entertaining
Uninteresting/Interesting

Attitude toward the company (Ac) (seven-point semantic differential scale), = .94 (adapted from Grossbart,
Muehling, and Kangun 1986)
1. Bad/Good
2. Poor quality/High quality
3. Unfavorable/Favorable
4. Inferior/Superior
5. Negative/Positive

80 The Journal of Advertising

Purchase Intent (PI), = .77


Very unlikely
to buy
1

Very likely
to buy
2

Downloaded by [Universite De Paris 1] at 13:42 19 May 2013

1. If you were to buy a music cassette tape during the next three months, how likely or unlikely would you be to buy it from the
company shown on the Web home page?
2. If you were to buy a music CD during the next three months, how likely or unlikely would you be to buy it from the company
shown on the Web home page?

You might also like