 |
Internet Marketing · eMarketing · Internet Advertising · Online Branding |
|
Internet Marketing · eMarketing · Internet Advertising · Online Branding |
|
|
|
||
| ADVERTISING BY ICONOCAST |
||
Advertising Pricing Models for the World Wide Web
Donna L. Hoffman and Thomas P. Novak
1
Advertising Pricing Models for the World Wide Web
Donna L. Hoffman and Thomas P. Novak
Novak, T.P. and D.L. Hoffman (2000), "Advertising and Pricing Models for the Web," in Internet
Publishing and Beyond: The Economics of Digital Information and Intellectual Property, Deborah
Hurley, Brian Kahin and Hal Varian, eds. Cambridge: MIT Press.
Introduction
The advertiser supported Web site is one of several business models vying
for legitimacy in the emerging medium of the World Wide Web on the Internet
(Hoffman, Novak, and Chatterjee 1995). Advertising on the Web takes on a
number of forms. Currently, the two most dominant forms of advertiser-supported
Web sites include sponsored content sites such as Hotwired, CNN, and Salon and
entry portal sites (for example, Yahoo, Netscape, and Excite), that function as
gateways to the Web and provide search and directory features to Web browsers.
The sponsorship model has attracted management attention because
advertising is expected to be an increasingly significant source of growing levels
of revenues in the new medium of the World Wide Web (Rebello 1996).
Sponsored sites are of interest because they are well suited to the Web
environment (Hoffman & Novak 1996), yet also retain important parallels to
existing media in the physical world. In theory, institutional advertising practices
and metaphors may be borrowed from traditional media environments to assist
initial commercial efforts. Additionally, although the online storefront model is
beginning to take off, many Web managers are hedging their bets by relying on
advertising revenue streams as a source of profitability for on line ventures.
elab.vanderbilt.edu
2
Against this backdrop, firms are trying to understand the elements that
make a sponsored site successful. As advertisers and marketers debate the best
ways to measure and to track visits and use on commercial Web sites, most firms
remain largely in the dark about the number of customers for their online
offerings. The Web advertising industry currently lacks standards for measuring
activities and use by potential customers on the Web. As a result, it is having
difficulty envisioning the Web as an advertising medium. There is also no
assurance that firms will be successful in generating significant revenues from
Web advertising in the future, particularly as electronic commerce efforts gain
momentum. Ultimately, we believe that this lack of standardization is likely to
limit the long-term viability of the Web advertising sponsorship model.
The lack of standardization exists on four fronts. First, there are no
established principles for measuring traffic on commercial Web sites that seek to
generate revenues from advertising sponsorship. Secondly, there is no standard
way of measuring consumer response to advertisement. Third, there are no
standards for optimal media pricing models. Finally, the complexity of the
medium in general hinders the standardization process.
From an advertising perspective, the Web medium has some similarity to
radio in that there are many different markets and they are clearly segmented
theoretically. But, the standardization of the radio media buy, that is purchasing
advertising time on a particular program on a particular radio station, eases
considerably the process of advertising in that medium. In contrast, the Web
presents a "nightmare" buy for agencies and their clients. For example, Focalink's
1996 database of over 600 commercial Web sites (Focalink 1996) showed that
3
there are more than 90 sizes for Web advertisement banners, that sites use many
different metrics to price advertising, that there is no consistency in definitions
even among the same or similar metrics, and that consumer demographic
information is virtually nonexistent. In the ensuing two years, the situation has
actually increased in complexity, as new advertising forms like so-called rich
media evolve from static banner ads.
Despite the lack of information in this chaotic emerging environment,
there is no dearth of activity. Total Internet advertising revenues approached $2
billion in 1998 (Forrester Research 1999) and the category appears to be growing
faster than traditional mass media advertising vehicles like television and print.
From a comparatively insignificant $550 million in 1997, the Internet advertising
industry has now logged over two years of revenue growth. Forecasts have online
advertising revenues surpassing $3 billion by 1999 and doubling to $6 billion by
the year 2000 (Internet Advertising Bureau 1999). At this rate, Internet advertising
will exceed billboard advertising expenditures (estimated at over $2 billion) by
the end of 1999 and surpass all forms of outdoor advertising (logging revenues of
over $4 billion) by the turn of the century. Note that total United States
advertising expenditures were forecast to top $285 billion in 1998 (Competitive
Media Reporting 1999).
As industry forces point toward advertising as an increasingly significant
source of revenues in the new online medium and with online shopping revenues
expected to reach almost $40 billion by the year 2002 (Forrester Research 1999),
it is no wonder that the advertising sponsorship business model is attracting
greater management attention, especially in traditional circles.
4
Yet, despite these heady forecasts, there still exists doubt among
advertisers about advertising sponsorship as a business model. This skepticism
may be traced to the fact that few have specified conclusively the precise manner
in which advertising on the Web might and should further a firm's strategic
marketing objectives. Clearly, standardizing the Web measurement process is a
critical first step on the path toward the successful commercial development of the
Web. Without standardization, ad hoc Web advertising pricing models will
continue to hinder the sponsorship model as a legitimate revenue stream.
This chapter examines current practice for advertising pricing models on
the Web and proposes models based on constructs that are arguably more suited to
the Web environment. The policy considerations that are likely to affect the
development of Web advertising standards are then addressed. Finally, the chapter
concludes with some thoughts on the best means to develop optimal Web
advertising pricing models.
Current Web Advertising Pricing Models
There is considerable confusion regarding the terminology currently in use
for Web advertising. The first step is the development of a common vocabulary. If
there is terminology from traditional media that is appropriate to use in the
context of Web-based advertising, it should be employed to avoid confusion and
to ease the adoption process of standards formation.
The current document forms of Web-based advertising are "banner
advertisements" and "target communications." A banner advertisement is a small,
typically rectangular, graphic image that is linked to a target communication.
5
Banner advertisements appear in various sizes, with 90 percent of banner
advertisements ranging from 120 to 500 pixels wide (with a median of 460 pixels)
and from 45 to 120 pixels high (with a median of 60 pixels) (Focalink 1996).
Banner advertisements typically provide little information beyond the
identification of the sponsor and serve as an invitation for the visitor to click on
the banner to learn more.
Target communications, in contrast may be fairly detailed, ranging from a
single Web page with basic HTML to a Web page enhanced by technologies such
as Java applets (a program written in the Java programming language that can be
included in a Web page to enhance functionality), streaming media (real-time
broadcasting of audio or video over the Internet), Shockwave (multimedia
playback over the Internet), or Web fill-out forms, to a series of linked pages, or to
complete corporate "Internet presence," “content," or "online storefront" site
(Hoffman, Novak, and Chatterjee 1995).
Banner advertisements are a primitive type of Web-based advertising and
are not likely to ultimately be the most effective new media form. However, as
the most prevalent form, it is appropriate to discuss Web pricing in this context.
Although additional Web-based advertising efforts will evolve, making more
general recommendations that can encompass other online advertising forms yet
to be developed is difficult.
Chatterjee (1998) considers banner ads to be a form of passive advertising
exposure, in that the consumer does not consciously decide to view the banner
advertisement. Rather, the banner advertisement is presented as an outcome of
accessing a particular Web content page or of entering a series of key words into a
6
search engine. Conventional market segmentation theory would lead to the
prediction that the more targeted the banner advertisement, the higher the click
rate.
Advertisements placed on home pages of general-interest sites or on the
entry page of a search engine would have lower click rates, therefore, than
advertisements that are consistent with the content of a narrowly targeted web site
or banner advertisements that are presented by a search engine in response to
specific keywords (e.g., ads for Lionel trains presented every time a visitor
searches for "model railroad" or for "Neil Young").
Paid links are a different form of passive advertising, and may be most
simply viewed as a text version of a banner advertisement. Paid links are often
incorporated in directories, which may contain large numbers of such paid links.
Chatterjee (1998) considers target communications, on the other hand, to
be a form of active advertisement exposure, since the consumer actively decides
to access the target communication by clicking on the banner advertisement, after
being passively exposed to the banner. Active advertisement exposure is under the
consumer's control; passive advertisement exposure is under the marketer's
control. The distinction between passive advertisement active advertisements
implies a crucial difference, therefore, between banner and target
communications. Furthermore, the concept of an active advertisement is a feature
that differentiates Web advertising from advertising in traditional media.
To date, most of the focus in Web advertising measurement has been upon
banner advertisements, most likely because banners advertisements, by their
passive nature, have many more parallels with traditional media planning than do
7
active advertisements. The factors that affect consumer attention to an
advertisement, so-called perceptual selection, in print media should also influence
perceptual selection of Web banner advertisements. These factors are closely tied
to the creative function in advertising and include size, position, motion, color,
and novelty (e.g. Wilkie 1990), all of which are considered relevant for predicting
the likelihood that a visitor will click on a banner advertisement.
Currently, exposure models, based upon CPMs (cost-per-thousand
impressions) or flat fees applied to site exposure or banner advertisement
exposure, are the dominant approach to Web media pricing. Fees based upon
actual "click-throughs" are also in use, in which the advertiser pays only when
individuals actually click on the banner advertisement with their mouses. These
click-throughs take the viewer to the advertiser's target communication. The
following sections consider exposure models, click through models, and other
possible pricing models. While it is premature to recommend any one media
pricing model, it is important to understand the relative strengths and limitations
of methods currently in use or that have been proposed.
Cost Per Thousand and Flat Fee Exposure Models
Flat fee pricing consists of a fixed price for a given period of time. Flat
fees were the earliest Web advertising pricing model. Flat fee pricing may be
implemented with or without traffic (the amount of individuals who visit a Web
site) guarantees. Naturally, it would be advantageous to the advertiser to request
guarantees of traffic level. The earliest advertising pricing approaches on the Web
simply used flat fees, such as advertisement cost per month, without clear
8
specification of the traffic delivered in that period of time. At a minimum,
accurate information on site traffic must be made available to the advertiser, so
that the advertiser may evaluate alternative Web media vehicles.
Assuming accurate traffic information, flat fee prices may be readily
converted into a CPM (cost per thousand exposures) model. CPMs may also be
enhanced by providing "guarantees" of the number of impressions in a given
period of time. The flat fee and CPM models are interchangeable, if traffic
information, specifying the number of (possibly unique) visitors to a Web site, is
available. If traffic information is not available, then flat fee pricing may still be
used although its value is then impossible to evaluate.
In 1996, ninety percent of CPMs for Web advertising (Focalink 1996)
ranged from $10 to $150, with a median of $60. The average CPM online is
currently $36.6 (Adknowledge 1998). In comparison CPMs for advertising range
from $6-$14 for national television, $8-$20 for magazines, and $18-$20 for
magasines, and $18-$20 for newspapers (I/PRO 1996).
The ultimate challenge will be the identification of the business models
that will be effective in the new Web environment. At present, the advertisersupported
business model is being driven largely by a broadcast paradigm, which
has initially gravitated toward CPMs as the appropriate unit of measure. In this
model, the belief is that exposure-based pricing takes into account different
advertisers' response functions and represents a rational way to price advertising
on the Web.
But, in fact, impression/exposure models go only part of the way because
the Web is different from traditional broadcast media. The Web is based on a
9
many-to-many communication model and traditional media are based on a one-tomany
communication model. In addition to exposure metrics, therefore,
interactivity metrics are also required. The CPM approach places too much
emphasis upon the banner advertisement and essentially no emphasis upon the
target communication, which is the real marketing communication that the
advertiser wishes the visitor to read.
In the CPM model, larger numbers of online visitors translate into Web
sites which are bigger winners because the one to-many model seeks a mass
audience for its message. The dangers of relying solely on exposure models means
that interactive managers will be driven to scale their sites to larger, mass
audiences with more homogeneous tastes, in order to attract more advertising
revenue. This goal conflicts with solving the more difficult problems of how to
measure interactivity and to price advertising according to the value of a
consumer's interactive visit to the advertiser.
CPM and flat fee models do nothing more than simply count the number
of visitors exposed to a particular banner advertisement on a particular site. But,
since consumer behavior on the Web depends upon a whole host of measurable
factors, including the type of site and the consumer's motivation for visiting it
(Hoffman and Novak 1996), a simple tally of visits is not sufficient to
demonstrate to the advertiser the value of their advertising expenditures. It is
meaningless to compare directly between Web sites the number of visitors
exposed to banner advertisements, without also taking these other factors into
account.
10
Models Based on Click-Through
Advertising pricing based upon click-through is an attempt to develop a
more accountable way of charging for Web advertising. The payment for a banner
advertisement is based on the number of times a visitor actually clicked on it. In
1996, the fee was approximately $0.25 per click (I/PRO 1996). Because clickthrough
rates have been dropping, perhaps due to consumer boredom, the fee now
ranges from $0.04 to $0.20 per click (see, for example, ClickQuick 1999).
A relatively small proportion of those exposed to a banner advertisement
actually click on the banner. DoubleClick (1996) reported that 4 percent of Web
site visitors who are exposed to a banner advertisement click on the advertisement
the first time they see it. The top 25 percent performing advertisements in the
DoubleClick Network had an average click rate of 8 percent, with some click rates
as high as 12 to 15 percent. Click-through rates decline after the first exposure,
falling to 2 percent for the second and third exposures, and to 1 percent or less at
four exposures. Payment based upon click-through guarantees, therefore, not only
that the visitor was exposed to the banner advertisement, but also actively decided
to click on the banner and to become exposed to the target communication.
Click-through payment may be viewed as payment for target communication
exposures.
This practice is not without controversy, however. Procter & Gamble was
the first to insist it would pay the Web site (Yahoo, in this case) only for the
click-through by viewers rather than for gross impressions of the banner
advertisements. (Associated Press, 1996). Some Internet publishers continue to
feel that this pricing strategy is unfair, arguing that the click-through is at least
11
partially a function of the level of creativity of the advertisement and the level of
interest generated in the viewer by it, which are not under the publisher's control.
On the other hand, as argued above, applying only traditional media exposure
models to the Web does not take into account its unique, interactive nature.
Additionally, the Internet is the first commercial medium in which it is actually
possible to measure consumer response, not just to assume it. Although the
click-through model may not represent the optimal approach to measuring the
value of interactivity, it offers a departure point from which to proceed.
Proposed Web Pricing Models
Interactivity
While payment based upon click-through guarantees that the advertiser
knows that there was visitor exposure to target communications, it does not
guarantee that the visitor liked the communication or even spent any substantial
time viewing it. It is proposed that an additional measure of the value of an
advertisement should be based upon the degree to which the visitor interacts with
the target communication. An interactivity metric might be based upon the
duration of time spent viewing the communication, the depth or number of pages
of the target communication accessed, the number of repeat visits to the target
communication, or some combination of these three elements.
Such a practice was announced for the first time in 1996 when a member
of the Internet mailing list Online Advertising Discussion List (1996) posted to
the list that Modem Media, the interactive advertising agency, had developed a
pricing model in which its clients will pay, not for exposures or clickthrough, but
12
only for activity at the client's Web site. This development raised anew the
controversy surrounding the best Web media pricing models, with Web publishers
arguing that the problem with activity-based measures, such as click-through or
interactivity, is that the Web publisher cannot be held responsible for the activity
related to an advertisement. An analogy is drawn to print, with the Web publisher
arguing that the print medium charges for advertisements, regardless of whether
they lead to sales.
Not surprisingly, advertisers and their agencies continue to argue that,
since the Web medium allows for accountability, it is possible and desirable to
develop models that measure consumer behavior. In the long run, the solution will
probably be found by accepting the reality that the medium and the advertisement
interact and that all parties share responsibility for outcomes.
Outcomes
Ultimately, marketers are interested in outcomes, and the ultimate
outcome is purchase. As Stephen Klein, former I/PRO manager, stated, "One
hundred thousand people going to a site is worth something, but a site that only
five people visit can be worth more if they are the right five people" (Murphy
1996).
The metrics discussed above relates to early stages of the purchase
process. Banner advertisements affect the consumer's awareness and interaction
with the target communication affects the consumer's comprehension and
understanding. Beyond these initial stages are the marketing objectives of attitude
change, purchase intention, and, ultimately, purchase.
13
An outcome-based approach to pricing Web advertising begins by
specifying precisely the marketer's goal for the target communication. Examples
of typical outcomes include influencing attitudes, motivating the consumer to
provide personal information, or leading the consumer to purchase. Whatever the
marketing objective, the Web provides a vehicle for integrated marketing
campaigns, which allows the marketer to track and to measure the effectiveness of
the advertisement.
A current example is per-inquiry (PI) advertisements. PI advertisements
pay a royalty only on actual product sales and require no other payment. Consider
the online affiliate programs offered by Cdnow and amazon.com, among many
other advertisers. In such programs, affiliates advertise products on their Web
sites that are sold by the advertiser and that the affiliates feel are appropriate to the
content of their Web site. If a visitor accesses the advertiser through the affiliate’s
Web site and purchases the product advertised on the affiliate's site, the affiliate
receives a referral fee or commission. Currently, referrals range from $0.50 to
$5.00 or more for each lead, while commissions range from 10 to 25 percent of
the purchase price of the product.
Although the pricing models most frequently applied to the Web are based
on traditional, mass media models, it may make more sense to incorporate the
direct response paradigm when considering outcomes.
Consider the following definition of direct marketing (Direct Marketing
Association 1996): any direct communication to a consumer or business
recipient that is intended to generate a response in the form of an order
(direct order), a request for further information (lead generation), and/or
14
a visit to a store or other place of business for purchase of a specific
product(s) or service (s) (traffic generation).1
The concepts of "direct order," "lead generation," and "traffic generation"
are immediately and obviously applicable in the many-to-many environment
underlying the Web. Outcome definitions and metric developed from considering
the Web as a unique hybrid of direct response and traditional communication
media will lead to the optimal set of models for measurement and pricing.
Integrating Exposure, Interactivity and Outcomes
Following the creation of measurement and pricing models, it will be
necessary to develop a set of integrated response measures, over time, and,
possibly, over sites that relate exposure and interactivity metrics to consumer
response. Exposure and interactivity metrics may take the form, for example, of
purchase behavior in an online storefront, attitude change, and the number of
visitors who request further information. The development of such metrics,
however requires two things: 1) identified visitors, and 2) multi-site data on every
Web site involved in the integrated marketing campaign. Until these data are
available, the measurement of outcome remains elusive.
In addition to the metrics described above, there are several other
behavioral and psychological measures that should be considered in the context of
Web advertising measurement. These additional measures include: primary
navigation patterns through the Web site; cross-site navigation patterns;
demographic, psychographic, and behavioral characteristics of visitors to a Web
15
site and to specific pages within a Web site; cognitive and attitudinal measures,
including flow; and visitor loyalty and repeat visits. It is anticipated that
innovative future pricing models will incorporate these measures in unique ways.
Policy Considerations
When developing Web measurement standards and pricing models, a
number of policy issues must be considered, including privacy and ethics. The
policy considerations are particularly important because consumer protection,
fraud, and deceptive claims problems are potential points of entry for government
regulators into the Web marketplace.
Privacy
Although a thorough analysis of privacy issues is beyond the scope of this
chapter, it is important to raise the issue in the context of Web pricing models.
Networked, distributed computing environments, such as the Internet, offer
unprecedented opportunities for the invasion of privacy. Information about
individuals is more accessible and more easily combined and integrated than in
the physical world. It is not so much that it is possible to learn things about
consumers that could not be learned before, but, rather, that gaining access to such
information might have been too expensive, too time-consuming, or too difficult
to gather previously.
In addition, it is not clear who would be able to have access to such
consumer information or what they might do with the information. In a different
context, serious privacy issues have arisen regarding patient mental health
information that has been entered into computer networks at the insistence of
16
insurance companies (Lewin 1996; Scarf 1996). This information was accessible
by a class of "health information trustees," whose inappropriate use of this
information in some cases has had serious and damaging consequences to
consumers. A parallel class of "marketing information trustees" could potentially
have access to vast databases of consumer transaction data.
In the context of Web measurement for marketing and advertising
purposes, the specific issues are the information that is gathered from consumers,
the awareness of consumers that it is being gathered, and the uses to which the
information is put. There is a tension between the marketer's need to know
information about individual consumers for the purpose of targeted marketing
efforts and the consumer's right to privacy. In our opinion, the ultimate solution to
this tension is the establishment of a full partnership with consumers, in which
they control ownership of their demographic and behavioral data and determine
the manner it will be used, if at all. This solution respects the many-to-many
model underlying the World Wide Web that permits consumers also to be
providers to the medium and allows consumers to remain active participants in the
interactive communication process.
For a demonstration of the type of information available to marketers
about visitors to Web sites, visit the Web site http://anonymizer.cs.cmu.edu:8080/
The "Anonymizer" site demonstrates the sort of information about the visitor that
is available to the Web sites that people visit. Depending upon the platform from
which the "Anonymizer" site is accessed, the information ranges from what
domain you are coming from, to what kind of computer you have, to the Web
browser you are using to your name and other personal details.
17
The approaches to privacy by advertisers that are most likely to attract the
attention of government regulators are those that ignore consumers' rights and fail
to enter into explicit agreements with consumers about their demographic and
behavioral data.
A recent study (Hoffman, Novak and Peralta 1999) found that Web users
value their privacy, particularly as expressed by visiting sites anonymously or by
adopting various aliases, depending upon the circumstances of the visit.
Furthermore, users desire "complete control" over whether a particular Web site
should receive any information about them. While users recognize that marketers
may desire demographic and behavioral data on visits for business purposes, users
do not feel that marketers have the right to sell these data to other firms. Web
users seemed willing to provide demographic information, if marketers identify
the information that is being collected and the uses that will be made of it. they
These findings suggest that privacy policies in this emerging medium
should be driven by the unique characteristics of the medium, such as,
interactivity, and the desires of its users, for control, for example, as they
experience that medium (Hoffman and Novak 1996).
Ethics
Researchers are beginning to address the question of ethical behavior in
the conduct of online research (Boehlefeld 1996; Duncan 1996; Thomas 1996). A
key result of this research, that "informed consent" is a critical component of
ethical research in many online environments, has general implications for the
way marketers may approach gathering data form Web visitors. Much more
specific consumer research is necessary, however, to determine the best ways to
18
develop and to implement such policies in commercially oriented Web
environments.
"Disguised advertisements" are another potential ethical concern. Suppose
that an advertiser-supported search agent site presented links to an advertiser's
Web site at the beginning of a list produced by a search request for a set of
keywords. In this case, while the person performing the research may believe a
link appears at the beginning of the list because it is the most relevant to the
search request, the top position of the link may be due to sponsor payments. Such
practices must be made clear to users of the search agent, as they have the
potential to deceive consumers and undermine trust in search agent sites.
Conclusion
The standardization of the Web measurement process is a necessary
precursor to the development of optimal Web advertising pricing models. Indeed,
without such efforts, the Web is not likely to achieve its full potential as a unique
and revolutionary commercial many-to-many mass medium. Managers must now
begin to address the issue of what the appropriate standards for Web advertising
pricing models should be. The distinction between passive advertisements that are
under the marketer’s control and active advertisements that are under the
consumer’s control has important implications for the measurement and pricing
process.
The best pricing models will be based on interactivity metrics. The
rationale behind this argument is that the degree to which the visitor interacts with
19
the target communication is a better measure of the value and effectiveness of an
advertisement than mere exposure.
Metrics based solely on impressions are necessary in the Web
measurement process, but must not form the basis of Web advertising pricing
efforts. Ultimately, what is required is a set of integrated response measures that
relate exposure and interactivity metrics to consumer response. These
"interactivity metrics" and "outcome metrics" must be included in any complete
program for Web measurement and advertising pricing.
Further research will be necessary to identify those metrics that are most
useful for judging the effectiveness of advertising, for determining the placement
of advertisements, and for determining optimal pricing schemes for efficient
media buys.
Our primary objective in writing this chapter is to stimulate further
research and discussion and help facilitate the process of developing optimal
pricing models for the Web. Because of the unique nature of the Web medium,
this process should proceed as a partnership among all stakeholders in the
industry, including advertisers, Web publishers, measurement, placement and
auditing agencies, and consumers (Hoffman and Novak 1997). In particular,
advertisers and commercial Web sites that sell advertising space must work
together to measure consumer outcomes in the context of direct response rather
than solely in terms of mass media exposure.
Only through cooperative effort among diverse constituents with
often-conflicting needs are we likely to make progress on the rough path toward
profitable commercial development.
20
References
Adknowledge, Inc. (1998), http://www.adknowledge.com/
Advertising Age (1996), "Fall Prime-Time Pricing Survey," Advertising Age,
September 16.
Associated Press (1996), "Procter & Gamble World Wide Web Ad Strategy
Raises Online Ire," San Francisco, Apr 28, 1996 1:17 p.m. EDT,
http://www2.nando.net/newsroom/ntn/info/042896/info5_380.htmI
Boehlefeld, Sharon Polancic (1996), "Doing the Right Thing: Ethical
Cyberspace Research,: The Information Society, 12, 141-152.
Chatterjee, Patrali (1998), "Modeling Consumer Network Navigation
in World Wide Web Sites: Implications for Advertising,"
Doctorial Dissertation, Owen Graduate School of Management,
Vanderbilt University.
ClickQuick (1999), “In Depth Reviews of Affiliate and Pay-Per-Click Programs,”
http://www.clickquick.com/
Competitive Media Reporting (1999). http://www.usadata.com/usadata/cmr/
Direct Marketing Association (1996), "DMA Report - Economic
Impact: U.S. Direct Marketing Today, http://www.the-dma.org
DoubleClick (1996).http://www.doubleclick.com
Duncan, George T. (1996), "Is My Research Ethical?"
Communications of the ACM, Special Issue on Internet in the
Home, 39 (December), 67-68.
Focalink (1996), MarketMatch Data Base.
Forrester Research (1999), http://www.forrester.com/
21
Hoffman, D. L., T. P. Novak, and Patrali Chatterjee (1995), "Commercial
Scenarios for the Web: Opportunities and Challenges," Journal of
Computer-Mediated Communications, Special Issue on Electronic
Commerce, 1 (December).
http://shum.huji.ac.il/jcmc/vol1/issue3/vol1no3.html
Hoffman, D. L. and T. P. Novak (1996), "Marketing in Hypermedia
Computer-Mediated Environments: Conceptual Foundations"
Journal of Marketing, 60( July), 50-68.
Hoffman, Donna and Thomas P. Novak (1997), "A New Marketing
Paradigm for Electronic Commerce," The Information Society, 13 (Jan-
Mar.), 43-54.
Hoffman, Donna L., Thomas P. Novak and Marcos Peralta (1999), “Information
Privacy in the Marketspace: Implications for the Commercial Uses of
Anonymity on the Web,” The Information Society, 15 (May).
Internet Advertising Bureau (1999) http://www.iab.net/
I/PRO (1996), "The Web in Perspective: A Comprehensive Analysis
of Ad-Supported Web Sites,: I/PRO Research.
Lewin, Tamar (1996), "A Loss of Confidence: A Special report.
Questions of privacy roil arena of psychotherapy," New York Times, May
22, Section A, p 1, col 1.
Murphy, Ian P. (1996), "On-line ads effective? Who knows for sure?"
Marketing News, 30 (20), Sept 23, pl, 38.
Online Advertising Discussion List (1996), Volume 1, Number 91, September 24
www.tenagra.com/online-ads/.
22
Rebello, Kathy (1996), "Special Report: Making Money on the Net,"
Business Week, September 22, pp 104-18.
Scarf, Maggie (1996), "Keeping Secrets," The New York Times, June 16, Sect. 6,
p 38, col 1.
Thomas, Jim (1996), "Introduction: A debate about the Ethics of Fair Practices for
Collecting Social Science Data in Cyberspace," The Information Society,
12, 107-117.
Wilkie, William L. (1990), Consumer Behavior, 2 nd Edition, New York: John
Wiley & Sons.
1 Direct Order includes all direct response advertising communications – through any medium –
that are specifically designed to solicit and close a sale. All of the information necessary for the
prospective buyer to make a decision to purchase and complete the transaction is conveniently
provided in the advertisement. Lead Generation includes all direct response advertising
communications – through any medium – that are designed to generate interest in a product or
service, and provide the prospective buyer with a means to request and receive additional
information about the product or service. Traffic Generation includes all direct response
advertising communication conducted – through any medium – that are designed to motivate the
prospective buyer to visit a store, restaurant, or other business establishment to buy an advertised
product or service.