Organization site - Summer 2001
Banner Advertising ... Pt 2
Banners on the home page of an online newspaper are not targeted well to the content of the "front page." At the time of this study, the authors found no evidence of effective advertising procedures in place for responding to the daily news routine. In other words, advertising departments at newspapers may not be prepared to keep up with content that updates rather frequently.
Targeted banner ads found at the first, second and third clicks show a relatively flat line. According to these findings, classified advertising sites of online newspapers (which may be operated in-house or by third-party vendors such as AdQuest and AdOne) do not even cross the 60 percent line for targeted to content.
Some of these targeted banners were house ads, promotions for the newspaper. House ads were coded as banners. Teasers to "inside" content sometimes looked like banner ads. They were not coded.
By the way, of the 75 sites coded, only one BMW banner ad appeared on a page that displayed BMW classified ads.
Wassmuth and Thompson (1999) have reported a trend that indicates the percentage of banner ads that contain specific content increases at deeper levels of the information hierarchy. By linking the findings of their study of banner ad content to this pilot study of targeting banners to page content, the researchers suggest the following: As banner messages become more specific after two or three clicks, the percentage of banner ads targeted to the content of the page on which they appear increases.
In the researchers' opinion, this is an important finding that indicates room for improvement in targeting practices of online advertising by newspapers and, in some cases, their third-party classified advertising services.
One possible reason for this lack of targeting performance may be related to the fact that few online newspapers are collecting information about their readers. Only one site required registration by users before entering the site: The New York Times.
And no ad viewed during the coding session for the New York Times was targeted to the content of the page. Three banners were seen on the home page, none after the first click, one after a second click, one after a third click and one after a fourth click. Not one of these banner ads was targeted to the content of the page.
This study did not evaluate the performance of "cookies." If the coder had returned to the New York Times site for a second visit, the cookie should have assisted the host (the newspaper) in targeting ads to the client (the coder's computer).
But the success of a cookie depends on the identity of the user. If one person registers but another person later accesses that site from the same computer (assuming login with the same profile), then the second person would see ads targeted to the "wrong" person.
The same problem would occur when personalized news is read by a different online reader.
Site sponsorship was found on only two of the 75 newspaper sites examined. Perhaps newspaper publishers resist the concept of site sponsorship.
Perhaps resistance from the newsroom is part of the answer. The authors suspect editors would be reluctant to invite site sponsorship because of the possibility that the sponsor may try to exert pressure to control content to some degree. This calls up major questions of ethics, credibility and accountability.
More likely, the researchers suspect, newspaper advertising departments have not developed programs for promoting and implementing sponsorship. In general, the current practice by online newspapers of targeting banner ads appears to be relatively weak. Techniques used for targeting ads to an online audience - registration, site sponsorship, and targeting banners to surrounding content - are underutilized, according to these findings.
An information location task was used in this pilot study as a way to examine online newspaper advertising beyond the home page. The researchers believe this is a strength of this research and a meaningful contribution to the literature.
However, this approach did not allow coding of "section" content, such as news, sports, and entertainment. Future research will examine those content areas.
Determining the effectiveness of sites that use a registration procedure to deliver ads targeted to users on subsequent visits is beyond the scope of this study. This pilot study coded "first visit" only. Research designs that inc]ude repeated visits to a site may be used in future studies.
The low level of site sponsorship may be related to the methodology used for this study, current practices by newspapers and their advertisers, or both. This study was limited to home page and classifieds. Future studies that access editorial sections of online newspapers may reveal a greater degree of sponsorship.
The authors believe qualitative research is needed that examines newspaper advertising department practices, procedures and policies for online ads. Of particular interest to the authors would be examination of the interactions between the advertising and editorial departments.
The literature review for this study supported the concept of targeting banners to the content of the page on which they appear.
However, evidence revealed by this pilot study shows weak support of this concept.
The researchers do not mean to suggest that newspaper advertising professionals are doing a poor job. Probably, there are no systems in place that even attempt to collate banner ads with page content.
With this study, the researchers do hope to raise awareness that a deliberate process of matching ad content with page content may improve the effectiveness of online banner ads.
For home pages of online newspapers, banner ads were not targeted well to page content. The percentage of targeted ads increased after the first, second and third clicks. But, even then, only about half (50 percent after the first click; 49 percent after the second click, and 56 percent after the third click) of the banner ads were targeted to page content.
Extending the findings of their study of banner ad content to this pilot study of targeting banners to page content, the researchers suggest: As banner mes sages become more specific after two or three clicks, the percentage of banner ads targeted to the content of the page on which they appear increases.
For example, a newspaper's home page may have a sampling of the day's content and no banner ad that matches any of that content. A click on the
International section, however, might lead to a page on which a banner ad for Delta Airlines international service appears. lf a story about France is accessed next, the reader might expect to see a more specific ad for Delta's schedule and ticketing for flights to Paris.
Future research will test this relationship between specificity of banner ad content and targeting banners to content of the pages on which they appear.
Of course, the question must be asked: Is there a need to develop systems for matching content of page to content of banner ad?
Why not simply target the ad to the user? For this study, "personalization" of news was not a factor. This pilot study did not evaluate the relationship between the newspaper site and individual users.
The researchers discovered some teasers to "inside" content that looked ] ike banner ads. Teasers were not coded. Sometimes banner ads were interlaced with teasers. The practices of creating teasers that look like banners and placing teasers among advertising seems to blur the distinction between editorial and advertising.
Is this good for click-through? Or does this train users in ad avoidance and, possibly, editorial avoidance? More research is needed on the identifiable or characteristic features of banner ads, particularly as related to the formal features of editorial content.
As with many pilot studies, sample size used was relatively low. Eighty-one of the almost 800 daily, general interest newspapers found on the Editor & Publisher Web site were selected. This seemingly low sample size, combined with the fact that the number of ads diminished after the third click, affected the statistics that could be run. Researchers should consider larger sample sizes for future studies of this kind in order to obtain higher numbers per cell (for a Chi Square analysis, for example).
The researchers recommend involving more than two coders for studies of this kind. Only two coders participated in this content analysis.
The researchers chose not to select a stratified sample of newspapers by circulation, an indication of the relative size of the organization. In their opinion, the Internet puts all content providers on "equal footing," at least in terms of the potential to be perceived as top-notch organizations. However, a stratified sample may be considered for future research.
Because the Internet makes all electronic publications "global," geographic representation was considered but rejected by the researchers for this study.
As is the nature of a pilot study, the researchers worked with a new coding instrument. The researchers are refining their measurement tool for content analysis of online newspaper content.
Unlike studies of home pages, which tend to examine a higher number of ads, this pilot study looked at ads on every level of the information hierarchy. Successive levels were examined as the coder completed a specific search task.
When the statistics printed, the researchers were thrilled to see a relatively large number of "p<.001" levels reported. However, scrutiny of the cell sizes used to determine those results prevented the researchers from reporting such robust findings.
The low number of ads found at some of the deeper levels of the site caused the problem. Some cells of the analysis were too small (N<5) to run inferential tests properly. This rendered Chi Squares and correlations powerless.
Therefore, only descriptive statistics have been reported.
This article reports a content analysis of current online newspaper practice regarding targeting banner ads to an audience through the use of site registration, cookies (not a factor is this study), sponsorship, and matching, or collating, banner ads to page content.
Unlike much of the previous research of online advertising, this pilot study did not sample "Top 50" sites. Usually, Top 50 lists are made up of a hodgepodge of sites that represent a variety of content domains. This is one of the first studies of online newspaper advertising based on a systematic sample.
Findings of this pilot study indicate only two instances of site sponsorship and one example of registration to enter the site.
Findings show an extremely low level of targeted ads on home pages of online newspapers. An increase in the percentage of banner ads targeted to page content occurs after the first, second and third clicks. But even then, only about half the banners are targeted to page content.
Coders were assigned an information location task (find a classified ad for a BMW) and instructed to code each banner ad encountered during the search. The increase in percentage of banners targeted to page content peaks at the 3-click level, where many of the searches were completed.
Sites that violated the "3-click rule" were not effective at targeting banner ads to page content at the fourth click and beyond.
The majority (87 percent) of newspaper sites coded did use banner ads. The majority (59 percent) of "pages" viewed included at least one banner ad. Overal1, only 27 percent of banner ads were targeted to content. A high number of untargeted banners were found on home pages.
Many online newspapers still are struggling to become selt:sustaining profit centers. The authors believe the study reported here provides insight about one type of symbiotic relationship between editorial content and advertising content that may improve newspapers' online products.
Andrews, Whit (1996). "Need to boost click through rates on your ads? Advice from some experts" [online]. Internet World, reprinted from Web Week, October 21, 1996, vol. 2, issue 16. [Available: www.internetworld.com/print/ 1996/10/ 2 l / undercon/clickthrough.html]
Angwin, Julia (1996). "It's tough to get noticed on the Web: Advertisers find the novelty has worn off" [online]. SF Gate (San Francisco
Chronicle), Tuesday, October 1,1996. [Available: http://www/sfgate.com/cgi-bin/article.cgi?file=/ chronicle/ archive/1996/l 0/0l/BU25590/DTL]
Banner Design (1998). "Effective Banners: "Definitions" [online]. [Available: http:// www.whitepalm.com/fourcorners/definitions.html]
Briggs, Rex, and Hollis, Nigel ( l 997). "Advertising on the Web: Is there response before click
through?" Journal of Advertising Research, April 1997,37(2):33-45
Bucy, Erik P., Lang, Annie, Potter, Robert F., and Grabe, Maria Elizabeth (1998). "Structural features of cyberspace: A content analysis of the World Wide Web," Paper presented at the AEJMC Annual Convention, Theory & Methodology Division, August 5-8,1998, Baltimore, Maryland.
Cho, Chang-Hoan (1998). "Applying the Elaboration Likelihood Model to audience processing of Web advertising." In Darrel D. Muehling (Ed.), Proceedings of the 1998 Conference of the American Academv of Advertising (pp.303).
Guthrie, J.T. (1988). "Locating information in documents: Examination of a cognitive model, Reading Research Quarterly, Spring, 23(2): 178199.
Guthrie, J.T., Britten, T., and Barker, K.G. ( l 991). ,,Roles of document structure, cognitive strategy and awareness in searching for information, Reading Research Quarterly, 26(3):300-324.
Guthrie, J.T., and Kirsch, I.S. (1987). Distinctions between reading comprehension and locating information in text, Journal of Educational Psychology,79(3):220-227.
Harvey, Bill (1997). ,,The expanded ARF Model: Bridge to the accountable advertising future," Journal of Advertising Research, March/April 1997,37(2):11-20.
Internet Advertising Bureau (1998). [Available: http://www.iab.com]
KRT (1998). Excerpt from an advertisement. Editor & Publisher, mediainfo.com: Journal of the online news industry, May 1998, p. 21.
Lainson, Suzanne (1998). ,,Re: Newspapers, classifieds and theWeb" [posted to online discussiongroup,Friday,March27,1998,13:51:54]. Avai lable: Archive for the Online-News discussion group: http://www.planetarynews.com/on-archive
McDonald, Scott C. (1997). ,,The once and future Web: Scenarios for advertisers," Journal of Advertising Research, April 1997,37(2):21-28.
"Media buying tips" (1998) [online]. Four Corners: Effective banners. [Available: http:/www.whitepalm.com/fourcorners/adbuyingtips.shtml]
Robinson, Kari Sargeant, and Kaye, Barbara K. (1997). ,,Determining the strength of new technology: The World Wide Web's effect on the advertising industry." In M. Carole Mackline (Ed.), Proceedings of the 1997 Conference of the American Academy of Advertising (pp.217-223).
Scott, W, (1955). "Reliability of content analysis: The case of nominal scale coding," Public Opinion Quarterly, 1955, vol. 61, pp. 483-92.
Thompson, David R. (1993). Mass communication and the newspaper of the future: Some effects of modality, story type, and search experience on reading time, memory, and information location (Doctoral dissertation, University of Texas, 1993).
Wassmuth, Birgit, and Thompson, David R. (1999). "Banner ads in online newspapers: Are they specific?" In Marilyn S. Roberts (Ed.), Proceedings of the 1999 Conference of the American Academy of Advertising (pp.263-271).
- - - - - - - - - -
David R. Thompson, Ph.D., is a faculty member in the Department of Communication Arts at Loras College (Dubuque, lowa USA). Thompson is an awarding winning teacher and researcher His specialties include computer-mediated communication, Web-based media, and other emerging forms of digital communication. Thompson has served as the director of an online newspaper. And he also has taught at Simpson College, the University of South Carolina, and Southwest Texas State University.
Birgit Wassmuth, Ph.D., is a Professor of Advertising at Drake University (Des Moines, lowa USA). Wassmuth is known internationally as an expert in graphic design, visual communication, advertising (print and interactive), desktop publishing, information graphics, integrated approaches to communication strategies, and Web site design. Wassmuth also has heldfaculty positions at the University of Missouri, University of Florida (Gainesville), and Temple University.