Much of the Internet's glitz is underpinned by an invention so unimaginative that I hesitate to call it an invention at all: the banner ad, cluttering up screens and slowing down surfing. As the Internet changes from an elite, publicly funded research network to a socially more diverse commercial environment, the economics of content production are changing with it. The noncommercial (academic and arts) model of content production--funding up front and free dissemination later on--is being superseded by a commercial model in which the distribution of content provides the major source of revenue. The Internet's transformation into a commercial environment will not come easily, however, particularly not for innovative e-Commerce. Technical and cultural legacies prevent it from being transformed overnight. First, there is no widely available technology allowing transactions to be made easily on a micro-scale (which might support a new, truly decentralized economic model). Second, Internet culture, with its slogan of information that wants to be free, is anything but enthusiastic about paying for content. Both factors are changing, though slowly; once embedded in machines and the minds of most users, routines acquire considerable inertia.
It would be shortsighted, therefore, to hail this "free information" as a sign of the emergence of a new economy and tout the old rules as obsolete. On the contrary. The old corporate models are simply being used in a new environment, though they've been dressed up a bit. It's the clash between old models and new technology that sparked the banner ad.
The content of a new medium is often an old medium, as Marshall McLuhan observed a long time ago. (1) And in the new medium's early days, the old medium often is simply transferred into the new environment, regardless of its fit. It took radio broadcasters a few years, for example, to understand that hearing music on the radio is not the same as hearing music in a concert hall. The crucial difference is that radio listeners can't tell the difference between a live performance and a live recording. What seems obvious to us today had to be learned first. Banner ads, in a sense, have the same cultural significance as the bygone radio orchestra. They illustrate at once how embryonic the (commercial) understanding of the new environment is and how powerful the old concepts still are. The radio orchestra helped to make radio popular and fill the void until enough experience was gained with the new medium to find better ways to attract audiences. Through banner ads, the old models underwrite the new environment. Commercial content is significantly financed by banner ads because such ads provide the bulk of online advertising revenue (about 70%, with the rest coming from sponsors). Total banner ad revenue for the Web's commercial content providers in 1996 was about $300 million. It rose to about $800 million in 1997, to about $2 billion in 1998, and is expected reach $4 billion by 2000. (2)
Banner ads represent an old economy model and, based on the numbers, they are one of the few areas that are prospering economically on the Net; and not just in the financial markets. Advertising requires a mass audience. After all, it was invented--and here the term "invention" can be used with a bit less hesitation--for mass media: newspapers, radio, and TV. These still are the major reference points. Is the Internet like print (text-based), like TV (screen-based), or is it a hybrid, a bit of both? Fashioning the new in the model of the old is a powerful strategy because even the most unimaginative business people can understand it. If grappling with new technology already is straining, the last thing a good solid enterprise needs is a new, untested idea. "Let's do what we do best!" This is one reason that network communication today is so different from the communitarian, many-to-many vision of the early Internet enthusiasts.
So, dull mass markets are being forced into existence. Portals are trying to bring order into what is perceived as a chaos and attracting scores of visitors to be sold to advertisers. Just like TV, or newspapers.
However, beneath this happy-go-lucky reproduction of the old commercial model, the "difference engine" of the Internet is working slowly but steadily. Tension is rising.
- Even in the ad industry, there is a consensus that banner ads do not work. To measure success, the industry has devised a still somewhat shaky indicator of an ad's effectiveness: the click-through ratio (CTR) (3). While banner ads never have been popular, recent studies show that CTRs are steadily declining. They are assumed to be as low as 0.1% (4). The fact that a lot of money is still being poured into these ads has more to do with conservative industry culture than with high-tech productivity.
- To fight this trend and capture the user's attention, banner ads have become more animated, propped up with pop-up windows and pull-down menus. In short, they have become even more distracting, which has created a cottage industry selling software to filter out the ads. This had more to do with a niche economy than with resistance to the commercialization of the Net. To the best of my knowledge, all ad filters are commercial, proprietary software. And they sell well. AtGuard, for example, was in March 1999 the second most popular Internet software sold through Buydirect.com (5). WebWasher, another ad-blocking software, is being used by roughly a million people since it was released in December 1998, according to Siemens, the German electro giant that developed the program (6).
- Search engines use keywords from queries to group users into marketing categories to which supposedly tailored advertisement is being shown. This often creates humorous interferences, revealing the primitiveness of the correlation algorithms. Occasionally, this results in lawsuits. In June 1999, a federal judge in California dismissed a request by Playboy Enterprises that the search engine Excite stop displaying ads for porn site when people keyed in trademark terms such as "playboy" or "playmate." The company tried unsuccessfully to dissociate itself from the porn advertisement environment it generates on a search engine. Of course, Playboy will appeal this ruling.
The above examples illustrate the tension between the new environment--the Internet--and the old model--banner ads. What is new is the Net's potential for instant connections between any of its parts, its instant feedback and self-referentiality. On the Net, you can know everything, provided it's online. However, this is only a potential. Users must realize the potential by adding their meaning to it. Traditional advertising is designed to be overtly meaningless and subconsciously influential. Banner ads try to do the same. But, have you ever found yourself wanting to know more about the Marlboro man? What works on a billboard does not work on-screen. Particularly, since it's the user who creates the meaning of the potential connection through a conscious decision, activating a link by clicking on it, or, mostly, by not clicking on it. In the extreme, the user can choose to subvert the potential altogether by filtering out connections before she or he even has to ignore their existence. However, really problematic for advertisers is that users can ignore banner ads much more effectively than billboards, because a banner ads requires some action from the users (click-through), while a billboard is designed to be ignored and nevertheless reach its audience.
But even the conservative corner of Internet advertising is slowly being transformed by the changing relational patterns of the Net. When everything can be connected with everything, on the fly, it becomes difficult to establish clear-cut boundaries. The Playboy lawsuit drags this practice to court, but for companies such as Doubleclick.com, this is business as usual. Doubleclick.com is an intermediary between buyers and sellers of advertising space. Its servers, which put banner ads onto the Web pages of registered companies, dished out 29.6 billion .gif and other files in the second quarter of 1999 (7). Neither the seller nor the buyer knows exactly which ad shows up where. While this does not change the banner ads themselves, it changes the mechanics of allocating ad space. This new infomediary brings together parties that have no prior knowledge of one another. And not knowing one another makes it more difficult for anyone to gain complete control.
By instantly creating connections online, the distinctions between previously separated categories and genres are being eroded. In a sense, all of the links that come up as a result of a search engine query are the same kind of information/advertisement. From a commercial point of view, it's a gradual difference between a link with a picture (an ad) and a text link (the search result), particularly considering what appears to be a common practice of search engines: selling the top positions on their hit lists.
But this is just the beginning of the dissolution of genres through advertising. Amazon.com has pioneered a hybrid of service, distributed economy, and advertising: the "associates program." The idea is easy enough. You can create a mechanism on your personal website through which books and other merchandise can be ordered at Amazon.com which then handles the shipping and payment and gives you a 5% -15% "finder's fee." Bruce Sterling is one of the many "associates." Having received his first $50 paycheck, he promptly broadcasted this event on the Net. For Amazon.com, this is great advertising. For Sterling, this is a "viridian" economy providing him with money--"about enough to order a couple of jumbo pizzas and a bottle of decent Chianti." (8) And for the people who bought Sterling's books it means, well, buying a book at Amazon.com rather than somewhere else. Such erosion of genres and convergence of languages will be, perhaps, one of the longest-lasting cultural effects of the Internet, driving and being driven by Net commerce. A precursor of this erosion and convergence can be seen on late-night cable TV: the infomercial. Is the infomercial the future language of the Net?
The Web-based art project Airworld by Jennifer and Kevin McCoy seems to indicate that it is so (9). Airworld investigates the slippery world of advertising and e-commerce, old concepts and new technologies, and the erosion of genres. The project uses the Doubleclick.com network to distribute 1 million banner ads over one month (starting mid-August 1999). Each is good-sounding but slightly strange business-speak slogan such as "option: business as usual" or "welcome, we are air." Doubleclick.com, which sponsors the project does not inform the sites on which the ads will be displayed that they will suddenly be playing host to a conceptual art piece. But more happens to these sites. In a separate process, their textual content is lifted and inserted into a database, a process similar to the one crawlers use for search engines. For the Airworld database, the text is broken into sentences and stripped of any reference to its source. It is from these free-floating linguistic pieces that the text of the Airworld site is automatically generated, dynamically for each visitor. A similar procedure generates videos using the altavista search engine for image mining. The sites on which the Airworld ads appear serve both as promotional vehicles and as generators of raw data. This self-referentiality seems to be typical for Net art projects that often loop the Internet through itself. Airworld creates a kind of meta-corporate website. It follows the standard information architecture of a corporate site with sections for solutions, technology, alliances, shopping, etc. For each of those sections, text is dynamically generated from the database. And because they are generated through a standardized routine by a machine, these texts are meaningless. Random chunks of business lingo are simply put in a sequence. An automated voice, a text-to-speech conversion program, stoically reads the text to the visitor. Jump cuts and nonsense are mechanically erased.
This introduces the second aspect of the project and it is here that its real discovery lies. In a world of intense commercial pressure and instant connectivity, language transforms. When everything is publicly accessible, everything becomes a kind of advertisement. An infomercial, more precisely (10). A doublespeak surreally devoid of meaning emerges and automatic text creation works. How about this: "Any reader should be advised that factors such as development of new capital funding, competition, lack of expert personnel, and other factors may materially affect in an adverse manner actual results of operations." Has this been written by a good corporate lawyer or a bad text-generating program? What the project brings to the fore is the emergence of a commercial culture so homogenous, so interchangeable that we no longer know what is human and what is machine generated. It's the old dream of artificial intelligence, but upside down. In the emerging global corporate culture it is not that the machines are becoming more like humans, but humans are forced to becoming like machines. Or, as one of Airworld banner ad reads: "Are you lost? Reorganize!" What the banner, like good advertising, doesn't say is that the reorganization is all easy now; it's made by a machine.
1. McLuhan, Marshall. Understanding Media: The Extensions of Man. New York: McGraw-Hill, 1964
3. Almost all commercial sites download tracking software--so-called cookies--on client computers to follow users' movements on a site or between sites in order to establish CTRs.
4. www.a-clue.com (May 17, 1999), www.bannertips.com (June 12, 1999), Internet Week (June 28, 1999).
5. Weaner, Jane: "Are the Banner Days Over?" ZDNN, March 10, 1999
6. Flynn, Laurie J.: Software Ad Blockers Challenge Web Industry, New York Times Online June 7, 1999. For individual users, WebWasher is provided free of charge.
7. www.doubleclick.com (earnings report)
8. Sterling, Bruce. "Viridian Note 00064: Viridian Finances" Nettime. (May 3, 1999).
10. I wonder if Bill Gates, after the experience of the Microsoft trial in which his personal email suddenly turned up evidence against him, has lawyers screening his outgoing messages, just to avoid future embarassment from Freudian slips.