“I know I’m wasting half my advertising dollars — I just don’t know which half.”

Most credit the advertising industry’s greatest paradox to John Wannamaker, king of American retail in the early 1900s. It was a simpler age of course, yet that statement summarizes today’s dilemma of advertising in a landscape of clutter. How do you reach a potential customer, and how do you know he’ll be listening to you when you do?

The problems facing marketers in the age of mass media are many. Just when it seemed there were enough TVs to broadcast a single message to the world, the television universe fragmented into a sea of cable channels and Web sites. Even when viewers could be identified, marketers had no way of knowing if viewers were zapping away from commercials with the remote, or skipping over them entirely with digital video recorders like TiVo or ReplayTV.

Techniques ranging from branded entertainment to guerrilla marketing have emerged to reach an increasingly well-protected audience. But the one that makes marketers feel most secure — the one that seems to be based in the pure cold reason of numbers — is market segmentation. Segmentation was first developed by the direct-mail industry as a way of saving postage and increasing response rates.

The first of these research firms amounted to little more than giant library card catalogs, with little files on each household in their region. The object of the game at that time was simply to minimize the costs of wasted postage and printing by sending mailings only to the households that stood a good chance of responding. At the time, there wasn’t enough direct mail advertising for it even to get the name “junk mail” yet. With the advent of computers, the industry took off, and direct mailings grew exponentially. Now, the object of the game was to hit households with messages that could penetrate the clutter and avoid the dustbin.

Companies like Acxiom and Claritas competed to collect consumer, census and any other data on every household in America, and sometimes beyond. By striking deals with some of the merchants they were serving, these two firms grew to behemoth proportions, taking in as much data from their clients as they were providing. In fact, all but two of the 9/11 hijackers were listed in Acxiom’s tremendous databases. Understandably, marketers in this confusing age — one made even more confusing by the highly mobile habits of mobile phone users — have gravitated to the certainty of number crunching and consumer modeling. Acxiom, for one, has broken down its millions of files into about 70 specific “types” of consumers.

These profiles include information ranging from what TV shows the target consumers watch or which books they read to how many doors are on their cars and whether they like dogs or cats. The research firms then use statistical modeling to infer other sets of facts from the ones they have. In the 2004 U.S. presidential campaign, for example, it was discovered that undecided voters in Ohio were much more likely to have call-waiting on their phones, and cats instead of dogs. Why? No one really knows. All that matters is that this statistical truth gave campaign managers the ability to target certain households, and not waste their time, energy and money on others.

For a wireless industry confronting unique competitive challenges and with a consumer base as fickle and disloyal as any, it’s only natural for marketing departments to exploit any tool at their disposal to reach potential consumers. And while direct mail advertising may be a good way to reach those few elderly or impoverished consumers who still don’t have a phone, I fear that the application of market segmentation techniques in the wireless universe could have a detrimental effect.

While segmentation might be effective in predicting a great deal about consumers, it may ultimately contradict the very ethos of wireless culture, for the wireless industry is different from any other. Market segmentation may be great for companies trying to figure out which customers should receive free samples of their soap in the mail; but using consumer data to target, say, SMS messages at likely consumers could be a disastrous move for the entire industry.

It goes beyond turning the mobile phone into an advertising platform, a strategy that most of us already realize will have to be handled carefully, if employed at all, since mobile users — unlike Internet surfers — are less likely to accept commercial interruptions when they’re paying by the minute.

But market segmentation, in particular, conveys an understanding of human behavior and predictability that challenges the underlying premise of the wireless revolution. Because of the spontaneity it allows, consumers have embraced wirelessness. A mobile phone permits people not simply to be late for an appointment, but to change arrangements, make sudden plans and encourage impromptu events. Some bit of the spirit of a flash mob trickles into every wireless interaction.

Market segmentation, on the other hand, depends on the predictability of consumer behavior. A targeted message also carries with it a subtler message that the recipient can be counted on to act in certain ways. Segmented marketing is intended, after all, to exploit what a marketer knows about a consumer in order to make her more like that. A person who has clicked on a few banner ads about dieting, or purchased a book on carbohydrates, is an ideal target for an onslaught of diet product ads. The consumer experiences segmentation and targeted messaging as having been “found out” and is then pushed further towards a static identity. Worse, segmentation is potentially divisive and alienating, since it defines users by what they want, rather than what they can do or contribute. Under the market researcher’s scalpel, society is sliced into evermore-precise consumer tribes. Mac users vs. Microsoft users. Ford drivers vs. Chevy drivers.

This works for traditional products, like sneakers or cars, because they serve as real, physical badges. Objects like these can serve as signifiers for a fixed or tribal identity. Wireless devices, however, are portals. Instead of promoting fixed identities, they are systems through which people can develop and manifest ever-changing selves. Wirelessness is about the expansion and plasticity of self-definition, and participating in culture on new levels. To turn mobile users into clearly delineated sets of target markets defeats the potential of this interactive, collective technology.

The information provided by market segmentation specialists can still be useful, but for the opposite reasons: instead of separating them further, these are groups we can connect. This is the connections business after all. Each segment is just a starting place, not an end in itself, as it is in so many other industries.

One of the only places where segmentation might apply in the wireless industry would be to the devices themselves. Certain handhelds can be targeted to certain people. But even then, the beauty of wireless devices is that they can be customized and changed based on the changing allegiances of the wireless user. A Hello Kitty faceplate one day, an American Flag the next. People who select and change their own ringtones every week shouldn’t be segmented into a fixed sense of self. (They’ll buy less ringtones!) It’s this very fluidity of self-image they’re paying you to support. A wireless device is not just another product, it is a symbol of infinite possibility.

Another area that seems, on the surface, particularly well-suited to the target marketer would be rate plans and service packages. And while, at first glance, offering a games package to a kid and a business package to a wealthy adult seems like a good idea, a company communicating selectively like this does so at a price. The consumer quickly realizes that she has been identified in terms of her usage patterns, age, wealth and whatever else the provider might know about her.

As a result, each keystroke, each game download, each refresh of a WAP site is, from then on, weighted with the additional consideration that these actions are being monitored, and that future offers will be based on them. Even if this effect is so subtle as to be subconscious, the categorization of consumers into evermore rigid personal profiles works against the underlying sensibility that wireless users have the freedom to define and redefine themselves with every new moment of interaction.

Even the early Internet benefited tremendously from users’ ability to recast themselves into new characters. Chat rooms were filled with individuals trying on new ages, sexes and temperaments. Where the Internet did this on the level of fantasy or avatar, wireless devices allow this to happen in a way that is much more connected to one’s real life experience. The North American mobile community UPOC, for example, allows young people to belong to any number of different or overlapping affinity groups. Hopping onto the Buffy the Vampire Slayer list shouldn’t necessarily peg an individual as a Buffy fan, as targeted ads based on this information might suggest.

The more users sense that their actions in the wireless space are tracked and then acted upon, the less they will look to it as a way of trying on provisional identities, engaging in spontaneous experiences and testing new activities. It’s as if everything will appear on their permanent record.

Indeed, basing a 21st-century marketing strategy on a technology originally developed to save on postage stamps just isn’t the most thoughtful or productive way of engaging with this new marketplace. It could reduce wireless participation from a mode of self-expression to a reflection of some pre-fab, computer-generated consumer profile and, in the process, change the mobile phone from a portal into a mirror.