Teaching the Internet to Value Our Time

Peter Coffee: Personalization and sophisticated algorithms can help Web developers meet individual needs.

"Information consumes attention," as Nobel economist Herbert Simon famously observed. "A wealth of information creates a poverty of attention." The Web, by this measure, makes all of us paupers: It offers so much data that we all wind up working harder to accomplish less, if our measure of achievement is whether we feel that were making well-informed decisions.

And we dont like feeling poor.

If youre looking for a competitive edge, nothing has more potential than increasing your customers return on invested attention—making them feel rich, in ways that mere money cannot. The question is whether the exponential improvement of data delivery technology can be tied, with equal success, to the task of showing each user the subsets of personal interest.

Can we make the Net pay attention to us and tailor what it offers based on what we want to know?

The Net is trying to pay attention, as measured by the explosion of effort devoted to capturing clickstreams. "By the middle of the next decade," projects data warehouse consultant Ralph Kimball, "clickstream databases will dwarf all the other conventional data warehouse examples. And the worst part is that the clickstream is really interesting."

What sharpens the edge of the problem is that the clickstream, unlike other bandwidth-busting traffic, resists compression. We can learn what people notice, or dont notice, in video scenes and the like, and we can apply ever-increasing processor power and ever more sophisticated fractal and wavelet algorithms to reduce the number of bits that we actually need to send to convey the essence of an experience. Personalization, by contrast, has its greatest promise in the highest fractal dimensions, so to speak: in the attributes that distinguish one "jazz record buyer" from another, for example, so that the same core best-selling products dont get recommended over and over again in a self-perpetuating process.

Nothing consumes more of our attention than streaming media, which we cant readily summarize or skim. At the same time, we value its high information content. IBM researchers are trying to resolve the resulting paradox with their VideoCharger technology: This monitors user interactions with video to develop a model that predicts what a user would find most interesting and broadens user options for direction and speed of viewing.

Some creative artists will resist such technologies, saying that they want their works viewed as they were conceived—or not at all. Piffle. The function of art, like the function of any other work of the mind, is to stimulate other peoples thinking—whether the end result is a decision or merely (if thats the right word) a more enlightened state of being.

Like books that let us reread, or skim, or return to a favorite passage at will, other media technologies need to become more responsive to our interests if we are to give them the favor of our attention.

And content on the Net, even if it does stay "free" in monetary terms, needs to work harder to be less expensive in terms of what it gives us back for our time.

E-mail eWEEK Technology Editor Peter Coffee