...or why Facebook and user-managed-metadata are more interesting than Microsoft, Yahoo! or (current) Google combined.
While the spotlight in the online world focuses on Google, Yahoo! and Microsoft, the long-term news is happening closer to home - where users interact among themselves through the Web and through online social networks. Companies with agendas mutter about user "privacy" as they jockey to promote or prevent industry consolidation - but the privacy discussion will change from "protection by the authorities" to self-determination as users learn to take control over their own data online. As they spread their "presence" online, they are looking not for privacy, but for recognition as individuals, by friends and by vendors.
Long-run, the whole world of advertising will change: The prizes being fought over right now may not be worth as much as the contestants hope. Google already has a firm hold on the best part of the market: where people are explicitly searching for something and declaring what it is. Beyond that, it's mostly inference: Where (and why) did they abandon a shopping cart? Were they serious when they clicked on that hotel in Las Vegas? (Statistics indicate that they were not, whereas people who click on Pittsburgh hotels usually have serious intentions of going there.)
But even as the giants fight over the non-search ad marketplace, new players are emerging to take behavioral targeting one step further.. Companies such as NebuAd, Phorm, Project Rialto, Front Porch and Adzilla are working with Internet service providers to track users directly (rather than via cookies) and show them relevant ads. This behavioral targeting approach undercuts traditional online publishers, who use content to lure users and to sell adjacent ads. ISPs and other behavioral targeters can sell advertisers direct access to the same users, bidding up the price of ad space but at the same time increasing its fungibility - and diminishing the importance of publishers' content.
Take user number 12345, who was searching for cars yesterday, and show him a Porsche ad. It doesn't matter where he is now - on Yahoo! or MySpace - he's the same number he was yesterday. As an advertiser, would you prefer to buy "someone reading a car review featured on Yahoo!" or "someone who visited two car-dealer sites for 40 minutes yesterday"? Of course, publishers - who control the ad space - can refuse to play, but it is tough to turn down the offer: "Look, you can sell access to your subscriber for 50 cents, but we'll bid $1 for him because we know where he was yesterday. Or we'll go find that same user on MySpace."
The market will get more competitive, and users will continued to be barraged by ads and will increasingly tune them out, even though the ads will be more "relevant. " Call that the public space, a world of billboards and cacophony. It fits the traditional definition of "privacy": The ISP and behavioral targeting networks don't know 12345's name and don't care. They just know what buckets he fits into. (Yes, if the FBI shows up, they can figure out who he is quickly enough, and of course companies make mistakes all the time and release data they intended to keep confidential, but for now consider the system as it is supposed to work.)
User-managed metadata
Now consider the other end of the spectrum: the new world of social networks, user-generated content and metadata. These aren't the "walled gardens" of yore, controlled spaces offered by outfits such as AOL. These are walled gardens custom-made by users for themselves.
Facebook, unwittingly or on purpose, has been teaching people to manage their own data about themselves. While Facebook's launch of the Beacon user-as-product-shill service was a PR fiasco, it familiarized millions of users with the notion that they can control information about themselves online and determine to whom it is visible. What might seem a horribly complex and tedious task to their elders - categorizing "friends," managing newsfeeds, handling intersecting communities of contacts - feels natural to the Facebook users of today. They want more granularity of control, not less.
Each user determines whom to let in to her own garden, whether it's friends or vendors. To illustrate, let me pick one example, Dopplr. (Not entirely at random: I plan to become an investor to help build out these functions.) Dopplr is a site for travelers. You can imagine or visit equivalent approaches for books (a hypothetical Amazon 2.0, a new and more personalized version with a user bookshelf feature), clothes (Glam.com and Stardoll.com), and even money management (Mint and Wesabe). Rather than promote their own content, these sites reflect the user's own content back to him or her. What could be more relevant? At Dopplr, I list my trips...and see how they intersect with my friends' trips. "Oh, we'll both be in London April 4? Let's get together!" Or, "Juan and Alice will be in town next Tuesday. Let's hold a dinner!"
So what's the business model? I'll "friend" British Airways, which will say, "We see you're going to Moscow next month. Why not fly through London and we'll give you 10,000 extra miles?" I'm no longer in a bucket of frequent travelers; I'm a specific individual with specific travel plans, which I make visible by choice to my preferred vendors. British Airways, of course, will pay Dopplr a handsome sponsorship fee to be eligible to be my friend (just as a local sports store or Nike rep might pay to sponsor a basketball game and be part of the community). Someday NetJets may show up, offering to ferry me and my friends to a conference we'll be attending together.
I'm far more likely to respond to BA or NetJets within a trusted site and for a specific offer, than I am to heed an ad from them while I'm reading the New York Times' latest report on the troubles in Russia. (As for Orbitz, my old standby: After five years, it still doesn't know my preferred airlines. Shades of the boyfriend who gave me pierced earrings back when I still had intact ears: I got them pierced, kept the earrings and lost the boyfriend.)
The new model creates a more trusted environment for reaching high-value frequent purchasers, whether of airline tickets, electronic gadgets, clothes or other item (but not with traditional ads; rather, with tailored offers and news of what they might want). Where does that leave the less-frequent purchasers? Probably looking to their friends rather than to advertisers for purchasing advice. I'm an expert on travel; my friends may look to me for hotel or pool choices. And when I'm in the mood to buy a book or a new computer, I'll check to see what my friends on Facebook are doing.
So the big news is already old news. It's not that traditional online advertising will go away....but its profitability will suffer as it becomes diminishing-returns efficient. Out of the spotlight, new value is being created somewhere else - in users' own walled gardens, which they will cultivate for themselves in real estate owned by the social networks and by highly targeted, user-data-rich vertical communities. The new value creators are not efficient ad networks, but companies that know how to create platforms like Facebook and communities like Dopplr.
Disclosure: Esther Dyson is an angel investor in a variety of Internet start-ups, including Orbitz (IPO'ed and sold several times), Flickr (sold to Yahoo!), Medstory (sold to Microsoft) and Postini (sold to Google). Her current (relevant) holdings or board seats include 23andMe, Boxbe, eventful.com, Evernote, Meetup Inc., WPP Group, Zedo and, soon, Dopplr.
This is background reading for my Op-Ed today in the Wall Street Journal.