Personalization, Portability, and Ownership

"Arguing in My Spare Time" No. 2.16

by Arnold Kling

May not be redistributed commercially without the author's permission.

It has been said of Microsoft that their software products have to go through three or four iterations before they begin to get it right. Their Internet strategy seems to have followed a similar course.

As I interpret their current strategy, based on sketchy newspaper accounts, Microsoft intends to try to participate in the Internet economy by providing critical infrastructure components based on Internet standards. They want to provide software for instant messaging. They want to provide ways for businesses to exchange data, based on XML. They want to provide ways for consumers to buy online without having to register at each Web site from which they wish to make a purchase.

Prior to the recent reports, Microsoft appeared to me to be somewhat of a laggard in its Internet strategy. Now, it seems to me that Microsoft has moved in front of many self-proclaimed e-commerce gurus. If my reading of Microsoft’s strategy is accurate, they are trying to position themselves to participate in a process of personalization that differs from the conventional wisdom on that topic.

Several months ago, I walked out in disappointment during a presentation given by the founder of MicroStrategy, a young billionaire named Michael Saylor. He exemplifies one of my new cliches, which is "If you’re so rich, why aren’t you smart?" His vision of personalization is one of database marketing on steroids. Corporations will entice consumers to reveal their preferences, and this information will be pooled and sifted by data mining software to come up with laser-guided marketing missiles.

This vision of personalization as a tool to empower the corporate empire has obvious appeal to business executives, which explains why Saylor’s company has been able to achieve at least an aura of success. However, it is not my vision of personalization at all.

The personalization I want is Internet infrastructure that empowers me. If data about me needs to be correlated and analyzed, I want to be in charge of that process.

To take a concrete example, consider the concept of a web-based personal calendar. The basic idea is that instead of keeping my personal calendar on my computer, I store it on a web site.

What functionality should my web calendar have? According to the Michael Saylor vision, and according to some of the initial implementations of web calendars, my calendar should send me information about special events. For example, I should tell my web calendar provider that I like Bruce Springsteen, so that in turn my calendar will inform me when he is coming to town for a concert.

Instead, what I want from a web calendar is to be able to share information with friends and family. I want them to be able to access my calendar, and I would like to be able to obtain permission from them to access theirs. For example, when we are driving up to King of Prussia to visit our friends Allan and Dianne, it would be nice to know if Bernie in Swarthmore is free for lunch that day so that I can stop by along the way.

In the corporate network environment, Microsoft has developed this sort of multi-access calendar as part of its software suite. Although the requirements of the Internet are somewhat different, it is not too much of a stretch to suppose that they could provide useful tools there as well.

What I want from a web calendar is closer to what Microsoft’s corporate office product offers than to what Michael Saylor wants me to have. To me, the Web calendar is a piece of infrastructure that I use in order to communicate with peers. I am looking for communication and convenience, not for database marketing.

A big challenge for Microsoft, or for any company that shares the vision that personalization belongs in the infrastructure of the Internet, is portability. My ability to obtain the benefits of personalization should not depend on my being logged on to a given computer.

For example, consider the use of "cookies." A cookie is an identification tag that a web site stores on my computer. Cookies are what enable Amazon.com to say, "hello Arnold Kling" when I log on. Cookies allow me to take advantage of "one-click ordering."

When I switch to a different computer, this personal identity is lost. In the case of Amazon.com, I can recover this identity by going to a special log-in screen. However, this is inconvenient, particularly when one might be registered with many web sites and be using many passwords.

Another problem with cookies is that some computers are used by more than one person. I have often wondered, for example, what would happen if I went to the public library and used the browser to go to Amazon.com. What if some innocent patron, not understanding how cookies work, had set up an account? Could I charge a "one-click order" to this unsuspecting neophyte? Order a gift for myself?

To achieve portability, I believe that what we need is a "smart card" than can store the equivalent of cookies. With such a smart card, you could switch from your usual computer to an alternative computer that has a smart card reader. For that matter, Internet-enabled phones, TV’s, and other appliances might accept smart cards. This would allow you effectively to carry your personal information with you.

In addition to portability, I believe that the infrastructure must provide for consumer ownership of data. This belief appears to fly in the face of the conventional wisdom, which is that the Internet today is the arena for a contest in which corporations are fighting for control of customer data.

In the standard paradigm, Amazon.com is competing with Barnes and Noble for information about my book purchases. However, as a consumer, I do not want either company to own that knowledge. If I find that Amazon.com does a really good job of translating past purchases into current recommendations, then I may want to send them the information about all of my past book purchases, including those made at Barnes and Noble.

If there is to be a central repository of my purchase history, then that should be a service provided to me. I should be able to determine when and where that information may be used. I might pay a storage fee to the repository. Alternatively, the fees for the repository could be embedded in the prices charged by merchants with whom I transact business on the basis of that information.

There are web-based services that provide backup for your personal computer’s hard disk. If I use that service, the last thing I expect the service provider to do is assert some ownership right to the data that it stores on my behalf. Similarly, if I find it useful as a consumer to submit my purchase history to a data repository, I do not expect that repository to engage in database marketing or other behavior that implies ownership rights to that data.

Personalization should be a feature of the Web infrastructure rather than a characteristic of individual web sites. Personalization that is specific to a given web site treats the provider of the web site, rather than the individual, as the owner of the individual’s data. Therefore, site-specific personalization is nearly an oxymoron. A better alternative would be for standards and tools to emerge that allow individuals and businesses to interact on the basis of individual data, without taking the ownership of that data away from the consumer. To the extent that Microsoft is headed in that direction, the current iteration of its Internet strategy is relatively forward-looking.