Gift Profiling

This past month, I have purchased quite a few gifts for my loved ones. It is a time consuming process (and my excuse for having such a long lapse between entries), but I really enjoy giving good gifts that are thoughtful and also useful. For some of my closest family, though, I can have a hard time finding gifts that suit my criteria -- the more you know about a person, the higher the bar is for thoughtfulness, and for someone like my mom (a writer) or brother (a historian), I may have only a cursory understanding of their real interests. I can only imagine it's just as hard for them to find things for me. This is one area where portable attention profiles, often refered to as APML, could really make things a lot easier.

The big idea with APML is that you will be able to see, control, and distribute the machine-created profile of your interests. If you browse Amazon while signed on, you've seen how this kind of data can be used, with their recommendation system. And the idea isn't terribly hard to understand: if you have some idea what a person has liked in the past, and you know what other people like, you can do some magical statistics to make recommendations for other thing the person would like. I know that for me, it's almost eerie; Amazon magically manages to suggest lots of things I already own and plenty that do look dangerously interesting.

Now, imagine if you could share your attention profile with friends and family. Then, armed with what you've been looking into and buying recently, what millions of other internet users have done independently, and their knowledge of you, they'd be ready to find on-target gifts that tightly match your interests. It would be like having your Amazon account without actually being able to buy things on your credit card.

If your social graph were also portable, I could almost see a social network emerging just for holiday shopping -- you sign up, select who can see your attention profile, and maybe even make a cross-retailer wishlist to help people along. Then, the network could keep track of who got you what (via something like Google's new Universal Activity Stream), keeping the gifts a secret from you but concurrently preventing duplicate gifts. Those of us who don't want to go out on a limb and skip the wishlists could still follow your interests and get you things you're likely to find useful.

Of course, there are still problems to be solved. As I wrote about before, social graphs and attention profiles are complex creatures that will need a lot of nuance to work correctly. Additionally, I suspect most people's attention profiles will require a little editing before they're willing to send them to their mothers, which means hard work designing and automating that process. However, the future looks interesting, at least.