Offline it’s easy to infer our relative level of interest in an object: the dog-eared book on the shelf is more of a favorite than the book with the unblemished spine that sits beside it; the movie that we own on vhs and dvd is dearer to our heart than a recent rental. On the basis of an individual object, our interests can be revealed by clues introduced through physical consumption.
It’s difficult, however, to deeply understand a person’s interests through the aggregation of offline attention. Certainly one can deduce a favorite genre by glimpsing a photo of another’s bookshelf, but free from additional insight - the kind generally amassed by our closest friends - understanding someone’s likes by viewing the aggregate of their objects reveals only an outline of their interests. To fill in the picture one needs to understand another on an intimate level.
What about online?
On an individual level it’s again easy to infer the level of interest in an object: I’ve added this to my Netflix que, purchased this on Amazon, and reviewed this on Yelp. Add to this the implicit web that can automatically track and aggregate actions, yielding stats for interests that were previously approximated by physical clues and there’s a good ability online to understand the relative level of interest in an object.
Interestingly, while it was required to intimately know an individual in order to understand their interests through the aggregate of offline attention, online, a computer and algorithm - antonyms of intimacy - can produce a rich description of one’s interests through aggregation.
Enter the Hierarchy of Attention.
For a given vertical, let’s say books, an algorithm containing the aggregate of our online interactions with books can produce an accurate picture of our interest in the category. Applying the correct weight to the coefficients of associated actions - reviews, purchases, implicit views, recommendations, comments, etc. - produces an algorithm that reveals interests. Organizing the associated actions by coefficient yields a hierarchy of attention - the order of importance of online attention as it relates to the revelation of one’s interests. A rating is more revealing than a purchase, a purchase is more revealing than an implicit view. And so on.
Building the hierarchy of attention is critical to constructing an image of one’s interests online and understanding one’s interests through the aggregation of online attention is a powerful concept. And until there’s a web of things none of this is possible.
