Managing Your Attention Portfolio

Bradford Cross wrote a thought provoking blog post today on the idea of anti-personalization. I encourage you to read his post. In summary, he discusses the concern raised by some that personalization may have negative consequences on our perspective, by only serving to reinforce our current beliefs with additional content that is similar to what we’ve previously consumed. He then shares his initial thoughts on ways to potentially break that cycle.

His post triggered a number of thoughts in my mind so I thought I would capture them here in an attempt to organize them and enable discussion.

Our natural inclination toward information sources supporting our current beliefs has an interesting new twist on the social web. Increasingly we are openly sharing content that we find compelling. We are signaling to others aspects of ourselves through the content we share. Those actions have consequences. We all have an innate desire, whether acknowledged or not, to both get along and get ahead. How we signal impacts our status in our social environment, whether online or offline. My sense is that we must be cognizant of these influences when designing future tools for filtering and prioritizing data streams. In a world where increasingly much of the available content is user generated, the source of the content may be as important as the content itself.

Brad used the term serendipity repeatedly when talking about those unanticipated discoveries of compelling content. He posited that maybe we can learn from serendipitous discoveries to find other compelling documents for a user. My sense is that a document-centric, predictive approach is not the most fruitful direction.

Here’s how I would recast the problem.

Let me start by asserting my belief that these serendipitous discoveries are inherently unpredictable. Such discoveries in my mind are positive surprises. They are unexpected. If one accepts the futility in trying to predict positive surprise, how do we move forward in maximizing our exposure to positive surprise?

I believe we need to shift our focus from documents to sources. To gain some perspective on this, let’s discuss the decisions each of us make when using Twitter. Twitter is powerful because it allows us to manage our attention portfolio in that information space in many unique ways. Every time we choose to follow another Twitter user, we are making an allocation of attentional resource to content they produce.

Managing your attention portfolio on Twitter (the set of users you follow) is similar to managing a portfolio of stocks. By looking at the user’s past performance, you make some estimation of your future gains without knowing whether the current trend will continue. You risk some of your attention so that you are able to capitalize on potential future rewards.

Right now, given current tools, my decision to follow someone is based solely on their past tweets. I scan through the content and make a decision about whether their value proposition is worthwhile. I focus only on this assessment because I can’t easily do anything more sophisticated.

How could we design a better attention portfolio?

For one, take into account the social context. Twitter users cluster into communities and information propagates. To access compelling content, I do not need to follow everyone within a given community of interest. I can capitalize on the flows of information and diversify my attentional investments. Communities are analogous to sectors in the stock market. Do you want to be heavily weighted in a basket of correlated stocks from a given sector? Or would you rather diversify across sectors? I have no easy way of making that assessment in the Twittersphere currently. No client supports that.

It’s possible that most don’t want that level of control even if it was available. But some might. I think we need to build a technology probe to flesh out the concept. Those insights will help us find the way.