Joel Liu's Library tagged Content → View Popular, Search in Google

Nov

2010

The Myth Of Serendipity

techcrunch.com/...myth-serendipity

- Preview

Content discovery serendipity personalization

In seeking to achieve serendipity, the individual reader becomes both the target of content delivery mechanisms and the genesis of what that content may be. This is why serendipity is so closely associated with personalization—it requires a high-resolution understanding of the user.

Serendipity and personalization are in fact two sides to the same coin. Personalization merely acknowledges intimacy, whereas serendipity pretends to have happened on it as if by accident.
Serendipity is really just an informed calculation based upon any number of our individually unique interests, habits, location, the time and date, and prior knowledge. This level of relevance is, of course, what the emerging personalized Web hopes to achieve for each user, whether for recommendations (GetGlue; Hunch), marketing and ads (Rapleaf; Facebook advertising), or news and content (my company, TrapIt).
3 more annotation(s)...
- Crowdsourced Serendipity
  
  Bridging the gap between editorial and social serendipity, the notion of crowdsourced relevance really only delivers a broad, lowest-common-denominator level of content discovery. While not without its usefulness to the degree that we want to be aware of what is most popular and most talked about, the trade-off is the lack of personalization.
  
  The pro here is the viral component, which makes up a great deal of our online content-discovery routines. Crowdsourced serendipity provides a tier of distribution in touch with a larger zeitgeist, from trivial cat videos to important broad-based news. The downside is that the lowest common denominator lacks any precision and therefore has limited utility.
- To some extent, this recognition is obvious. Our interests change and evolve over time. Yet for the kind of precision that seeks to provide consistent serendipity in the ways we have been discussing, the indicators need to be equally sensitive.
  
  The content that I want, and better yet, the content that I don’t even know that I want, is an ever-changing proposition based on any number of factors. To achieve that level of sophisticated customization requires a sensitive understanding of context for any proposed “serendipity engine”, both a context of the content and the user.
- Recognizing that serendipity is a constantly moving target of context, the best we can hope to achieve are fleeting moments relevance.