On serendipity

One of our more recent NoTube demos was inspired by an XKCD cartoon which prompted us to ask: can we make a large video collection interesting enough so that people keep browsing rather than give up? As we’ve said before, many people find choosing in large collections difficult. Although there is clearly a place for search in finding known-item video content, the user has to know what they’re looking for.

The aim of the demo is to surface niche and diverse programmes of interest buried in the long-tail. It is designed to aid serendipitous content discovery by helping people to browse better by following interesting connections between programmes using content-based links. In this respect, it is the polar opposite to choosing what to watch based on what’s popular. The connections are created using a BBC-specific subject classification system, traditionally used by professional cataloguers for subject indexing/classification as part of the cataloguing procedure in the BBC’s internal TV and radio programme catalogue. The classification system reflects the last 50 years of BBC programming: it is extremely granular, TV-centric and includes numerous quirky terms.

For the purposes of the demo, we are using the classification system as a form of Linked Data. From any programme the user can follow suggestions for similar programmes based on the number of common classification links between them. The idea is to support “hours of fascinated clicking” through the video collection, similar to the way that following links in Wikipedia articles can take users on surprising and unexpected journeys through the content.

Screenshot of NoTube browser

NoTube demo for browsing a large video collection

Libby and I are currently in the final stages of preparing our demo for user testing, to see if it really does keep people clicking and helps them to find new programmes. This planning has led me to ponder the rather vague and ethereal concept of serendipity, commonly defined as the pleasure of unexpectedly making a fortunate discovery while looking for something unrelated. Serendipity is often cited as a key benefit to end users of apps built on Linked Data, which aims to connect things better to create new – and sometimes surprising – links. Given the positive emotions of delight and surprise associated with discovering something interesting and new by accident, there is clearly a role for serendipity in increasing user satisfaction. However, in designing our user study, it soon became clear that measuring the contribution of serendipity is not straightforward.

The merits of the online versus the offline world for offering opportunities for serendipity is a topic that is hotly debated. For example, last year’s news of the probable decision not to print the next edition of the Oxford English Dictionary, but make it available online only, triggered some lively discussion about implications for serendipity.

As some have pointed out, online recommendation systems cannot really provide true serendipity, because serendipity occurs by chance, in circumstances where there is no expectation or intention:

We may stumble upon interesting songs, pieces of news or movies using these tools, but we will be expecting it. By intentionally looking for these events using the aforementioned tools, it’s no longer an unsought event. (From The Humanising Technology blog)

However, all commentators appear to agree that experiencing serendipity, whether in the real world or online, is a ‘good thing’ for exposing us to a healthy diversity of opinions and ideas.

Since the NoTube project is concerned with Linked Data and Recommendations as research topics, the theme of serendipity is of particular relevance to us. As I mentioned in a previous post, one of the things we want from a good recommendation system is for it to help us broaden our horizons and discover new, enjoyable things that we would probably not have discovered on our own. The second of the user evaluations within our part of the NoTube project, aims to look more closely at which specific aspects of the connections between programmes afforded by Linked Data are actually of most interest to people.

This entry was posted in Demos, Linked Data, Recommendations, Thinking Out Loud, User Experience. Bookmark the permalink.

5 Responses to On serendipity

  1. Pingback: Tweets that mention On serendipity | NoTube -- Topsy.com

  2. Pingback: VocabControl » Serendipity and large video collections

  3. Pingback: Conversational Metadata | Conversational Aesthetics

  4. Tim says:

    Jesús, in his blog that you link to, suggests cultivating serendipity by understanding the user’s background, presenting relevant resources related and unrelated to the areas the user is expert in, and helping the user look for information from different points of view. Granted, doing all this makes recommendation algorithms harder to understand and implement, but those sound like plausible approaches.

  5. Pingback: Preliminary findings of N-Screen user testing | NoTube

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s