Reflecting on my information seeking patterns in the context of discovering new music
Before I dive into the specifics of search and information behavior, I want to acknowledge how our discovery of new music has evolved over the years. Do you remember watching MTV endlessly or Shazam-ing to discover what’s playing?
Snapping back to last week, I was asked to choose 5 tracks for my wedding film. “Try to give us some Indian Indie music as well”, they added. That genre was unexplored territory for me, I would need to spend many hours searching for the songs that will forever frame the background of my happiest moments.
An interesting thought occurred to my student brain, “might as well get a case study out of this exercise”. So, I decided to conduct a diary study and observe my search behavior as I went through this process over the next few days.
Diary Studies
A diary study is a longitudinal, qualitative method that provides insight into the mundane activities of a user interacting with a given system. I used the Snippet technique to collect the details of my interaction with YouTube and Spotify over a span of 5 days. I also used note-taking, think-out-loud behavior, voice memos, and screenshots.
Here are some of the questions I aimed to answer through this study –
1. How do I frame my query? Does the platform help me with it?
2. What are my motivations to refine the query?
3. Do I lean towards a faceted search or an exploratory one?
4. How are the results presented to me by the system?
5. How do I share my information findings with others (in this case, my husband and the videographer)?
6. Are there any anomalies in attitude or actions that I have not observed in a similar interaction before?
The Information Foraging Theory
I used the Information Foraging Theory put forward by anthropologists Russell et al. (1993) and Sandstorm (1994) to analyze the collected data.
Platforms like Spotify and YouTube are full of information. The only way I would know what to click on and what to ignore is through the concept of ‘information scent’. I chose to click on a certain artist or playlist based on my perception or estimate of –
- How likely is the style of that artist in sync with what I want?
- How long will I take to find the song I’m searching for if I click on that album/playlist?
I was more likely to click on something if the information scent (surrogates like artist, album, playlist) matched my arbitrary estimate of value (language, genre etc.) that I might get from the information source.
But how did I know if a particular ‘patch’ would deliver the information I am seeking?
- Through surrogates like the cover image, song title, artist, album, genre, and tags like ‘Wedding’
- Quantitative data like number of songs in an album or playlist that I’m familiar with, number of likes/views by others.
- Checking if the length of the song is optimal for the length of my video (this was a secondary requirement)
Opportunity costs
As per the theory, every time I chose a particular song, it was a lost opportunity for another song. This led to the behavior of listening to a candidate song for only a few seconds to determine if it matched my requirements. This behavior is analogous with reading the summary of an article to see if it’s worth spending more time on.
I further divided my data into between patch activities and within patch activities, which helped me break down my observations for information seeking and satisfaction.
An example of a ‘within patch activity’ would be extracting information like lyrics, tone of voice, tune of the music from a song, to help me decide if it’s the appropriate choice.
Enrichments
Running out of time, I started to create ‘Enrichments’ or strategies that help me search efficiently. This was in the form of query re-framing and system recommendations.
Once I figured out that I can choose a song by watching other YouTube wedding videos by famous videographers, rather than using search keywords like ‘Indie wedding songs’, my search process and outcome were enhanced. The readymade playlists and ‘songs by mood’ feature on Spotify can also be viewed as an enhancement that I used to get the best out of my search.
Finally, I used the NASA Task Load Index to self report on the cognitive load for this end to end process.
My Takeaways
As researchers and designers, our understanding of search is often limited to elements like the search bar, autosuggestions or filters. Taking the time to look at the overall activity of ‘Seeking Information’ changes your perspective on something as simple as searching on an interface.
Further, analyzing the data as different components like ‘exploring’, ‘seeking’, ‘gathering’ and ‘using’ offers a fresh take on understanding the process of a user.
Hope this article motivates you to peek into your own search patterns!
Read the full article here
