New Metaphors, New Models
Down the YouTube Rabbit Hole
I have been guilty of plunging into those YouTube ‘Rabbit holes’ before, and it is definitely an odd experience when you think about it. It might have started out on a tutorial or music video, but about 2 hours later, I was watching a guy get a haircut using a bowl or a lady shoving cucumbers up her nose instead. How did I get here? The recommended videos algorithm has a way of not only “expanding” (as YouTube describes it) our interests, but also keeping us glued to the screen. It is like very much like going down a tunnel where 20 minutes turn to 2 hours before one can snap out of it.
Over the past years, I have become more conscious of this and also very curious about the algorithm. How does it work? Apparently, it is designed to mimic the brain, and I do observe a parallel with the YouTube recommended rabbit hole and how a random chain of thoughts might unfold in our mind. Idle thought can wander aimlessly to great lengths, but is seldom productive.
So thinking about metaphors and mental models, I decided to use this opportunity to explore our interaction with the YouTube algorithm. It is the algorithm that manifests on the recommended videos and modulates user behavior. Often, I see people in the comments section of some obscure video claiming, “How did I get here?” or “It’s 3 am, and I have no idea why I’m watching this lol.”
Ideation and Brainstorming
I framed my challenge as — how can we think of new ways to navigate through YouTube videos? How to not get lost in the recommendations? The feeling of being lost as a recurring user experience is something worth redesigning. Also, I think it might be beneficial to explore how we can interact with and explore our own behavior in relationship to the influence of an algorithm. I learned more about how the algorithm drives (and is driven) by behavior from this article by Edward Muldrew.
To ideate and brainstorm, I used Dan Lockton’s Design with Intent Toolkit. I picked some cards that prompted me to think about navigating through a digital interface in a different perspective. I stressed on the word “navigation” in particular and explored different metaphors associated to it. These were few of the cards that I took inspiration from.
Thinking of “navigation” in terms of maps, branches, desire paths, and real-time feedback like a GPS… sparked some new ideas for imagining how Youtube might look like alternatively.
Since the YouTube algorithm already maps what we might click on, we can use that to generate a “landscape”. The landscape is another metaphor that can be drawn from three (or more) composites — user’s past activities, content categories associated to user, and social trends. “Desire paths” could be traced out of watched history. Path or progression is a metaphor for time spent on the platform, i.e. journey undertaken down the YouTube Wonderland. Desire corelates to search and choice.
Sketching
Through the sketching process, I thought about how I could visualize an algorithm that has no inherent form but takes on the personalized form of the user’s behavior over time and other trends and metrics on the platform. Ultimately, each person would end up with their own “landscapes” and maps of their usage. Some tools like a timer and total number of videos might be a good way to keep track.
Usually in such browsing sessions, the user has no destination and the motivation is to discover content. But that is not the whole picture. People also spend too much time, and motivation can also be to seek distraction, not content in particular. For such instances, I think it is very important to start building a more apparent relation with the algorithm — because the algorithm holds the strings we are attached to whenever we use the platform. What the algorithm does must become more transparent and we need to find new ways to understand how it influences us.
