Using data to uncover how users perceive algorithms on Instagram — and its influence on the user experience.
In the spring of 2016, the popular image sharing and social media platform Instagram announced a future implementation of personalizing algorithms on their platform. This quickly sparked a fierce debate among Instagram-users on Twitter and Instagram, about how this would irreversibly change the user experience of Instagram. Together with my fellow student Josefine Grunnet, we started to collect massive amounts of data from this debate, in order to later explore and analyze how, and indeed why, users reacted so strongly to these proposed changes.
This led to some very interesting insights about how users perceive algorithms on social media, and later also reflections about how design methods can be utilized in order to create a better understanding of the relationship between user and algorithms — and therethrough — improve the user experience.
The following will give an overview of our research process, our insights and reflections about how the implementation of algorithms relate to the user experience of Instagram — and social media platforms in general.
Context
On March 15th, 2016 Instagram announced via their blog that they would move away from their traditional chronological feed and instead begin to order users’ feed based on the likelihood that a given post would be of interest — i.e. they were planning to implement a curating algorithm on their platform.
To improve your experience, your feed will soon be ordered to show the moments we believe you will care about the most. — Instagram
Following the announcement from Instagram a debate quickly developed on Twitter and Instagram using the hashtag #RIPINSTAGRAM insinuating that, for these users at least, this would be the end of Instagram.
This presents a intersting case to dive into, as these negative outcries from the users of Instagram reveal some very useful insights, which can be used when designing and implementing personalized feeds on social media and digital platforms.
In order to analyze why some Instagram users reacted so strongly to the announcement of a personalized feed, we collected roughly 4.000 tweets and 4.700 Instagram posts containing the hashtag #RIPINSTAGRAM.
Findings
In the case of personalized content and curating algorithms, the algorithms work by receiving various forms of input from the users, and based on these inputs the algorithms generate a personalized feed. Of course the technological specifics of the algorithm has some influence, but so does the user, as it is the data from the user which determine how the feed will look like. Thus, put very simple, the user and algorithm take part in a mutual relationship where they both help to shape the feed.
However, contrary to the idea of a mutual relationship between user and algorithm, the (negative) perceptions held by users all present the algorithm as taking sole control over the feed and removing the influence of the user.
(…) now they’re going to use an algorithm to show us photos that they think we will like. It is impossible for an algorithm to know anything about what our brains are thinking or want to see while we see this content. When Facebook implemented this strategy they used it just to sell advertising. Completely inhibiting our ability to discover new content and art. — Instagram user.
Analyzing the debate, we found four key insights about how users perceive the effect an algorithm would have on the current user experience of Instagram:
- The algorithm will control everything you see as a user.
- The algorithm removes all agency currently held by the users.
- Pay to play — the algorithm will lead to the invisibility of small companies and artists, and instead favor ads and sponsored posts from big corporations.
- Purely based on likes, the algorithm will favor already popular and mainstream content (such as make-up tips, puppies and cats) over new creative approaches — and make “everything the same”.
The introduction of an algorithm into the personal feeds is perceived as removing the influence of the “ordinary” user, the small business or the underground artist and instead promoting big businesses. Although the users have the ability to actively shape the algorithm through their actions, they seem highly unaware of their own influence, which in turn creates a feeling of being shut out.
My reflections — going beyond research
However, how does these findings relate to the user experience of social media platforms — and furthermore — how can we utilize these insights when we go beyond research and start designing?
How can we utilize these insights when we go beyond research and start designing?
Coming from a user-centered design background, it is obvious to me, that if the aim is to improve the user experience, designers and developers need to take these insights into consideration when they are designing and developing new algorithms for their platform — especially on platforms that rely on user-generated content and social interaction.
These insights highligt a new opportunity for developers and designers to improve the user experience and implementation of algorithms on social- and digital platforms.
My suggestion is a move within the design and implementation of personalized feeds. A move that goes beyond the usual black-boxing of algorithms and instead embrace the prevalent ideas of user-centered design and empowerment — And through user involvement and visibly interactive design, designers and developers should enable and encourage users to engage with algorithms and equip them to actively shape their own news feed.
In order to avoid a situation where users fail to see their own potential in shaping and changing the content they are presented with, it would be optimal to involve the users throughout the process of designing, developing and implementing algorithms on social media and digital platforms and not leave them feeling as passive bystanders.
— thereby companies, such as Instagram, do not simply utilize advanced algorithms to personalize the content shown to their users, but they also maintain (or even improve) the user experience of their platform
Final notes
Following our analysis of this debate and the users strong reactions to the announcement from Instagram, our study and findings have been re-written into a blogpost and later as a peer-reviewed paper co-authored with Martina Skrubbeltrang Mahnke — a researcher at the IT University of Copenhagen.
