Instagram "Dislike" Feature
Role: UX Designer
Time Frame: ~ 8 Weeks
Background: In this project I collaborated with 3 other designers to explore Instagram to see in what ways we might innovate or implement a feature that would improve the experience for users. Through research we settled on making the algorithm modification features already present on Instagram more visible to users via a "dislike" button.
Market Research
When exploring Instagram, me and my fellow designers found 2 notable features of interest to us that we thought we could improve. The first feature was the "Interested/Not Interested" buttons where you could sort a post into either in order to intentionally personalize your algorithm. The second feature was the "Content Preferences" setting where you manage your algorithm and preferences.
You can save posts into "interested" or "not interested" to indicate to IG's algorithm the type of content you do and don't wish to see.
The content preferences tab allows you to modify the type of content you want to see, including what you've saved into "interested" or "not interested".
Amongst ourselves, these features were not familiar to us and were only discovered after deliberate exploration. But of course, that does not automatically mean that these automatically unknown to more users, so we wanted to do some research to see how familiar other IG users were of them. In general, our focus became on how people directly or indirectly modify and influence their personal algorithms.
User Research
Research Goals
Goal One: Assess user engagement with suggested content
We figured that content management and algorithm preferences came into play when engaging with recommended content, outside of your main feed. That would mean suggested posts, IG reels and the discover page would be the main sources of user interactions we would have to study. Understanding how users deal with unwanted content would be key.
Goal Two: Gauge awareness of “Interested/Not Interested” feature and content preference settings
Next we wanted to see how aware users were of those features that we highlighted. If they were aware but did not use them, we would explore why. If users did not know of them, we would assess the reason for them not knowing and figure out if it was due to lack of interest, lack of accessibility, or both.
Goal Three: Assess algorithm accuracy and interest in algorithm management for users
Finally, we wanted to gauge how users perceived the accuracy of their algorithms and to gauge any interest in users being able to more directly modify it.
Summary of User Interview Insights
Insight One: When encountering unwanted or disliked content, users will primarily scroll past it.
Blocking or muting a post or account was less common and reserved for content that would provoke a strong negative reaction.
Insight Two: Most users were completely unaware or had little awareness of the highlighted features.
Many were not aware of the interested/not interested feature and the few that were, did not often use it. Most of them were also not aware of the content preferences setting on Instagram.
Insight Three: Most users believed their algorithms to be decently accurate, but also expressed a general interest in being able to modify their algorithms more directly.
Most users thought that their algorithms were at least decently accurate and suggested relevant content, but there were still times were it would give content they didn’t like or there were times it was boring or repetitive content. Many also expressed interest in either more transparency or direct control over their algorithm.
The Design
Based on these research insights, we decided that scrolling away as a method of indicating dislike or disinterest should remain. We don’t want to interrupt a process that works for users. But for those times where the content provokes a strong negative reaction, or times when their suggested content was repetitive, boring or uninteresting to users, we would provide a more visible and accessible option to modify your personal algorithm.
To simplify, we had 2 key design goals:
-
Implement a feature that would make content management more prominent to the user.
- To not disrupt the typical process of scrolling past disliked content when implementing it.
After brainstorming different ideas, we came up with the "Dislike" button.
The dislike button takes the shape of a heartbreak icon. When a user presses it, a pop up appears that shows the user relevant topic tags associated with the post. The user can remove any topic tags that they do not like or are no longer interested in seeing, and then save it to their content preferences.
This isn't a traditional dislike button in which the user is informing the original poster, but rather, something that only affects your personal preferences and does not notify the original post.
So we made a heartbreak icon that was on the opposite side of the “like” icon to give it enough distance for users to understand it was not related to the like button. Also, after pressing the button, a pop up would appear where you could remove suggested preferences and save that so your algorithm could be updated.
Prototype
This video showcases the user journey when a user engages the dislike button and saves changes to their contend preferences.
User Testing
Given our limited time frame, for our user testing, we focused on testing the usability of the button and the content preferences pop-up.
We gave users the task of disliking the "Road Trips" post and removing "travel hacks" from their content preferences and saving. All users found the task pretty simple to undertake.
Conclusion & Future Improvements
Overall, I'm proud of the design that we came up with, but of course, given the limited time frame and resources there are more things that I would like to test.
First, in terms of outcomes I would like the measure, I would want to understand how effective the content preference tab was in updating a user's personal preference. Another thing I would add to the content preference tab is to find a way for users to specify their reasons for wanting to "dislike" a post. During our user interviews, several people talked about why they would scroll past, mute, or block a post. Some reasons included boring or repetitive content, distaste or discomfort, or a dislike of the account itself.
Going into greater depth of the reasons for disliking a post would allow us to refine the content preferences tab so users can specify their reasons and the algorithm can more accurately update.