I’m tired of seeing Netflix recommending something I’m not interested in. However, I can do nothing to make a change. Gradually, stuff that I don’t like occupied my homepage, and sometimes I quitted Netflix because I can’t find what I want there.

I searched online to see if others share the same concern. And it turned out to be a great number of complaints, with different contexts but all leading to the same unsolved problem.

In order to further understand user scenarios and pain points of undesired contents or recommendations, as well as their expectations with some new features, we conducted in- person interview with 7 Netflix users and defined the problems and needs.

We found out 5 major problems:
1. Undesired recommendations in different categories
2. Confusing “Continue watching” (display a show that have completed and automatically remove it)
3. Unable to remove undesired items from lists (i.e.“Continue watching”,“Recommendation”)
4. Disordered and confusing “My list”
5. Difficult to locate a particular movie/TV show on theNetflix app

We brainstormed some solutions that can tackle these problems:
Feature 1: Enable user to organize “My List” - P1,P3, P4, P5
Feature 2: Add list “Watch history” - P1, P5
Feature 3: Enable user to manage recommendations - P1, P3, P5 Feature 4: Enable user to manage “continue watching”- P1, P2, P3, P5 Based on the graph below, we decided to focus on solutions F3 and F4.

The final deliverable is a personalized feature that helps users effectively manage their recommendations, whileoffering a higher likelihood of exposing users to contents that they prefer.

Since “Continue watching” is generated by user behaviors while recommendations are developed by Netflix, we made the flow and design separately.

To find out the optimal solution of our design, we explored different kinds of interaction. We came up with various ideas and select two of them to design. I was in charge of designing the Solution 1 and Zoe created the Solution 2.

We leveraged simple interactions such as long press and tap in Solution 1.

We selected solution 1 between these two solutions, since its interactions are simple, accessible (easy to learn, use and remember) and intended enough.  

We wanted to iterate out design to further understand user’s intentions of removing an item and help Netflix to adjust recommendations. We optimized the design by allowing collecting user feedbacks and providing user inputs after an item was removed.