It's hard to go wrong with personalized recommendations and the thumbs up / thumbs down is a useful feature for that. However, it isn't tailored for browsing as the engagement doesn't produce immediate results. Rather, user has to wait for the algorithm to come back with recommendations.
'More Like This’ stands apart as a remarkable feature that lets the user drive their own exploration journey without forcing them to articulate their subjective preferences. The user can choose the movie that they liked and can actively browse based on their preferences. The recommendation algorithm takes the role of a smart assistant helping them along the way until they find their favorite movie to watch.
“My movies don’t fit into specific genres or categories. I like certain elements of a movie that I am unable to describe, but I know the movie is good when I eventually find it.”
The users cannot always articulate their preferences. They might have a vague idea about what they are looking for and know when they have found the movie that they will like. Letting the users kickstart their journey with this feature will be a great asset for Netflix.
As Netflix tries to expand in emerging markets, there will be cold start problem for the recommender system as there isn't enough data for the newly acquired users. 'More Like This' will be an active way for users to browse based on their favorite titles, and this can be a strong indicator of their preferences.
Prime Video Explorer comes closest to the More Like This feature, with an added ability to browse based on multiple titles.
Other features considered were Pinterest's More Like This, Google Image Search, Amazon Instant Recommendations, Wayfair's View Similar, Myntra's (leading online apparel retailer in India) View Similar and Related Items.
... enable user to indicate which attributes matter to them?
... enhance discoverability of 'More Like This' in browse results?
I created 5 guiding principles before starting out on the designs to remind myself of what I am trying to accomplish in the end implementation. These also help identify best design among explored options.
To tackle the first HMW, I mapped out the user flow to set context of a search leading to the More Like This page. Here I show a snapshot of explorations for More Like This page starting with the layout explorations and preference filters.
Preference filters in combination with grouped recall (option # 4) is the design direction aligning with the principles of Simple, Clear, Contextual and Informative. I decided to iterate further along this direction.
I designed the More Like This feature to enhance the user control in accessing movie title and helping them express their nuanced preferences. The feature is best accessed on mobile as it is well suited for browsing new titles when the user is on-the-go.
Take a look at how I envision the enhanced More Like This feature to work.
Managing design changes designs especially when I disagree with changes.