How would you design a recommendation system for Disney+ for new customers (with less data)?

Question

How would you design a recommendation system for Disney+ for new customers (with less data)?

Asked at Walt Disney over a year ago

1.5k views

Product Design category

Asked at

1.5k views

How to answer Product Design questions

Interview Guide

Answers (2)

To answer this question, Log in.

Top Product Design interview questions

How would you design a bicycle renting app for tourists?62 answers | 82.5k views
Build a product to buy and sell antiques.54 answers | 66.8k views
How would you design a web search engine for children below 14 years old?36 answers | 42.9k views
See Product Design PM Interview Questions

Invite members
Invite by email

Add another

sahil kapoor · Answer 1 · 2024-04-15T22:37:20+0000

Clarifying questions

Are we talking about recommendations where the user is just signing up for the first time or a general recommendation system?
The goal is just to focus on new users interested in the platform more, and leaving to increase retention. yes
Are we focusing on mobile apps to web app

User Group

New User -

1. Based on the age

2. Based on the Activity

Based on Activity

1. Power User - User 1 streaming app Dily

2. Semi Occasional User - User Streaming services 1 week

3. Occasional User - User Streaming services 1 month

I think for this example, I would like to focus on Semi Occasional User

Needs

1. They are looking for movies and TV shoes

2. They are inserted in particular content ( licensed and owned by Disney)

3. They want to discover new content that they like

4. They want seamless integrations and want to watch on phone, TV, etc

5. Seamless Payment

6. They want a quick recommendation

Priorities User Needs

Must have

1. They are looking for movies and TV shoes

2. They want to discover new content that they like

3. They want seamless integrations and want to watch on phone, TV, etc

4. Seamless Payment

Should have

1. They want a quick recommendation

Could have

They are inserted in particular content ( licensed and owned by Disney)

Solutions

1. Build a survey recommendation to gather what content the new user likes when they sign up - Comedy,

2. Starring collecting data on the show they searched and watched and start recommending them on that interest (genre, TV show or movie, etc)

3. Introduce a feature where they can see what their friends are watching

4. Recommend them Disney + exclusive and licensed content

Prioritizing solution

Solution	Reach	Impact	Effort
Build In Rec	High	High	low
Search and watch data	High	Hight	Medium
Friends and Family Sharing	High	Medium	High
Disney + exclusive and licensed content	Medium	High	Low

Based on the above - I'll build for 1,2,4 first

Tradeoffs:

Survey-Based Recommendation:

Tradeoff: While surveys can provide valuable initial insights into user preferences, they may also introduce friction during the onboarding process if not carefully designed.

Consideration: Balance the depth of the survey with the need for simplicity and user experience. Long surveys could deter users from completing the sign-up process.

Data Collection and Privacy:

Tradeoff: Collecting and analyzing user data to personalize recommendations is powerful but raises privacy concerns.

Consideration: Ensure transparent data handling practices and offer clear opt-in/opt-out choices to maintain user trust and compliance with privacy regulations.

Friends and Family Sharing:

Tradeoff: Implementing social sharing features can enhance engagement but requires careful management of privacy settings and user controls.

Consideration: Allow users to customize sharing preferences and privacy settings to accommodate varying comfort levels with social interactions.

rohit kumar · Answer 2 · 2024-03-01T10:57:51+0000

For a new customer in Disney + we might not have enough historical data to provide him/her with appropriate recommendations for shows or movies as we can do for other old users for whom the recommendations are based on their past browsing/streaming/renting data.

This is called a "user cold start problem" in terms of the recommendation system and it could be mitigated by using one or a combination of the following options:

1) As part of user onboarding, the user can be asked a few questions to baseline their viewing pattern.

Eg.

a) A multiple-choice question for the type of content they like to watch: comedy, thriller, etc.

b) Which sports do you watch

c) Favourite actors/actress

2) Using their profile information to categorize them with similar viewers and use the viewing history of the group to recommend similar content to the new user

Eg. Age group, Geographical location, Gender, etc can be used to find similar viewers whose historical data could help provide viewing recommendations to the new user.

Once the new user starts interacting with the system, his/her own viewing history shall be created and henceforth used for further recommendations