Question

You have decided to build feature 'Sharing recommendations' for Netflix. What happens next?

Answer 1

I will first clarify my assumption for what "Sharing recommendations" means in Netflix - Sharing recommendations is a feature which allows users to share - 

a. My List/Recommendations (Curated list of movies/series which user adds to his/her list which he wants to share)

b. Netflix Recommended Lists ( Based on user streaming behaviour, Netflix recommends a list of movies/series which the user wants to share)

The feature will allow users to share the list of movies to another user/users. This helps in satisfying the core needs of the user such as

a. Social Validation - What are others watching?

b. Make my viewing experience in Netflix better by seeing what others are watching

Goal of this feature - Sharing Recommendations for Netflix is a feature which will help improve engagement in the platform. The North star metric for this feature would be to measure the Netflix Streaming time/user in Netflix.

Who is this feature aimed at - Netflix already provides a fairly long list of recommended movies based on personalized taste and recommendations. This is a feature which is aimed at active users of Netflix and improve their engagement in the platform even more. Active users are the users who will be interested in sharing and viewing what they are seeing and what others are watching.

Solution 

Approach 1 - Allow users to share recommendations/my list via social media 

Ability to share my list/recommended list as a link via social media 

Ability to view the list shared and add to my profile

Approach 2 - Allow users to share list publicly and allow other users to search and discover the shared recommendations

Ability to share my list as searchable and discoverable via Netflix

Ability to search and discover shared recommendations

These use cases need to be further prioritized based on - 

a. Impact on Customer

b. Impact on Business

b. Dev Effort/Implementation Cost

c. Time to Market

d. Other concerns like Privacy/Legal Issues (Since this is about sharing user recommendations , worth thinking about this as well)

Amongst these use cases, I will prioritize Approach 1 where users be able to share their lists via social media and anyone be able to view the list and add to their account for the MVP. Approach 1 also provides the maximum customer impact to minimum dev effort needed to build. 

Although Approach 2 is interesting , the key challenge in Approach 2 is the potential noise it would create if even a small percentage of Netflix active users start sharing their recommendations. Even if 1% of users start sharing their recommendations, it would be close to 200k lists shared which makes user discovery painful. (150 million subscribers. Assuming 20% of users are active users. Assume 1% of user start sharing recommendations)

Metrics to track

a. Feature Metrics

Streaming time increase/decrease due to the shared lists/recommendations

Click through rate increase due to shared lists/recommendations

Number of lists shared via social media channels

Number of times a lists has been accessed and via which channels(Facebook, whatsapp, messenger etc.)

Number of lists subscribed and added by user to their netflix account

b. Indirect Platform Benefits

Recommendations CTR%

Retention in the platform MOM

Launch Strategy for this feature - The way to launch this feature would be via A/B testing and tracking engagement metrics improvement for a test and control group before releasing this feature to everyone.

Comments

Hi, The focus of the question is how will you partner with other stakeholder to get this feature developed. Knowledge of working with engineering to build a roadmap, proper AB testing and evaluation of metrics, ability to pick the right metrics.

Answer 2

Not sure if this is the right approach to answer questions like this... but this is my take; feedback is welcome. 

Clarify questions

What does sharing recommendation feature do? It allows user to create a personalize list (URL) of movies and series and share it with friends via email; social media (FB, twitter etc)

What is the Goal of the feature? Goal is to increase user engagement on Netflix

Which platform? Available on Desktop & Mobile

Any specific country? US only to start with 

Users

Active users of netflix: users who watch movie or series on daily; weekly basis

Dormant users: Users who have signed up but would watch something say once in 2 months 

Use-Cases

As a user I would like to know what other people are watching to avoid FOMO

As a user I would like to contribute to the share playlist which seems to have similar interest as mine. 

(I am stuck here; can't think of more use cases)

Cut through prioritization

Let's take the first use-case as in order to measure second use-case; the first use-case is a prerequisite 

Solution

Solution	Impact to user	Impact to business	Effort to build
Allow user to search recommendations within the platform	high	medium	High
Allow share within the Netflix platform by providing share functionality via the FB; Twitter link; email option	high	medium	low (as open source apis exists)

Based on above; I would choose approach 2.

Metrics

* How many people are creating playlist/recommendation list

* How many people are sharing the list via social media / emails 

* How many people are clicking on the list; CTR of the shared list

* How many people are watching via the list ; CPA of the shared list

* Avg user watch time who clicks on the playlist - has it increased?

Answer 3

Clarify (Answers in italics)

1. What is the feature? Sharing recommendations allows users to share content with others. No details on how this is done. This can be done through lists, individual shows/movies, direct message, etc...
2. Who is this feature for? No specific target
3. Monetization? Not for now

Personas/Problems

1. Casual user: shared recommendations to, and can use recommendations to find relevant content
   1. Finding relevant content can be difficult
   2. Interacting ang improving upon Netflix's proprietary algorithm doesn't always help
2. Power user: creating the lists/media to share with others
   1. Organization of content they find
   2. Sharing content they find
   3. Measuring visibility and interactivity of their shared content

Prioritized Needs & Potential Solutions

Listing out the needs and solutions in priority order. We start with the sharing side first, and then move to the viewing. 

1. Sharing content
   1. Direct shares with individual users
   2. Categorization based shares (eg. sharing a movie into the #horror hashtag for everyone to see)
   3. Network based shares (based on your friends and network) - but doesn't make a lot of sense for Netflix
2. Organizing shared content - how to easily share lots of content?
   1. Create shared lists
   2. Easily allow users to share this content through easy to use named features and consistent UX
   3. Allow users to share consistently with the same "group" of users
3. Viewing shared content 
   1. Exploration of specifically shared content
   2. Surfacing shared content algorithmically to users

Measure

1. Shared content per user - measure the number of average shares
2. Viewed shared content per user - measure how much people are actually using shared content
3. Shared content view time - aggregate Netflix metric for how sharing affects total time spent on Netflix