Let's pretend for a moment you are a product manager at Netflix, and your task is to add a feature that allows users to video chat (like Zoom or FaceTime) with someone and watch a synchronized stream of a show together. What would that spec look like before we got started with it?
Here's how I would write it (where I don't know things, I'm just making it up):
Watch Together Product Spec
BLUF
Netflix users frequently express a desire to watch shows with friends and family remotely while engaging in real-time discussions. Previous attempts to implement such a feature have not met the desired results. The goal is to provide users with the ability to watch Netflix content together, regardless of their devices, and have a synchronized video chat similar to FaceTime.
Problem Statement
As a watcher of Netflix shows, I want to be able to watch with a friend the same content at the same time and talk about it.
A feature that Netflix has frequently attempted but has yet to get right is having living room experiences where watchers can chat about a show in progress with their remote friends and family. We have tried just synchronized video, watch parties on Xbox, leaving text notes for future watchers, and more, but have yet to see the results we want. Yet, it is still one of the most common requests from users.
This feature attempts to put the best of all of these and put them together so that our watchers can have a true living room chat about their favorite shows as they are witnessing it together. X% of our users are watching Netflix on video-chat enabled devices, and we are going to leverage that as they have a video chat (think FaceTime-like) while watching a synchronized video.
Goals and Non-Goals
Goals
- Allow two people to watch a show together
- Allow users to watch together regardless of their device as long as it has video and/or audio
- Allow users to watch shows together as long the show is available in both users' region
- Allow users to pause, scrub, and skip shows and have it work for both users
- Allow users to set their own subs and dubs, which can be different for each user
Non-Goals
- Allow more than two users
- Solve for users that don't have video devices
- Encourage users to chat without watching a video
Key Metrics
- Time spent watching Netflix using Watch Together
- Time spent in chat total
- Users who use this feature within a week of signup: indicates a user may have signed up to use this feature
- Month-over-month retention on feature: indicates users find it useful to use it more than once
- CSAT: after users finish a chat, a one-question survey will ask them how it went. If bad, we will ask why.
Rollback Criteria
After three months, we will look at initial metrics to see if users are rating it well and if they are using it more than once. If it looks good, we will promote further and follow up. If it is not looking good, we will meet to see what we can improve. We do this again at six months.
If, by nine months, we do not see X% growth, X retention, or X CSAT, we will begin winding down the feature.
Timelines
Made with time.graphics
| Event | Date |
|---|---|
| Product Review | Sept 13 |
| Beta Release | Oct 25 |
| General Availability | Nov 26 |
Mock UX
Note: this mocks the user experience but not the final designs.
User opens their app to the show they wish to watch with a friend. They click on "Watch Together"
A show modal shows up. User invites their friend to watch with them, and the friend accepts
User watches a show with a friend
Photo of woman by Jake Nackos on Unsplash
Designs belong to Netflix with only minor edits by me
Outstanding Questions
Legal
Both users of this feature will have their own independent subscriptions to Netflix and watch distinct streams that function the same. However, we need to investigate if the act of synchrony amongst watchers in different places is a legally distinct action according to our contracts with content providers.
WebRTC, Vendor, or Build-It
We still have not settled on how precisely to stream the video chat. Having nearly the same predictable latency as the Netflix video will be important because if the sync drifts too much, it will be annoying to users. WebRTC is preferred as it would peer the connection and therefore cost nothing to Netflix to run. We could also buy an off-the-shelf solution such as AWS Chime. Lastly, we could build an in-house solution to save costs and make the stream the most predictable.
