Before I jump into improving the video experience in FB, I would want to clarify few things.
1. What it means by video experience? Is it video calling experience or the video streaming that are being posted on FB?
2.Improve video experience does that mean the current experience is not up to the target metrics? Do you see any gap? What is it?
Interviewer:There has been few difficulties which we have been facing and ultimately the customers are not happy.
1. User satisfaction.(NPS Score)
2. User Retention. (I will park this for now and talk about it more when the need arrives.)
FB is a social networking site whose goal is to make the communication and collaboration hassel free.
And Video calling/streaming helps in attaining its goal by knowing the people better and improving their relationships.
1. Frequent Users:
Who check FB at least twice in a day and perform some activity like browse through news feed, like or tag friends , post pictures or status on their wall.
2.Non frequent users(Silent users):
Who check the FB once or twice in a week and spend minimal amount of time on it(~10-15 mins). These kind of people are silent users who are not into doing some activity but simply browse through the feed.
3.Non techy users: There users are someone who are new to FB or using internet like grandma’s and mom’s but are interested in learning more and stay connected to their loved ones and friends.
I would like to scope down the users to 1 as we are working on improving the video experience which would be more applicable to active users.
1. Choppy video streaming.Low quality/low resolution videos.
2. More buffer time which makes user impatient.
3. Poor bandwidth/ network connectivity.
4. The device that users are trying to stream videos.(mobile/desktop/tablet)
1.I would like to come up with progressive data processing. Where initially we show the low resolution video and by the time it buffers it will try to load the entire video which minimizes the buffer time and improves the quality of streaming.( Maps to Challenge 1 and 2)
2. The pixel size varies from device to device and hence the resolution varies. The video streaming on mobile needs some special attention as it has to show a high resolution video in a low resolution phones whose pixels are very less when compared to desktop and tablet.This compression and loading of videos loads the videos faster as the size of the video would be smaller.
We will have to tie up with 3rd party to work on building a compressor which compress the videos and reduce the resolution to fit in the mobile resolution which is still a high resolution for mobiles.(Maps to challenge 1 and 4)
3. Have more data centers for improving the performance of videos.( Maps to challenge 3)
I would trade-off above solutions based on Usage frequency and implementation cost for Pilot 1 patch.
1. Usage frequency-high(As the quality of the video improves the more chances of viewing more videos.)
Implementation cost- Low( should have some algorithm implementation which will not need any additional infrastructure or resources.)
2.Usage frequency-High( Interests user to watch the videos with reduced delay in transmission.)
Implementation Cost- Med( Involves tie ups with third parties or additional resources and time to invest on using compressors.)
3.Usage Frequency- High( faster access to content and videos.)
Implementation cost – High( setting up new data centers is expensive.)
SO i would prioritize solution 1 based on above metrics.
I would like to improve the video experience of FB by working on improving the user engagement and user satisfaction metric that is to implement progressive data loading/processing which results in higher resolution and lower buffer time of videos.It improves the quality of videos which makes users watch more videos and improves stickiness factor.