Me: As per my understanding, Timeline the list of posts by a user that are displayed in a chronological order typically on a user's profile. Additionally, all the things that define a user such as their bio, profile and cover photo, all their photos, movies, books and interests are also part of timeline. Is that correct? (Can someone correct me if I'm wrong please?)
Me: Ok, so Timeline is everything that a user cares about and all the things about themselves that they are willing to share.
Is there a known issue with the current Timeline such as a metric dropping or do we know what aspect of Timeline we want to improve?
Interviewer: Its fine right now but we want to improve the experience for users.
Me: Ok. The goal of timeline is to make sure users update their timeline from time to time so that their up to date information and interests will help other users and helps facebook learn more about the user and their interests. So I would say the goal should be to improve effectiveness. (Is this a specific enough goal?). And as rich timeline helps users find their potential friends and learn more about their friends, it directly helps Facebook's mission of bringing people closer.
Next I want to talk about different user personas that use Timeline and pick one of them and report their needs.
Interview: Ok, go ahead
Me: Main user personas that use Timeline are 1) A user who wants to update their own timeline with new photos or interests any changes in their life.
2) A user who wants to look at another user's timeline that they found on marketplace or a group to find out more about their potential client.
3) A user who wants to add a potential friend and wants to find out more.
4) A user looks at a comment on a post or in an event or a page and wants to message him/her.
Out of these users, I would like to pick the (2) user that wants to find out more about a user who is not their friend since its an important user segment as we expand more marketplace and other community focused products.
Do you agree?
Next I will go over their under met requirements
1) They want to learn about the financial ability of the user if they're a potential buyer or renter
2) They want to find identity of the person in case there are no photos posted by the user .
3) They want to know if there are any type of connections so that they can find someone that you know that might know them so that the user can message them and find out more. Eg: User posts on rentals asking for a roommate. Some people comment on the post. Poster wants more tools to decide how to filter the list.
I will pick the 3rd one since thats a fast growing need and I think it will have the highest user benefit.
Next I will go over some solutions for this:
1) Potential connection: Showing tertiary connections, someone that your friend knows and this user Y knows as well and show them the list so the user X can message them and learn more about user Y.Like the rental example.
2) Similarity score: Show them a "similarity score". Taking all the data points the user has shared publicly we can show them how "similar" these users are so that the user gets
I will implement the 2nd solution because this score will influence users who use our products such as marketplace buying and selling, rentals, jobs, groups etc to add more information about themselves so that the similarity scores are more representative of them and help them in potential sales/finding apartments/roommates and getting more responses from groups and job applications. This helps our goal to improve effectiveness of timeline and have fresh profile information. I will measure the success of this feature by creating control and test groups and measuring avg number of profile updates per MAU in each cohort so that we can see if more users are updating it. As a secondary metric, I will measure avg time taken from creation of post to marking it as sold IF timeline was viewed. Hypothesis here is that users who saw timeline with the similarity score generated more messages and closed the deal quicker than the users who saw timeline without the similarity score. This measures the effectiveness of the improved timeline.