Questions:

• Is “Facebook Search” referring to the overall search bar at the top where users can search across multiple facebook products (people, events, groups, marketplace, etc.)
  • Interview (I): Assume it is for the overall search experience
• What’s the current goal for Facebook as a company? How does Facebook search fit into the current company goal?
  • I: Your call
• What’s the goal for this improvement for Facebook Search? Are we aiming to drive engagement overall? engagement within a particular product? increase discovery-> adoption of new product (like adding a new group)?
  • I: Your call
• What platform are we focusing on?
  • I: Your call

Assumptions:

• “Facebook Search” feature that we are improving is for the overall search experience on FB
• This exercise will focus on…
  • the mobile platform since mobile usage/adoption is on the rise and most facebook users will be accessingFB on their mobile platform (vs. desktop)
• Facebook Goal: Drive meaningful connections & engagements within your community
• FB Search Goal: Help users find the meaningful engagements that they desired in the FB community

Approach:

1. Define user segment that we want to hone in on
2. Understand use cases and pain points for the specific user segment
3. Identify potential ideas to solve the identified pain points
4. Prioritize and select an idea for my product vision
5. List and prioritize features
6. Trade-off / risks

User Segments by activity frequency

• Daily FB users
• Weekly FB users
• Monthly FB users
• In-frequent users (5-10 times a year)
• Potential Churner (less than 5 a year)

Normally, I would want to look at FB data to see what are the sizing for each bucket and their engagements on the platform. Given that we don’t have this data today and our goal is to improve search so we can help increase engagements in the FB platform, I would not pick daily and weekly users since they are probably already engaging with the platform quite frequently (i.e # of posting/comments/likes per month is high for these two groups). I would pick Monthly FB users since…

1. their engagement is probably low (compare to daily/weekly segment), which creates a large opportunity for us improve their engagement
2. their monthly activity means that we have more opportunity to engage and influence their activity on facebook, vs. in-frequent/potential churners

search use case and pain points

• hard time finding only latest posts/activities from my friends to get latest updates
• difficult to find a person that I am looking for to add
• want to find a group to join, but don’t know what to search for / too many results
• want to find an event worth going to but dont know which events are my friends going to
• difficult finding the item that I want to buy on marketplace
• hard to find a photo or video that my friends are in

I’d like to focus on improving the experience of “finding an event worth going to but dont know which events are my friends going to”. I want to focus on this for the following reasons:

1. by helping the user find an event that they want to go with their friends, it aligns well with FB and our feature goal by helping users to find meaningful engagement on FB
2. this type of search activities are something that can happen more frequently than the other search use cases (e.g. it is more likely that I want to find things/events to go to with my friends every week vs. trying to find a group to join or finding a friend to add). If I can make this experience delightful and easy, it can turn into a habit where users can come back weekly to find events to go to with their friends
3. lastly, events change all the time, so there is a bit of a fomo effect that can drive users to come back to use this new search feature; otherwise, they will miss out

Product Visions:

Find the best events near you that you and your friends will love to go to

Feature Ideas:

4. Event suggestion based on your and your friend’s activity on facebook
5. Event suggestion based on what your friends and friends of friends (2nd degree) are going to
6. Hot / popular Events near you (based on fb rsvp)

The feature I would build to improve the search experience for events would be #1. By basing the recommendation based on your activity and friend’s activity, we can open up the # of events that can be recommended (vs. #2 may be limited if your friends aren’t going to an event). Additionally, #1 will enable our users to be seen as more of a “recommender” as they can now suggest events to their friends knowing that the event should be relevant to your friends. If it goes well, this can create a positive psychological behavior that can sustain this habit (e.g. I want to keep being the “recommender”).

From a UX standpoint, I imagine the user can click on the top search bar and it will take them to a search experience where the top will show a list of events in a horizontal manner. For each recommended event, there should be an obvious CTA to share this event with your friends, as well as some reasons why this event was recommended (e.g. related to your groups, related to an event that your friends just went to, your friends are already going, etc.)

Risk / Trade-off

• There can be potential user privacy concern as we will need to read your activities and your friend’s activities (like comments, groups, postings, etc.)
• Once a user and their friends click RSVP, we don’t know if the event is actually as good and relevant as expected. so we will need to figure a way to gauge this so we can understand whether our recommendation is actually working
• By taking up a more prominent area in the search UX, this can also have a negative impact on other search activities, thus decreasing engagement of other FB products.

Question: How would you improve FB search? Clarification: Can we assume that we are talking about the general FB app search tool? Interviewer: Yes Clarification: Is there any particular reason that we are required to improve the search? Does the usage of the search decreased? Interviewer: There are no particular indicators. Assumptions: The FB search is the main app search on all platforms. The search is a central feature in discovery of all FB app content. Does this sound correct in terms of the focus of our discussion? So, now that it’s clear what we are going to focus on, I would like to outline the structure for the way I am planning to go about this problem. Structure:

1. Goals and mission
2. Target Users / Segments
3. Pain points
4. Solutions
5. MVP (?)

How does this sound to you? Do you want me to focus on anything else?

1. Goals and mission – FB mission is to give people the power to build community and bring the world closer together. I think that it is clear that all FB tools were built for the communities and there is a lot of content available.
   1. The search is a key factor in today’s usage experience in finding just the right content, community and information to allow better engagement with the product. So by improving the search mechanism we are actually focusing on improving the overall engagement of the FB users, which is one of the main goals of FB noverdays with all the other platforms available.
   2. Product goals = allow a more accurate contextual hit rate between search keywords and the results
2. I am thinking about the different types of users that this product have and they are divided into two main buckets:
   1. Searchers – the ones that are using the search
   2. Developers – the ones that are adjusting the search algorithms to better address the needs of the users queries

Prioritise: I am considering focusing on the developers here, as providing more accurate results will greatly impact the usability of the feature and will align more precisely to the initial goal of the product in hand. Does this make sense?

1. Pain points – now I am going to brainstorm a couple of pain points for this type of user.
   1. Understanding the context of the search / keywords
   2. Ability to improve the algorithm based on accuracy of the hit results
   3. Measuring the engagement with the search results

Prioritise: I am thinking that we should prioritise based on frequency of the problem. I am sure that, if possible, we would like to know on each search, whether the user was satisfied with the results that were provided by the search algorithm. The other two pain points are a bit more tricky to address, hence I will prioritise them lower due to complexity:

• Context of the search will require some more complex feedback from the user and might have some privacy issues involved as well.
• Measuring engagement might be more relevant for a different overall engagement domain and will be more difficult to measure in the context of the search.

Do you want me to add anything else? If not, I will be happy to suggest a couple of solutions for this specific pain point.

1. Solutions
   1. Search feedback mechanism
   2. Search satisfaction polls
   3. Results ranking mechanism

Prioritise: For those solutions I would go with: the level of effect that a certain feature can have and the overall effort to implement this feature. I would also consider the trade of interfering with the overall ease of use of the search. I think that a search feedback mechanism will be the best solution to enable a basic interaction with the user and getting results feedback that can be later on used by the search algorithm to improve personalised results. The other two solution might be also relevant however:

• Search satisfaction polls might be more intrusive for the customer and might be out of context. People will not remember exactly what they didn’t like and this might create more frustration.
• Results ranking mechanism might be a very efficient way to understand whether the search algorithm is hitting the results correctly, however this will require a lot of interaction with the user. One of the possible solutions might be to have the feedback mechanism addressing, from time to time, not the overall results but rather a specific part of the results, like the videos for example.

To summarise we can create a basic, familiar search response mechanism that will be a visual tap of smiles or any other visual elements. This can be similar to feedback from Google maps navigation results, Stars as in app ranking and other visual elements like Likes. Those can be A/B tested later on.

larification:

What are we aiming by improving the Search Feature?

We are want to make the user experience more memorable and comfortable.

Which medium are we targeting?

We are going to target the Website version.

What exactly is the Search Feature in FB used for?

It is the feature given at the top, which is used by the user to search for friends/ other users in FB by typing their name. It is also used for searching groups, posts, photos/videos, Events, Places, and even the games.

Do we mean the Search functionality in the Marketplace, Live sessions too?

Yes.

With the present clarity, let me jump on to discuss the user Segments that we can target.

Customer Segment:

Age Group: 17-

Age Group: 17-25

Age Group: 26-50

Age Group: 50+

Of all the above customer segments, it is the Age Group between 17 and 25 the most important. This is because those age groups are College going students and new Joiners in workplaces who tend to be Tech-Savvy and Socialize more in general and, thus, use FB pretty extensively.

Improving their user experience will undoubtedly help our growth multiply manifolds.

Pain Points:

!) They tend to use Facebook extensively and later feel that their time is getting wasted too much.

2. Most of the time, the videos listed in Facebook’s Watch section are not related to the user.

Ideas:

For the 1st Problem:

If a user gets something valuable from the time spent in our app, the thought of “Wasting their Time” will be eradicated.

No matter what the demography is, any person would love to earn while spending his time for enjoyment. That too for a college-going student having a way to earn as well as spending time leisurely will be like heaven. Thus, we could introduce the following idea to help the user have a monetized benefit:

• For every successful search, a user performs and buys a product through our search we could offer them a coupon with a validity of 2 hours and one-time use option to have a discount of saying some 5%.
• For the same, we could partner with some growing E-Commerce platforms and charge them for every successful transaction. This way, they’ll get user acquisition done efficiently, and we’ll have our info for our data collection about every user.

User Flow:

• The same way how a person searches for his friend’s account, they can search for a product.
• In the results side, we’ll display the websites of the partnered E-Commerce platforms.
• By clicking the result, the user will get carried to the said e-commerce site and thus get availed a discount.

Thus, the user gets a monetized benefit out of the time spent through Facebook, and we get our data, and the partnered websites get their user acquisition.

Potential Pit Fall:

Though we get to know the interests of our users, by taking them to a third party website, we are distracting the user. This, in turn, will reduce the session length one or the other way.

Success Metric:

We could look at the following metrics to know how welcomed our product is:

• # of Successful conversion as the metric to look upon.
• # of searches per user per week.

For the Second Point:

As of now, Facebook’s Watch Session shows a user n number of videos listed one by one. It also has a search option to the left panel. It would be of immense help for the user if Facebook allows us to search the top liked videos of our friends. In this way, the relevance of the videos showed will improve. The user will also like to get to know the videos that their friends preferred.

Potential Pitfall:

Privacy Concerns from our users

Success Metrics:

• # of searches
• # of reactions (Both likes and comments)

Prioritize:

Considering the two solutions, Though the second one highly aligns with our goal, Privacy concerns will be a considerable drawback. This will make the Users take it in a wrong way rather than enjoying it. While though the first one’s implementation is a bit more complicated, the fact that the users getting coupons will increase our engagement and will also help us in many ways.

Summary:

So, One of the best ways to improve the search feature to benefit the user is to introduce the listing of products that a user can buy from partnered third party e-commerce sites when someone searches.