How Are You Being Evaluated
This product execution interview question tests whether you can decompose a problem, think in a structured way and articulate your solutions.
- Describe how Facebook currently does search and how it monetizes it.
- Analyze how Facebook compares to Google and mention relative strengths and weaknesses.
- Describe use case scenarios that can take advantage of Facebook’s strengths.
- Prioritize use case scenarios to implement.
- Ideate solutions for these scenarios.
- Think of ways to monetize these solutions.
- Wrap up
INTERVIEWEE: Facebook currently has some search capabilities, and my impression is that they are monetizing it through search ads. Is this correct?
INTERVIEWEE: So, we are looking for ways to increase revenue generated through search ads?
INTERVIEWER: Well, search ads are one way to generate revenue. But, I would like to hear from you other ideas beyond search ads.
INTERVIEWEE: Okay, understood. So I would like to start by doing an overall analysis of search in Facebook today vs. the competition, then describe user search scenarios, prioritize them, and brainstorm ideas on how to monetize them in ways beyond search ads.
INTERVIEWEE: Since I have not done many searches in Facebook on the past, I would like to run a few now to understand better how it works. Could I take a few minutes and do that now?
INTERVIEWER: Of course, please go ahead.
(The interviewee runs two searches, one to search for ‘running shoes for women’ and the other to search for ‘auto maintenance services.’)
Okay, I just did two sample searches: one for finding women’s running shoes and another for auto maintenance services, which I currently need. From what I observed, the results are not actionable. The search organizes the results into three sections: Marketplace, Photos, and Links. The Marketplace results are for used shoes, so it is not relevant since I want to buy new shoes. The Photos section lists posts from people with photos of shoes, but that is not what I am looking for. These results are not useful for several reasons:
- the search engine is not taking context into account, for example, it should deduce that I am looking for new shoes, not used ones, and provide useful information like price and recommendations;
- the results are not exhaustive; the user cannot rely on the results to make an informed decision,
- the results are not personalized, at a minimum, the results should take into account my location,
- and there is no prioritization given to the results, for example, listing by lowest price and the highest number of recommendations is something users expect today.
As a comparison, if you run the shoe query in Google the results are shown within a photo carousel with brand, store, price and recommendations information. You see multiple options to make comparisons that helps you narrow down which shoes to buy. In the case of searching for auto maintenance services, Google displays the closest auto repair shops around me. These results are useful. They enable the user to narrow options to choose from, a very actionable result.
Now, before talking about search use case scenarios, I need to point out that unless the search algorithm is improved, it is unlikely users will see Facebook as a search destination. The Facebook search engine needs to prove its value to users first.
Moreover, another hurdle is the fact that Google is the de-facto top-of-mind destination for search. Changing people’s mindset about Facebook as a search destination would require changing people’s habit of always using Google.
I think Facebook should leverage its advantage in knowing personal and granular behavioral data about its users to provide better search results. For example, users favor recommendations from friends when searching for products, services or entertainment. And they like detail experience descriptions about a product or service, something Facebook users already shared with posts. Google could try to do the same using Google+, but as we know Google+ has not taken off.
With this in mind, I would like to brainstorm about different user search scenarios and prioritize the ones that could leverage Facebook’s social network data.
Could I take a minute to think about possible use case scenarios?
(The interviewee draws the following word association graph to think of different use case scenarios. She marks with + those that will be more fruitful.)
Okay, here are a few use cases that I think would be fruitful to consider.
- People traveling to a new city or country for tourism usually need help in finding places, events or restaurants in the area. Recommendations from friends or a large pool of people, with photos, and detail entries of their experience, would be useful.
- A second scenario is helping users find good cooking recipes recommended by the user’s friends or other Facebook users.
- A third scenario is looking for a good personal service provider, like a stylist or massage therapist. Having recommendations from friends or others about the person that directly provides these services, rather than recommendations on the business they work, are better.
- A fourth scenario is finding possible roommates profiles that meet a user’s criteria.
I would prioritize these scenarios based on which ones leverage Facebook’s proprietary social data. Of these four, I think recommendations while traveling and finding a personal service provider meet these criteria. But, I prefer the traveling scenario, since it is more suited for a mobile search, and mobile access to Facebook is greater than 50%.
In the traveling scenario, Facebook could provide search results to restaurants, events, and places to visit based on friends’ recommendations. Perhaps recommendations from previous visits to the same city or maybe recommendations of what to do in that city. The Facebook search engine would need to know which friends had visited the same place and whether they posted about them. If there are not too many postss from friends, the search algorithm could aggregate data from all Facebook users that visited the same location. One concrete implementation for the mobile app could be to add a new entry under the Explore section and call it Travel Companion. After tapping the Travel Companion option, a map appears with pins close to the user’s current location, highlighting places such as restaurants, museums, bars, hotels, and theaters. The map could also show photo bubbles of friends that have been at different locations in the same city. A filter control could be added to let the user choose which types of places to show on the map. The map could be displayed in two modes: full screen or half screen, with the lower half presenting more details about the pinned locations. Tapping on a theater or museum list entry could take the user to a page that shows the current show times and interface to order tickets. A similar idea can be used for restaurant reservations.
Now, regarding how to monetize this feature, I think a transaction, search ad, and call-to-action model can be used.
Here are some examples:
- When a user buys a ticket from the theater or museum, a transaction fee can be charged from those establishments.
- Charge a search ad fee if the user clicks on a sponsored ad that appears as part of the search results.
- Charge a call-to-action click fee when the user clicks on a CTA button in a search result.
For this type of search feature to be successful, there needs to be a lot of data available from Facebook users. Finding posts with location information is not a problem since users are in the habit of tagging posts with their location. But, getting users to enter recommendations about places they visited could be harder especially once the experience is over. One solution could be to send reminder notifications to users that bought tickets or eat at a restaurant to provide feedback. The feedback can take the form of a simple simple star rating system with additional space for comments.
Would you like me to elaborate on this?
INTERVIEWER: I think we are running out of time so how about just summarizing your analysis.
INTERVIEWEE: Sure. Today, Facebook is not top-of-mind as a place for doing searches, Google is. The search results that the Facebook search engine provides are not very useful. To change users’ mindset about Facebook as a search destination, Facebook needs to improve its search engine. This is a precondition to monetizing search in Facebook. One way of improving search is by leveraging Facebook’s knowledge of personal and detail behavioral data about their users. The types of search queries that could benefit from this knowledge are recommendation-based searches. I brainstormed different scenarios and proposed a solution for a typical scenario: a traveler going to a new place and looking for recommendations for places to visit, restaurants, etc. The solution consists of displaying a map with photo bubbles of friends and pins of locations they visited close to where the user is. If the number of friends is too small, the results can be complemented with other Facebook users’ recommendations. Finally, I suggested three monetization models that can be applied to this feature: transaction, search ad and call-to-action fees. I think this solution can be used as a proof of concept that Facebook can provide better search results than Google, when social data is key to finding targeted results.