Reducing Churn Rate for New Users
Product & Team: Messenger Kids, Growth
My Role: Product Designer
About this Project
The following case study is a project that I completed as part of the Messenger Kids Growth team in an effort to reduce the churn rate for new Messenger Kids users. I was the sole Product Designer for this project and was involved during the entire product development cycle, from ideation to launch.
Design Process
Problem
Parents aren’t adding contacts for their kids after onboarding. This is a problem for the Growth team because we want to retain new users and connections is one of the biggest factors controlling retention. Kids who aren’t connected to their friends and family will use the app less and are likely to switch to another app or service where they can reach all of their close contacts.
Initial User Research
Before diving into the design, as a first step, the team and I wanted to understand why parents stop adding contacts for their kids after onboarding. To do so, I ran two user studies: 1) a survey with employees who use Messenger Kids and 2) interviews with external Messenger Kids users.
Findings
Parents would like their kids to be connected with more of their friends, however, they found connecting kids to be challenging because they don't know where to do so or who is available.
Existing Flow for Parents to Add Contacts
After onboarding, friending for parents is only located in the Parental Controls Dashboard, which is buried within the Facebook Menu and difficult to find.
Solution
Based on our user research findings, in order to solve this problem we need to make friending more discoverable for parents so that it's easier for them to add friends for their kids and know who is available on Messenger Kids.
Design Question
How can we increase the discoverability of friending and encourage parents to add contacts for their kids after onboarding?
Goal
Determine whether or not increasing the discoverability of friending for parents will lead to increased friend requests sent after onboarding, more contacts for kids, and ultimately lead to higher retention.
Performance Metrics
Number of friend requests sent, number of connections made, monthly active users.
Brainstorming Solutions
I lead a brainstorming session with the growth team members to source and discuss ideas to for how we can increase friending discoverability..As a team, we identified many different potential levers to explore that would solve this problem.
Choosing a Lever to Explore
I wanted to identify the highest impact, lowest effort option in order to validate the solution and efficiently use resources. After evaluating all of the options, I chose to explore the Parent's Facebook Feed as the focus area for this project.
Updated Design Question
How might we leverage the Parent's Facebook Feed to increase the discoverability of friending and encourage parents to add contacts for their kids?
Design Considerations
Before starting my explorations I met with key stakeholders to identify any design constraints and principles that I should be aware of...
I met with the relevant Facebook and Messenger Kids product teams to discuss relevant product standards and guidelines since this project involves both products
I met with Legal, Privacy, and Safety XFN to learn about any relevant regulations and risk areas since this project involves sharing kid information
Design Constraints
Legal requirements - can’t show kid profile picture in feed, only kid name allowed
Design system compliance - compliant with Facebook and Messenger Kids design systems
Time and resource limitations - fast turnaround required, so utilize and / or modify existing Facebook feed elements as much as possible
Design Principles
Maintain parent trust and kid safety - protect sensitive kid information and ensure that parents are comfortable with this info in Facebook feed
Ensure parent comprehension - parents need to understand that the promotion they are seeing is for adding contacts for their kids on Messenger Kids and not for themselves or other Facebook users
Maximize performance - connections are an essential element to the value provided to users and are key to Messenger Kids’ success
Optimal Upsell Type & Format
There are numerous types of Facebook Feed upsells (i.e. Full Page Popover, Footer, etc.) and for each upsell type there are many different variants. Selecting the appropriate upsell type depends on our Team's goal (adding contacts) and the the specific variant depends on our content needs (image, text, etc.). With our Team goal and content needs in mind, we identified the 'Simple Feed Unit' and 'H-Scroll Feed Unit' as the best options for our upsell.
Initial Explorations
After deciding on the optimal upsell type and format for the feed unit, I started putting together explorations and landed on three design directions that the team could follow:
H-scroll feed unit focused on Kids
H-scroll feed unit focused on Parents
Simple feed unit focused on connections
H-Scroll Feed Units significantly outperform Simple Feed Units, so we will move forward with the H-Scroll option if our research finds that Parents are ok with seeing kid connections in their Facebook Feed.
Choosing a Design Direction
We had external users evaluate and compare various upsell prototypes to gauge comfort, comprehension, and preference and provide us with directional insights for our final design. These were our findings:
Parents understood that this upsell was to add contacts for their kid on Messenger Kids
Parents are comfortable with seeing potential kid connections in their Facebook Feed and viewed this information as helpful
Unit focused on kid preferred; kid name with account managed by parent performing the best
The avatar performed the best, as some parents felt uncomfortable with seeing kids' profile pictures or just the kid name
Based on these findings, the direction we chose to move forward with is the H-Scroll Feed Unit Focused on Kids using an Avatar to represent the user.
Developing the High Fidelity Prototype
After aligning on the design direction to pursue with product and XFN partners, it was now time to start designing the high fidelity prototype for the H-Scroll Feed Unit. During this process I used the Constraints, Principles, and User Needs identified in the Understand and Ideate phases to guide my design decisions.
Feed Unit Container
The container needs to be FDS compliant and meet our team’s content needs:
Space to explain purpose of feed unit
Space for privacy / visibility info
Space for Messenger Kids branding
Compatible with H-scroll cards
Button leading users to friending page in Parent Dashboard
H-Scroll Card
The card needs to be FDS compliant, optimized for performance, and meet all of the team’s content needs:
Maximizes # of cards in single screen
Space for small profile picture
Space for kid name
Space for parent name
Button for adding friend
Card is dismissible
Profile Picture
Research showed that parents are sensitive about profile pictures, so we wanted to ensure that our designs addressed those concerns:
Utilizes existing, approved artwork
Establishes brand identity
No gender or bias with outfit / action
Differentiates contact suggestions
Optimized for small, round picture
Proposed Design Ready for User Testing
Based on the requirements gathered in the previous design stages and after getting the design approved by all the necessary product partners this is the Facebook Feed Unit design that I proposed, ready to be tested with users...
Launch-Ready User Testing
Since this feature is highly visible and involves sharing kids’ personal information, prior to launch we conducted another round of user research to ensure parents felt comfortable with the final design.
Findings
Overall, the reception of the design was overwhelmingly positive, however, there were a small number of users who felt uncomfortable because they misunderstood the visibility or just didn’t want to see this information in their feed.
Improvements
To address these concerns, the following improvements to the design were made:
Make privacy info more visible
Allow users to hide unit
Final Design & Launch
After implementing the design improvements and finalizing the designs, I worked with Engineering to build out this feature and conduct QA to ensure the build matched the mocks. Below is the final product that we launched!
Results
We conducted a 3-day experiment and the results confirmed our initial hypothesis. Increased discoverability of friending successfully drives contact requests sent by parents, connections made, and ultimately monthly active users.
A/B Test for Privacy Notice
The privacy notice addresses parents’ concern over seeing kid information in feed and helps them feel more comfortable sending out the connection requests.
Moving Forward
Build off of the success of this initial version of the feed unit. Address the areas for improvement identified above to optimize the performance.
Click-through rate (~6%) is at an acceptable level when compared to similar upsells, however, it are still lower than expected, highlighting opportunities to improve contextuality, candidate list, and candidate ranking.
The overall conversion rate of impressions to monthly active users for this experiment is low. However, breaking down the data by user segments, we found that the impact of the upsell on needy users (users with < 5 friends) was more than 2x for monthly active users, highlighting opportunities to improve targeting.
Once the feed unit is optimized, the plan is to convert this to an Evergreen Upsell (no expiration) due to its impact and importance to the growth and success of Messenger Kids.
What have you learned from this project?
This project involved 3 product teams and 2 design systems and required a lot of coordination to create a design that met all of the different team and design system criteria. What I learned through this process is that collaborating early and often can save you time and also enable you to ship the best product possible. If I would have reached out to product partners earlier, I wouldn't have wasted time on explorations that weren't compliant from the start. Additionally, if I hadn't reached out to growth designers on other teams and gotten additional eyes on my designs, I would have missed out on a lot of learnings from previous experiments and the final design would not have been as effective and led to less impact.
Moving forward I will make sure to align with all relevant partners and research related work before diving into the designs. Effective collaboration is a vital part of being a good designer and I will make this holistic approach a top priority in my design process moving forward.