Boosting engagement by 25% using Recommendation & Gamification system
Miko is a brand of educational robots designed to engage children through interactive conversations, storytelling, and educational games.
​This case study focuses on the Platform OS team's efforts to enhance content engagement on the Miko 3 robot
Team
Sharang (Product Design Lead)
Aditya (Senior Product Manager)
Sanya (Product Designer)
My Role
User Research, Design strategy, Concept Development, Testing & Validation
Problem Statement
The Miko 3 platform features 20+ game apps, 180+ voice graphs, and over 15 third-party brands (with each 20-30 brands) along with 100+ personality-driven HRI content.
The platform was experiencing suboptimal user engagement with a significant gap between the potential user base (MAU) and consistent daily engagement (DAU).
Furthermore, an average of 3.45 sessions per week indicates that users are not fully utilizing the content implying potential issues with content discovery, user experience or content relevance.
User Research
We conducted online interviews with 60 beta users in the U.S., evenly distributed across three age groups: 4-6, 6-8, and 8-10 years.
These sessions provided insights into how children interact with our product across different developmental stages.
Additionally, we conducted in-person sessions with 10 children in Mumbai and Bangalore, allowing us to observe real-world interactions and gather qualitative feedback in diverse cultural settings
Apart from that deriving insights from Customer support tickets and Amazon reviews relly helped a lot
Methods used
1. Observational Study with children
2. Post purchase JTBD interview with Parents
3. Mix Panel Data
4. Customer Support Tickets
5. Amazon Reviews​​​
​​​
Insights
Difficulty in navigating content across genres
Children face uncertainty about what to explore next leading to unstructured learning experiences
How might we empower kids to explore and engage with content in a way that is intuitive, enjoyable and relevant?
Recommendation System
Despite optimizing Miko’s navigation with best practices, building a mental model for the vast and diverse content could only be best addressed through a recommendation system
A personalized recommendation system was designed to suggest stories, music stations , voice skills, and games. The AI system refines its recommendations using usage analytics. It operates on a hybrid model combining content-based suggestions tailored to the user's journey with collaborative filtering based on similar users' explorations.
There were three key touch points for recommendations
1. Post each App session (music, stories, games)
2. Context specific recommendation
3. Proactive system-initiated suggestions
Gamification System
To enhance engagement, a structured learning experience was introduced through the Daily Missions module.
This module encourages children to explore different features and functionalities by offering rewards for completing tasks, using specific commands or reaching new milestones.
By doing so it helps kids become more familiar with the device’s capabilities and enriches their overall experience
Results
Over the past three months, we executed a phased rollout of two new modules. We first launched our Recommendation system and, after two months, introduced a Gamification system to further incentivize user interaction.
Testing and Measurement Process:
We segmented users into cohorts based on acquisition dates and behavioral patterns, allowing us to perform rigorous A/B testing.
This cohort analysis helped us isolate the effects of each module and understand how different user groups responded over time.
Results:
-
Recommendation System:
-
DAU increased by ~1,200 users
-
Average sessions per user increased by ~0.26
-
Average session length increased by ~1.6 minutes
-
-
Gamification System (launched after two months):
-
DAU increased by ~800 users
-
Average sessions per user increased by ~0.3
-
Average session length increased by ~1.8 minutes
-
While these results are promising, this work is still a work in progress.
