Dynamic Segmentation For Over A Million Users.
Objective
Segment and cohort over a million users.
Obstacle
Starting from scratch and building data architecture.
Outcome
Dynamic segments and cohorts created for easier personalisation.
Background
Bloom, the world’s first self-guided therapy app, is on a mission to gain a deeper understanding of their growing user base of over 1 million individuals. With such a large and diverse audience, the team is eager to uncover the key factors that drive user conversion, what motivates users to take the step from exploring the app to becoming paying subscribers? Understanding these triggers is essential for optimising their onboarding process and improving their conversion rates.
Challenges
User Behaviour: Therapy is a deeply personal and complex subject, with no standardised path for how individuals seek or access help. Unlike other digital products, where higher usage often directly correlates with increased subscription levels, Bloom found that this wasn’t necessarily the case for their app. Users might engage heavily with the content but not convert to paying subscribers in the way one might expect, making it challenging to draw straightforward conclusions about what drives behavior.
External Factors: Bloom was blind to the external factors that often influence someone’s need for therapy, life events, stressors, or personal circumstances. Without visibility into these drivers, the team was left to rely on assumptions or incomplete insights. Recognising the limitations of guesswork, Bloom sought firm, data-backed answers to better understand their users and the nuanced factors impacting engagement, conversion, and long-term value.
Solution
One Source Of Truth: We kickstarted their modern analytics pipeline by combining and modelling user behavioural data from Mixpanel and subscription payment data from RevenueCat. This integration provided a unified view of user activity and financial performance, laying the groundwork for actionable insights. From the modelled data, we built an interactive leadership dashboard that automated the tracking of key KPIs across acquisition, retention, engagement, and revenue, offering a clear and comprehensive top-level overview.
Analysis: With this high-level visibility established, we conducted a deep-dive analysis to uncover hidden levers that could drive improvements in subscription uptake and renewal. By pinpointing these opportunities, we enabled the team to make informed, data-driven decisions to enhance user engagement and maximise long-term growth.
The Proof Is In The Numbers...
5
1M+
35
Implementation
Segmentation: We created distinct user segments based on behavior, for example those who used the app regularly to reinforce positive habits. Within these segments, we further defined cohorts based on specific conditions or user needs and categorised them by value. This allowed Bloom to gain deeper insights into user engagement and identify high-value groups, enabling more targeted strategies for content, and product optimisation.
Activity Bucket: We developed an activity score bucketing system that provided a clear understanding of user engagement and helped pinpoint the features and content types most strongly correlated with conversion and lifetime value (LTV) at different stages of the user journey.
Set For The Future: The entire project was completed in just five months and included the setup of a bespoke data pipeline with automated dashboards, the creation of an activity scoring model to evaluate user engagement, and in-depth data analyses.
