From Cancellation to Conversion
Rethinking user definitions to create new value proposition
Companies today use different types of tools to help increase conversion and retention and although the healthcare industry has been slow in adopting the digital revolution, users expect an easy and smooth experience when using an health app.
A well established B2C health app with the requirement for registration and deposit in its onboarding stage was struggling to identify the reason behind its high cancellation rate once the free trial period was over.
Like in all industries, conversion lies on understanding the users motivation and need but in this case, the traditional data analysis showed no clear conclusion as to what was the reason behind the cancelation of the subscription.
So how do you find the real reason behind the cancelation? how do you shed light on the real user experience and find new traces of information in a rooted well-structured product funnel?
The funnel had a "black box" - the dry data did not provide a clear indication of the user's experience. By identifying user activity patterns and classifying them, Raven was able to create an overview of all the activity combinations which resulted in cancellation.
Each activity combination served as specific cases that can be interpreted and explain the motivation behind each step of the user.
The interpretation enabled setting new user parameters which better reflect the user's behavior,and therefore allowing the company to offer customized solutions to increase conversion and insure user retention.
Raven's data team established a strategy that explores more unique usage patterns and offers new user behavior interpretation. For that, Raven developed two different Machine Learning and AI Models to help classify the customers which have canceled their subscription: 1) User-text-conversion model - a qualitative hierarchical clustering based on LDA modeling for user text analysis, which classifies the users into groups by their conversation patterns. 2) User activity model - turning user activities into neural nets using GAN( generative adversarial network) and LSTM (Long Short Term Memory networks) methods.
Rave'sn data analysis process led the company to clessify users in two types groups. One group was characterized by a clear intent to use the app for medical purposes while the other used the app for its gamification features. The new user identification parameters served as the foundation for the company strategy to create new customized features in the app. overall, the company's value proposition has amplified and led to an increment of 50% in user cancellation. The company today, using Raven's methodology, can already recognize and predict this type of users and offer them a customized solution.