This project analyzes user activity data from TELEPROMPTER, a company supporting online content creators with smooth video recording tools since 2018. With over 10,000 new users every month and a database exceeding 15 million records, the aim is to identify key factors influencing conversion from free trial to paid subscription — and ultimately, improve business performance.
Goal: Increase the conversion rate from trial_started → converted, turning more trial users into paying customers.
Impact:
- 💰 Higher recurring revenue
- 📈 Increased customer lifetime value (CLV)
- 🔁 Reduced churn rate
- 🔍 Data Cleaning & Feature Engineering
- 📊 Exploratory Data Analysis (EDA)
- 🤖 Logistic Regression Modeling (with custom thresholds)
- 📉 Confusion Matrix Visualization
- 🌍 Conversion Segmentation (by OS, Country, Time)
- 🔁 Churn vs. Reactivation Analysis
- Certain OS versions and countries show significantly higher conversion rates, suggesting technical or behavioral segmentation opportunities.
- Trial users who return after cancelation remain a small segment — highlighting retention challenges.
- Users on non-annual plans show higher churn than returning users, signaling a need for better long-term offers.
- 🐍 Python:
pandas,numpy,matplotlib,seaborn,scikit-learn - 📊 Power BI (for dashboard)
- 🧪 Jupyter Notebook
- 🔄 Git & GitHub for version control
- 🖥️ PowerPoint (for presentation)
- 🚧 Deploy conversion prediction logic into live product funnels
- 🧪 Run targeted A/B tests for top-performing OS segments
- 📬 Collaborate with marketing to build behavior-based email campaigns
- This analysis was initially developed for an internal competition organized by data36.com data science club, where it earned me the Senior Special Prize ranking.