Project Overview

This project analyzes 12 months of real web analytics data to understand how users move through the acquisition and conversion funnel, where they drop off, and what drives growth. The analysis covers user acquisition by channel, device-level behavior, mobile analytics, conversion funnel performance, and an A/B test simulation for a checkout optimization intervention.

Problem Statement

Growth teams need to understand not just how many users they are acquiring, but where those users come from, how they behave, and why most of them never convert. The goal of this project was to map the full funnel from session to transaction, identify the highest-leverage drop-off points, and design a data-backed experiment to address the biggest gap.

Duration

The project spanned several days, covering SQL data extraction, exploratory analysis, funnel modeling, statistical testing, and dashboard design.

Data

The analysis uses the publicly available Google Analytics Sample dataset from BigQuery Public Data. It contains session-level records from the Google Merchandise Store covering August 2016 through August 2017, with device, geography, channel, and transaction data across 900K+ sessions.

Approach & Methodology

Data was extracted and flattened from BigQuery using SQL, then analyzed in Python using Pandas, Matplotlib, Seaborn, and SciPy. The methodologies included funnel analysis, channel attribution, device segmentation, mobile analytics, and two-proportion z-test significance testing for A/B test simulation.

The Approach and Process

Exploratory Data Analysis

• Established core growth KPIs: total sessions, unique visitors, conversion rate, bounce rate, average revenue per session, and average session duration
• Analyzed monthly session volume and unique visitor trends to understand growth trajectory and seasonality
• Evaluated month-over-month growth rates to identify acceleration and deceleration patterns

****Direct link to full analysis notebook: GA_User_Growth_Funnel_Analysis.ipynb