Google Ads Performance Analysis

Overview

This project analyzes Google Ads campaign performance data covering the period from August to September 2013. The objective is to evaluate campaign effectiveness, identify key drivers of profitability, and provide actionable recommendations for budget allocation and optimization.

The analysis was conducted in R and combines exploratory data analysis, data visualization, KPI monitoring, and panel-data econometric modeling.

Objectives

• Evaluate campaign-level performance.
• Compare return on investment (ROI) across campaigns.
• Analyze conversion rate and click-through rate dynamics.
• Identify high-performing and underperforming ad groups.
• Investigate the main factors influencing profitability.
• Generate practical recommendations for campaign optimization.

Dataset

The dataset contains Google Ads performance metrics at campaign and ad-group level, including:

• Clicks
• Impressions
• Click-Through Rate (CTR)
• Average Cost Per Click (CPC)
• Cost
• Conversions
• Conversion Rate (CR)
• View-Through Conversions
• Conversion Value

An additional ROI metric was calculated as:

ROI = ((Conversion Value - Cost) / Cost) × 100

Methodology

Exploratory Data Analysis

• Campaign spend distribution analysis
• Weekly ROI trends
• Conversion rate tracking
• CTR vs Average CPC scatter analysis
• Ad-group performance benchmarking

Econometric Analysis

Random-effects panel models were estimated to examine the relationship between ROI and:

• Click volume
• Impressions
• CTR
• Conversion Rate
• Time trends
• Campaign-specific effects

Model selection was supported by panel-data diagnostics and Hausman testing.

Key Findings

• Campaign 15 consumed the majority of advertising spend but delivered weaker efficiency.
• Campaigns 1 and 3 consistently outperformed other campaigns in terms of ROI.
• Conversion Rate emerged as the strongest predictor of ROI.
• CTR alone was not a significant predictor of profitability.
• Performance deterioration over time suggested possible ad fatigue and audience saturation.
• Several ad groups demonstrated strong scaling potential due to high CTR and low CPC.

Tools & Libraries

• readxl
• dplyr
• janitor
• lubridate
• plm
• ggplot2
• scales
• lmtest
• sandwich

Repository Structure

├── analysis.R
├── AdWords_Report.pdf
├── README.md
└── data/

Results

The project provides:

• Campaign performance evaluation
• ROI and conversion analysis
• Ad-group benchmarking
• Marketing insights and optimization recommendations
• Panel-data econometric modeling results

Author

Leonidas Zangelidis