Marketing Campaign Data Analysis

Analyzed 174,226 rows of digital marketing campaign data using R — evaluating creative performance, campaign efficiency, CPA, CVR, and time-period trends across 10 ad creatives and 3 campaign types.

---

📌 Project Summary

This was a take-home analytics test analyzing a real marketing dataset. The goal was to answer 7 business questions about ad creative performance, campaign efficiency, and seasonality trends using R.

Key findings:

• 🎨 10 creative versions were run across 5 colors (Black, Grey, Orange, Teal, White) in standard and new variants
• 📈 New creatives generated marginally more clicks (mean 93.63) vs standard (mean 92.99) — Grey**new showed the largest individual improvement
• 💰 Conversion campaigns drove 5,483,600 conversions — nearly 4x more than Brand (1,381,526) or Fan Acquisitions (1,381,300)
• 📊 CPA ranged from ~$18.69 to ~$19.84 across campaign types — sc_BBY_NOLA_branding had the lowest CPA at $18.69
• 📆 CVR spikes every 3rd month (March, June, September, December) — suggesting seasonal patterns worth modeling with ARIMA
• 🔁 Impressions follow a 2-high, 1-low monthly pattern — average impressions dip below 95,000 every 3rd month vs a baseline of 103,782

---

🗂️ Repository Structure

Marketing-Data-Analysis/
│
├── README.md
├── TakeHomeTest_Code_File.R    # Full R analysis code
└── Data_Analysis_Report.pdf   # Written findings and visualizations

---

🛠️ Tools & Skills Used

Tool	Purpose
R	All analysis and visualizations
tidyverse / dplyr	Data manipulation and filtering
ggplot2	Bar charts and trend visualizations
readxl	Importing Excel dataset
lubridate / xts	Time series handling

R concepts demonstrated:

• Data cleaning and column standardization (tidy_names, tolower)
• Filtering and grouping with dplyr (filter, group_by, summarise)
• String pattern matching with grepl
• Custom metric calculations (CPA, CVR)
• Side-by-side ggplot2 visualizations with gridExtra
• Time series trend analysis across monthly dimensions

---

🔍 Analysis Questions Answered

Q1 — How many versions of creative were run?

10 versions across 5 colors in standard and new variants: Black, Black**new, Grey, Grey**new, Orange, Orange**new, Teal, Teal**new, White, White**new.

Q2 — Did new creatives have any effect?

New creatives showed a marginally higher mean for clicks and impressions but no significant overall effect on conversions. Grey**new was the standout — showing meaningful improvement in both clicks and impressions relative to standard Grey.

Q3 — CPA and CVR by campaign type

• CPA = Total Spend / Total Conversions — ranged from $18.69 to $19.84
• CVR = (Conversions / Impressions) × 100 — exported per campaign

Q4 — Which campaign type drove the most conversions?

Campaign Type	Sum of Conversions	Mean Conversions
Conversion	5,483,600	94.37
Brand	1,381,526	23.80
Fan Acquisitions	1,381,300	23.78

Conversion campaigns generated ~70 more conversions on average than Brand or Fan Acquisition campaigns.

Q5 — Are there time period trends?

• Spend: Stable at ~$467/month with negligible variation
• Impressions: Follow a 2-high, 1-low pattern — dip below 95,000 every 3rd month
• Clicks: Stable at ~92–93/month with no seasonal signal
• Conversions: Marginally lower in March and June, correlated with impression dips
• CVR: Spikes above 0.0510% every 3rd month (March, June, September, December) — suggests quarterly seasonality. ARIMA modelling recommended for deeper forecasting.

Q6 — Creative colors run in February?

8 creatives: Black, Black**new, Orange, Orange**new, Teal, Teal**new, White, White**new

Q7 — What questions would you ask to project next month's CPA?

1. What is the spend budget for next month?
2. Will the client maintain the same conversion-heavy campaign split?
3. What month is it — slow month or fast month based on seasonal trends?
4. Is it a peak demand period for this product?
5. Is a competitor entering the market?
6. What is current consumer demand like — stagnant, growing, or expanding?
7. What is general market sentiment around this product?

---

💼 Key Takeaways

• Conversion campaigns are the clear winner — 4x more conversions than other types at similar CPA
• Grey**new is the best performing creative update — significant improvement over its standard version
• Quarterly seasonality in CVR is a real pattern worth building a forecasting model around
• New vs standard creatives showed minimal aggregate difference — individual creative color matters more than the new/standard split

---

*This project was completed as a take-home data analysis test. Dataset also included.