Skip to content

pathilink/criteo_case

Repository files navigation

Criteo Analytics Case Study 2025

Status Badge

This project is an analytical solution for the Criteo case study, with the aim of generating strategic insights from simulated digital campaign data.

🎯 Objectives

  • To assess the exposure of the Brazilian market to its main advertisers.
  • Analyze the impact of year-end sales on the fashion sector.
  • Estimate Q1 2025 revenue for US advertisers.
  • Provide strategic recommendations based on the data analyzed.
  • Explore additional insights through complementary analysis.

🧾 Data Structure

Column Description
Year Campaign year
Month Campaign month
Country Country of advertiser
Advertiser_ID Unique identifier of the advertiser
Industry Industry of advertiser
Clicks Clicks on ads
Displays Ad impressions
Client_Sales Sales generated for the client
Client_Revenue Revenue generated for the client

🇧🇷 Concentration of Clicks in Brazil (2024)

How many advertisers are responsible for 40% of clicks in Brazil?

  • Strategy: cumulative sum of clicks per advertiser ordered in descending order.
  • Result: 5 advertisers account for 40% of clicks in 2024.
Advertiser_ID Cumulative clicks %
A9 13.82
A24 23.13
A13 30.70
A33 35.71
A5 39.60
A22 43.36
A99 45.14
A75 46.82
A139 48.39
A151 49.95

Cumulative Click Share graph

👗 Seasonal Impact: Fashion in France

User behavior in the fashion industry during the end-of-year period (November and December).

Metric Jan-Oct (average) Nov-Dec (average) Variation (%)
Clicks 15,465.19 17,355.28 +12.22%
Displays 3,787,375.21 3,552,459.70 -6.20%
Client_Sales 320.72 366.55 +14.29%
Client_Revenue 25,476.96 46,567.11 +82.78%

$$ \text{Variation (%)} = \left( \frac{\text{Holiday Avg} - \text{Before Avg}}{\text{Before Avg}} \right) \times 100 $$

Fashion France

🇺🇸 Estimated Revenue Q1 2025 - US

As the data for 2025 is not available, I used the data for 2023 and 2024 for the projection.

  • Revenue Q1 2023: US$ 81,929,089.52
  • Revenue Q1 2024: US$ 298,461,921.57
  • Annual growth: +264.29%
  • Estimate Q1 2025: US$ 1,087,275,827.81

$$ \text{Growth Rate} = \left( \frac{\text{Q1 2024} - \text{Q1 2023}}{\text{Q1 2023}} \right) \times 100 $$

$$ \text{Q1 2025 Est.} = \text{Q1 2024} \times \left(1 + \text{Growth Rate}\right) $$

Revenue US

📈 Advanced Analysis

✔️ Campaign efficiency by Advertiser (Brazil Top 10 by Clicks)

  • CTR (Click-through rate): Clicks / Displays
  • CPC (Cost per click): Client_Revenue / Clicks
  • CPA (Cost per acquisition): Client_Revenue / Client_Sales
Advertiser_ID Clicks Displays Client_Sales Client_Revenue CTR CPC CPA
A9 7720264 1392250876 309903 17,246,832.61 0.01 2.23 55.65
A24 5198795 652614462 393661 12,744,086.39 0.01 2.45 32.37
A13 4227974 975029266 125080 6,914,913.10 0.00 1.64 55.28
A33 2800175 407697994 110680 14,463,447.89 0.01 5.17 130.68
A5 2168307 250716739 168229 4,638,622.80 0.01 2.14 27.57
A22 2101798 408402680 49751 3,451,414.38 0.01 1.64 69.37
A99 993325 122836378 11889 2,552,285.60 0.01 2.57 214.68
A75 939631 176917053 6237 1,666,896.87 0.01 1.77 267.26
A139 875097 123817233 4403 1,491,770.65 0.01 1.70 338.81
A151 872824 139038811 13627 555,114.52 0.01 0.64 40.74
  • Efficient campaigns (⬇️ CPC and CPA)
  • Very expensive campaigns (⬆️ CPC and CPA)

metrics

🌍 Top Revenue by country and industry

Country Industry Industry_Revenue Total_Revenue_by_Country Percent
US AUTOMOBILE / MOTO / BOATING 596,372,055.65 1,427,777,297.12 41.77
FR REAL ESTATE 358,910,782.30 1,119,818,654.50 32.05
JP TRAVEL 207,887,315.78 354,035,987.77 58.72
DE FASHION / LUXURY 134,671,582.57 397,887,676.04 33.85
GB TRAVEL 129,095,404.93 496,625,694.51 25.99
BR FASHION / LUXURY 63,804,384.77 192,178,725.62 33.20

🧠 Conclusions & Recommendations

  • The click base in Brazil is concentrated in a few advertisers → diversification is key.
  • Fashion in France is highly seasonal → prioritize personalized campaigns in November/December.
  • USA has consistent growth in Q1 → reinforce campaigns at seasonal events.
  • Clustering can be used for targeted strategies by advertiser profile.

🛠️ Technologies used

Python Badge

👩🏻‍💻 Author

Linkedin Badge

🔓 License

License: MIT

About

Insights from simulated digital campaign data.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published