Crunchy Bits - Campaign Performance A/B Analysis

A/B Testing & Sales Performance Analysis of Marketing Promotions

Company

Crunchy Bites

Goal

A snack chain plans to add a new item to its menu. However, they are still undecided between three possible marketing campaigns for promoting the new product. In order to determine which promotion has the greatest effect on sales, the new item is introduced at locations in several randomly selected markets. A different promotion is used at each location, and the weekly sales of the new item are recorded for the first four weeks.

Tools used

Programming Language: Python
Data Analysis: Pandas, NumPy
Visualization: Tableau, Matplotlib, Seaborn

Dashboard

This dashboard provides a data-driven analysis of promotional effectiveness across different market sizes, store ages, and time periods. It highlights key insights from A/B testing, comparing Promotion 1 & Promotion 3 against the control (Promotion 2)

Note: The dataset does not include a control set so Promotion 2 will be used as the baseline to determine campaign performance due to promotion's low but stable outcome.

Key Analysis and Takeaways

Promotion Performance Analysis

Promotion 1

Consistently performed the best, achieving an average sales lift of +22.75% over Promotion 2 (control).

Promotion 3

Showed positive lift (+16.98% over Promotion 2) but was less effective than Promotion 1.

Promotion 2 (control) Consistently underperformed, making it the weakest promotional strategy.

Market ID Performance

MarketID 1, 10, and 7

Had the highest sales lift from Promotion 1, indicating strong customer engagement.

MarketID 9 and 4

Showed the lowest improvement, suggesting promotions were less effective in those locations.

MarketID 2

Lacked Promotion 2 data, making it impossible to measure relative lift.

Final Business Recommendations

1. Scale Promotion 1 Across High-Impact Markets

2. Optimize Promotion 3 for Long-Term Engagement

3. Improve Targeting in Low-Performing MarketIDs

4. Conduct Cost-Effectiveness & Profitability Analysis

Market Segmentation Insights

Large Markets

Experienced the biggest gains from promotions, with Promotion 3 slightly outperforming Promotion 1 in some cases.

Medium Markets

Responded best to Promotion 1, with a +21.89% sales lift, while Promotion 3 had a lower impact (+16.26%).

Small Markets

Showed similar results for both promotions, with Promotion 1 (+18.40%) and Promotion 3 (+17.12%) performing nearly the same.

Sales Trends Over Time

Promotion 1

Gained momentum over time, peaking at a +26.30% sales lift in Week 4.

Promotion 3

Impact fluctuated, with a notable drop in Week 3 before recovering in Week 4.

Week 3

Saw the weakest performance for both promotions, suggesting possible customer fatigue, competitor activity, or seasonality effects.

Visualizations

Queries

Find the location with the highest total sales

SELECT LocationID, SUM(SalesInThousands) AS TotalSales
FROM marketing_campaign
GROUP BY LocationID
ORDER BY TotalSales DESC
LIMIT 1;

------------------------------------------------------------------------------------------------------------------------

"locationid" "totalsales"
209 380.36

Find the week-over-week percentage change in sales for a specific location

SELECT week, SalesInThousands,
LAG(SalesInThousands) OVER (ORDER BY week) AS PreviousWeekSales,
(SalesInThousands - LAG(SalesInThousands) OVER (ORDER BY week)) /
LAG(SalesInThousands) OVER (ORDER BY week) * 100 AS WeekOverWeekChange
FROM marketing_campaign
WHERE LocationID = 1;

------------------------------------------------------------------------------------------------------------------------

"week" "salesinthousands" "previousweeksales" "weekoverweekchange"
1 33.73
2 35.67 33.73 5.751556477912852
3 29.03 35.67 -18.615082702551163
4 39.25 29.03 35.20496038580778

Identify the most effective promotion type by calculating the average sales per promotion

SELECT Promotion, ROUND(AVG(SalesInThousands)::NUMERIC, 2) AS AvgSales
FROM marketing_campaign
GROUP BY Promotion
ORDER BY AvgSales DESC;

------------------------------------------------------------------------------------------------------------------------

"promotion" "avgsales"
1 58.10
3 55.36
2 47.33

Compare sales trends between small, medium, and large market sizes over time

SELECT week, MarketSize, ROUND(SUM(SalesInThousands)::NUMERIC, 2) AS WeeklySales
FROM marketing_campaign
GROUP BY week, MarketSize
ORDER BY week, MarketSize;

------------------------------------------------------------------------------------------------------------------------

"week" "marketsize" "weeklysales"
1 "Large" 2973.32
1 "Medium" 3552.31
1 "Small" 843.68
2 "Large" 2883.58
2 "Medium" 3550.25
2 "Small" 880.13
3 "Large" 2967.19
3 "Medium" 3528.60
3 "Small" 830.23
4 "Large" 2955.52
4 "Medium" 3444.15
4 "Small" 890.52