Embedded Files
Retail Analytics: Customer Segmentation and Store Trial Performance
Retail Analytics: Customer Segmentation and Store Trial Performance
Quantium Simulation | Case Study in Grocery Retail Strategy
Quantium Simulation | Case Study in Grocery Retail Strategy
Project Overview
Project Overview
This project was completed as part of the Quantium Virtual Experience Program, simulating the role of a retail data analyst. Over two tasks, I explored customer purchasing behavior and assessed the impact of a new store layout trial in the chips category, with the goal to deliver clear, data-backed insights to support strategic decisions for the client’s upcoming category review.
Tools used: Python | Pandas | Matplotlib | Seaborn | SciPy
Focus areas: Data cleaning | Customer segmentation |Control store matching | Uplift testing | Strategic analysis | Visualization and Reporting
Business Context
Business Context
Quantium is a global leader in data science and artificial intelligence, known for helping businesses harness data to drive strategic decisions. This project was completed as part of the Quantium Virtual Experience Program, which simulates the kind of challenges faced by Quantium’s retail analytics team.
The simulation focused on the chips category within a major retail client, where I was tasked with analyzing customer purchasing behavior and evaluating the impact of a new store layout trial. The work reflects real consulting scenarios — from segmenting customers and benchmarking performance to delivering insights that inform category strategy and executive decision-making.
Executive Summary
Executive Summary
This project explored customer purchasing behavior and evaluated a store layout trial to inform chips category strategy. The work was divided into two parts: segmentation analysis and trial impact assessment.
Part 1 – Customer Segmentation & Purchasing Behavior
Part 1 – Customer Segmentation & Purchasing Behavior
Objective: Identify key customer segments and purchasing patterns to guide category decisions.
Approach:
Cleaned and validated transactional data (outliers, missing values, formatting)
Engineered features such as pack size and brand name
Segmented customers using LIFESTAGE and PREMIUM_CUSTOMER attributes
Analyzed total spend, frequency, and product preferences across segments
Key Findings:
Mainstream Older Families and Retirees drive the highest sales and volume
Young Singles/Couples are the largest group but spend less per customer
Mainstream customers pay the highest unit prices; Premium customers buy less frequently
Strategic Recommendations – Task 1:
Prioritize Mainstream Older Families and Retirees in promotions and product placement
Off-locate premium-priced products near high-traffic areas frequented by Mainstream Young Singles/Couples
Use segmentation insights to guide assortment, pricing, and store-level targeting
Part 2 – Trial Store Impact & Control Benchmarking
Part 2 – Trial Store Impact & Control Benchmarking
Objective: Evaluate the effectiveness of new store layouts using uplift testing.
Approach:
Selected control stores using correlation and magnitude matching
Built scalable functions to compare trial vs control performance
Assessed monthly trends, uplift significance, and drivers of change (volume vs frequency)
Key Findings:
Store 88 showed sustained uplift in sales and customer volume
Store 77 attracted more traffic but saw lower spend per customer
Store 86 had a short-lived spike, with performance returning to baseline
Strategic Recommendations – Task 2:
Roll out the new layout to stores with profiles similar to Store 88
Refine layout strategy for Stores 77 and 86 to boost basket size and sustain engagement
Monitor post-trial performance to validate long-term impact and guide broader implementation
Supporting Visuals & Insights Task 1
Sales peaked in March 2019, following a sharp dip in February, suggesting a possible seasonal or promotional rebound.
The rolling average line smooths short-term fluctuations, revealing a stable upward trend across the year.
These insights underscore the importance of timing promotions strategically to align with natural demand cycles and recovery periods.
Mainstream customers dominate sales across nearly all lifestages, especially among Older Families and Retirees, reinforcing their role as high-value segments for the chips category.
The Customer Lifestage Average pie chart shows that Young Singles/Couples make up the largest customer group (21.5%), yet contribute less to total sales — revealing a gap between population size and purchasing power.
The Sales by Customer Segment pie chart confirms that Mainstream customers drive the highest share of total sales (41.8%), followed by Premium and Budget tiers.
Mainstream Young Singles/Couples form the largest customer group, followed by Mainstream Older Families—indicating strong population density in these segments.
Despite their size, Young Singles/Couples contribute less to total sales and quantity sold (as shown in earlier charts), suggesting lower engagement or spend per customer.
Premium customers are consistently fewer across all lifestages, reinforcing their niche status.
Mainstream customers consistently dominate across most lifestages, contributing between 29% and 58% of total volume.
Budget customers show strong presence in Older Families, New Families, and Young Families, often exceeding 40% of the segment’s volume.
Premium customers, while present, represent a smaller share—typically under 30%—highlighting their niche role in volume contribution. The percentage labels make it clear that Mainstream customers are the backbone of chip consumption, while Budget customers play a key role in family-oriented segments.
The heatmap reveals that Mainstream Young Singles/Couples pay the highest average unit price (≈ €4.07), significantly more than their Budget and Premium counterparts. This suggests that Mainstream midage and young singles/couples are more willing to pay a premium per packet, likely driven by impulse buying or brand preference.
Interestingly, Premium customers tend to pay slightly less per unit, which aligns with their lower purchase volume and frequency. This behavior may reflect more selective or occasion-based purchasing.
These insights highlight an opportunity to off-locate premium-priced products near high-traffic areas frequented by Mainstream young singles/couples, leveraging their higher willingness to pay and maximizing commercial impact.
Supporting Visuals & Insights Task 2
The heatmap highlights the relative performance of each trial store across three core KPIs.
Store 88 shows the strongest uplift in total sales and customer volume, with a modest gain in transactions per customer.
Store 86 demonstrates the highest uplift in customer volume and a positive shift in transactions per customer, though total sales remained flat.
Store 77 shows uplift in sales and customer volume, but a slight decline in transactions per customer, suggesting limited behavioral change.
These results reinforce Store 88 as the most promising candidate for rollout, while Stores 86 and 77 require further evaluation.
Store 88 showed a clear and sustained uplift in both sales and customer volume throughout the trial, consistently outperforming its control store. The strong March spike followed by April stabilization confirms the layout’s lasting impact, making Store 88 a prime candidate for rollout.
Store 77 experienced a clear uplift in both total sales and customer volume during the trial, while its control store (233) declined steadily. However, the growth was driven by increased foot traffic, as average spend per customer fell, suggesting the layout attracted more shoppers but didn’t boost individual spend.
Store 86 experienced a temporary surge in total sales and customer volume mid-trial, peaking in March before declining to match the control store by April. This pattern suggests initial trial engagement that was not sustained. The control store remained relatively stable, reinforcing that the trial effect in Store 86 was short-lived and may require refinement or deeper diagnostics.
Reflection & Conclusion
Reflection & Conclusion
This project strengthened my ability to turn raw data into actionable insights. Task 1 highlighted how segmentation can guide targeted marketing and pricing. Task 2 deepened my understanding of uplift testing and trial evaluation. Together, they reinforced the value of data-driven strategy in improving customer engagement and commercial performance.
Page updated
Google Sites
Report abuse