Instacart Customer Segmentation & Revenue Optimization

Overview

Built a customer analytics framework for Instacart's marketing team, analyzing 32.4M order-product records across 206K customers to identify high-value segments and optimize marketing allocation. The project delivered customer personas, revenue concentration insights, and actionable recommendations for ad spend, targeting, and cross-sell strategy.

Business Problem

Instacart had a large customer base but lacked clarity on which customers drove the most value. Revenue was concentrated in unknown groups, and the marketing team needed data-driven segmentation to allocate spend more effectively and identify growth opportunities.

Business Objectives

Segment customers by behavior and value
Identify top revenue personas
Understand when and what customers buy
Optimize marketing allocation using data

Data

32.4M order-product records
206K customers across 3.4M orders
5 merged datasets (orders, products, customers, departments, demographics)

Analytical Approach

Data was cleaned and merged using pandas, then analyzed through a customer segmentation framework with behavioral flags and revenue tier concentration analysis. 27 visualizations were produced using matplotlib and seaborn.

Scatter plot showing correlation between income and spending power by region — High-income customers ($120K+) spend similar amounts per order as middle-income customers, revealing untapped revenue potential.

Key Findings

Bar chart showing 79.5% of Tier-1 revenue from older adult households — Tier-1 revenue is heavily concentrated in older adult households, as 79.5% of top-tier revenue comes from the 56+ segment.

Top Insights

56+ adults drive 80% of Tier-1 revenue → primary target segment
Families generate 87% of order volume → core revenue engine
Demand peaks 10am–4pm and weekends → optimize ad timing
High-income customers underspend → upsell opportunity

Regional department mix heatmap showing minimal variance across regions — Regional department mix shows minimal variance, supporting a centralized national growth strategy.

Recommendations

When (Ad Timing)

Shift ad spend to low-utilization windows (12am–6am Tue–Wed)
Deploy automated push and discount campaigns
Reduce spend during already saturated peak periods

Who (Targeting)

Prioritize family households as core revenue engine
Target 56+ and health-focused families (highest ROI)
Upsell high-income customers with bulk and premium offerings

Deliverables

Final Excel report delivered to the client containing all visualizations, analytical breakdowns, key questions answered, and marketing recommendations.

For a full technical breakdown including code, queries, and methodology, view this project on GitHub.

← All Projects