FlexField Fitness Strategic Partnership Analysis | Datathon Semi-Finalist

Project Overview

Data-driven analysis to identify the optimal strategic partnership for FlexField Fitness, a declining gym chain facing customer retention challenges. Analyzed 3,494 customer records across four potential partners (Chef's Meal, PulseGear, CoreBoost) to recommend a partnership strategy that addresses 30% membership dropout rates.

Business Problem

FlexField Fitness experienced declining membership due to competition from budget gyms and digital platforms like Peloton. The company needed to determine which partnership—food delivery, athleisure apparel, or sports supplements—would best improve customer loyalty and create compelling joint products.

Technologies Used

• Python: pandas, NumPy, scikit-learn, statsmodels, scipy
• Statistical Analysis: Chi-squared testing, OLS regression
• Data Visualization: seaborn, matplotlib
• Development: Jupyter Notebook

Technical Implementation

Data Processing & Analysis

• Datasets: 4 Excel files with 3,494 total records
  • FlexField Fitness: 897 gym members
  • CoreBoost: 886 sports drink consumers
  • PulseGear: 888 apparel customers
  • Chef's Meal: 823 meal delivery users
• Cleaned data using pandas, handled missing values with SimpleImputer, removed duplicates
• Applied one-hot encoding for categorical variables (Gender, Fitness Goal, Dietary Preferences)

Statistical Testing

Chi-Squared Analysis to test association between fitness goals and partner products:

• CoreBoost: χ² = 224.73, p-value = 2.47e-41 ✓ (highly significant)
• Chef's Meal: χ² = 2999.53, p-value = 0.319 (not significant)
• PulseGear: χ² = 3382.01, p-value = 0.468 (not significant)

Regression Modeling

Built OLS regression models using statsmodels for each partner:

CoreBoost Model (R² = 0.162):

• Hours at Gym significantly predicts consumption (coef = 0.796, p < 0.001)

Chef's Meal Model (R² = 0.777):

• Calorie Intake (coef = 0.021, p < 0.001) and Whole foods preference (coef = 10.28, p < 0.001) are key predictors

PulseGear Model (R² = 0.287):

• Gender is strongest predictor of apparel spending (coef = 211.94, p < 0.001)

Customer Segmentation

Analyzed spending patterns by fitness goal:

• Build muscle: 356 customers (40% of total)
• Fat loss: 183 customers
• Performance apparel: $218,919 total spend (highest category)

Visualizations

Created data visualizations using seaborn and matplotlib:

• Consumption patterns by fitness goal and product type
• Linear regression plots of gym hours vs. product consumption
• Grouped bar charts comparing demographics across segments

Decision Framework

Multi-Criteria Decision Matrix

Evaluated partnerships on 4 weighted criteria:

• Strategic Alignment (30%)
• Customer Engagement Potential (30%)
• Revenue Growth & Profitability (20%)
• Feasibility & Implementation (20%)

Results:

• CoreBoost: 8.2/10 ✓ (Recommended)
• PulseGear: 6.6/10
• Chef's Meal: 6.0/10

Key Recommendation: CoreBoost Partnership

Why CoreBoost:

1. Strongest statistical association with fitness goals (p-value near zero)
2. Direct link to workout performance and recovery
3. Year-round revenue stream (mitigates seasonal attendance drops)
4. High personalization potential for different fitness goals

Key Insights

• CoreBoost scored 37% higher than alternatives on strategic metrics
• Identified 30.35% average membership dropout rate requiring intervention
• Each additional gym hour correlates with 0.796 bottle increase in sports drink consumption
• Performance apparel market represents $218,919 secondary opportunity

Business Impact

Delivered actionable recommendation with statistical validation (p < 0.001) that addresses customer retention through performance-enhancing products aligned with fitness goals. Partnership strategy positions FlexField to compete against digital fitness platforms by creating integrated gym + nutrition experience.

---