Our client, one of the world’s largest premier hotel chains, aimed to increase customer order spending by providing recommendations for combo offers.
Overview
Solution
To achieve this goal, we undertook a comprehensive approach:
- Exploratory Data Analysis: We conducted a thorough exploration of order data to identify trends in item combinations within various orders.
- Seasonal Analysis: We leveraged order data across different seasons, recognizing that certain menu item combinations varied with seasonal shifts.
- Market Basket Analysis: For each combination of menu items taken two at a time, we conducted market basket analysis. We calculated support, confidence, and lift metrics using transactional data.
- Threshold Determination: By establishing thresholds for support, confidence, and lift, we determined the ideal combinations of menu items to recommend to customers.
Output
Our analysis yielded valuable insights and outcomes:
- The likelihood of finding Belgian waffles in transactions with orders for eggs is 19 times higher than the typical probability.
- Offering new combo deals led to an incremental order revenue gain of 1%.
LHS | RHS | Support | Confidence | Lift |
Belgian Waffles | Eggs to order | 0.1% | 49% | 19.95 |
- Calculation for Incremental Revenue Gain per Day: = (0.9 * 400 * 70 + 200 * 30) – (400 * 50 + 200 * 50) = INR 1,200
- A 10% discount was applied to the combo deal.
- Acceptance probability increased by 20%.
- The price of individual items was set at 200.
This case study demonstrates how our market basket analysis approach enabled our client to make informed decisions, resulting in increased customer order spend and improved revenue.