Olist - Brazilian e-commerce challenge

Increasing Olist's profits using in-depth data analysis

Language: Python

1 Problem

How to increase Olist’s (Brazil’s largest department store) monthly profits while maintaining a healthy order rate?

2 Solution

This Data Analytics project makes use of the Brazilian E-Commerce Public Dataset to provide applicable solutions for increasing Olist’s monthly profits in a healthy and sustainable way.

3 Objectives

To address the problem, two objectives have been set:

1. Identification of the main sources of loss (low-performing sellers)
2. Simulation of two loss-reduction solutions

4 Results

4.1 Identify the worst-performing sellers

The worst-performing sellers have been identified as the sellers with the highest monthly orders, usually with more than 80 orders per month.

4.2 Loss-reduction solutions

Solution 1: limiting strategy

By limiting a seller’s number of monthly orders to 30 when their share of 1-star reviews are >10%, Olist’s monthly profits sustainably increase by 0.8%.

Benefits:

• no sellers are banned –> Olist doesn’t lose customers
• Reduce “bad sellers” negative impact by 3.5 factor

Trade-offs:

• Low impact on monthly profits: only 0.8% increase

Solution 2: ban strategy

By banning the worst-performing sellers from the online platform, Olist’s monthly profits increase by 17%.

Benefits:

• High impact on monthly profits: 17% profit increase

Trade-offs:

• Olist loses customers and monthly orders