Big Data

The Client

A leading North American automated retail company managed 40 thousand automated kiosks throughout the United States and Canada processing over 2 million transactions per day. Surveys found that allocation of product in kiosks was not optimized to satisfy customer needs. 

The Problem

Kiosks were not optimally loaded to ensure availability of content by location.   The challenge was to identify a root cause and create a solution to allow users to find merchandise of their interest on the first visit to the kiosk. The challenge was to identify root cause and prevent.

The Solution

We mined the available transactional sales data, demographic data, and store metadata to identify a root cause of the problem. 

A kiosk-specific rental forecasting algorithm was defined and developed that helped ensure that desired product was available at each kiosk at the correct times. 

In addition, the algorithm could determine the likelihood of product being returned to a specific kiosk. This combination of new information allowed field employees to optimize the product to stock in each kiosk.

Data Collection & Automation

Retail Transactional Data

Demographics data via 3rd party

Store Metadata

The Results

Customer achieved 10% margin improvement by utilizing automated data collection, data mining, and implementing forecasting algorithm that allowed expected product to be available at the right time on a kiosk-by-kiosk basis. 

At a glance

INDUSTRY

Retail

CHALLENGE

Kiosks being either overloaded or underloaded. Need to identify root cause and prevent.

SERVICES USED

Data Analytics Platforms

ETL

Data mining algorithms

Automate data collection

RESULTS

10% margin improvement by utilizing data science and implementing forecasting algorithm that enhanced Customer Experience.