July 10th, 2016 | Analytics, Big Data, Data Science

Applying Machine Learning to the retail sector

Europe’s third largest food sector franchiser growing with Machine Learning techniques with Intellignos

One of the world’s largest discount supermarket chains applies Machine Learning techniques for its 846 stores and online business starting in Argentina.

Due to their success in the last decade, the company was swamped in data. With more than 5 million clients (an astounding 12.5% of the country’s population) in their CRM and database, it was extremely difficult to make sense of so much information, let alone generate successful marketing initiatives through advanced segmentation and predictive analysis.

In a highly competitive market, the more you know about your consumers, the better your chances are to capture market share and improve revenue.

The problem was clear: an untapped database with a huge amount of information and the need to provide high value to customers.

With this, Intellignos set sight on the solution: a set of Machine Learning techniques that would enable the company to exploit all its data to its maximum capabilities.

But…”what is Machine Learning?”
“Machine Learning is a set of statistical modelling techniques, which applied to data, use algorithms to detect behaviors in the past and apply that knowledge to situations in the future.

With Machine Learning we can generate predictive models, so as to respond to business problems and help better decision-making.”

The solution was centered in the use of all demographic and transactional data in order to improve:

  • Offers and impact of marketing communication
  • Improve purchase experience and benefits for heavy buyers
  • Reduce Churn and customer inactivity.

Different approaches were used to address each of these goals, some of them:

1. Cluster Analysis
It’s a Machine Learning model that allows to partition a database into homogeneous segments based on behavior, in order to group individuals who are similar to each other according to certain features.

Through this analysis, individuals who resemble each other regarding demographic and purchasing behavior characteristics were grouped, which allows to generate:

  • Differentiated marketing actions.
  • Promotions, discounts or commercial activities.
  • Loyalty actions based on behavior.

2. Cross-sell by Segment
It is a statistical model that includes all data from consumer behavior. The learning about customers, their purchasing habits, tastes and needs, facilitates the sale of complementary products the customer uses or intends to consume.

In this way, for each segment there is a set of products to suggest – together with the best strategies to carry out this approach – so as to maximize the value of each client.

3. Customer Activation
Customer activation was approached through a set of machine learning techniques that allow, through the study of past events with behavioral characteristics associated with them, understand which of these help explain its occurrence and apply that knowledge to the future.

For those customers who are inactive and whose behavioral characteristics have resemblance to a group of active customers, the company can perform specific actions that incentivize consumption in order to reactivate these clients.

4. Market-Basket analysis (bundle creation)
It’s an exploratory type of tool, applied to details of purchased items by customers in their transactions, in order to provide visibility into which products tend to be purchased together.

Through a series of patterns or rules which contain information about products typically bought simultaneously, bundles, promotions and product-groupings are created in order increase volume and profitability.

These initiatives led by Intellignos’ Data Science department, together with the ongoing support from Intellignos’ business team have helped the company to grow and generate better tools for sales planning, Marketing ROI and resources distribution.


Join our Newsletter and receive periodical news and updates: