84.51° Projects and Automated Machine Learning

The website of 84.51° provides a few revealing numerical facts that convey the enormous size and scope of their data science efforts:

• 1250 consumer packaged goods partners

• 60 million households

• 1 billion personalized offers delivered to customers last year

• Over 10 petabytes of customer data analyzed

• 3 billion customer shopping baskets analyzed

• 138 different machine learning models in production

Many of the group’s predictive models are used every day by Kroger. For example, the sales forecasting application creates forecasts for each item in each of more than 2500 stores for each of the subsequent 14 days. In most companies, these types of sales forecasting models are updated rarely or never, but sales forecasting for Kroger is dynamic. These forecast models are updated on a nightly basis based on the most recent data. Using another 84.51° capability, “Kroger Precision Marketing” analyzes the relationships between media exposure and store sales. It uses customer purchase data to make brand advertising more addressable, actionable, and accountable. Over the past three years media campaigns for over 1000 brands have been orchestrated using the results of this data science-driven analysis.

Click here to read the full Forbes article.