Retail Giant Optimizes Performance for over 13,000 Employees using AutoML

This retailer is considered to have revolutionized the customer experience with millions of visitors per year. This retail and entertainment center has changed the way that retail and entertainment interact with one another and is considered one of the top tourist destinations in the world. 

Challenge

With approximately 13,000 employees, it was important for this retailer to have the ability to understand and predict how much foot traffic would come through the mall on a daily basis. That way, they could optimize the efficiency of scheduling their workforce, ensuring that they had enough staff available to adequately handle the traffic, but not waste money on overstaffing.

To address this issue, this retailer ran an effective, albeit manual forecasting model in the form of a Python script to predict daily traffic, based on historical traffic counts that were determined by parking lot data. Because this was a time and labor-intensive process, the report was only run once a quarter and predicted daily traffic for the next three months. Looking this far into the future was problematic as it did not take recent mall trends, events or promotions into account.They needed a better, more cost-effective solution, one that could be implemented quickly, would run automatically, and was easily accessible to decision-makers. And that’s where CCG came in.

Solution

CCG was engaged by Microsoft as a trusted partner to leverage their cutting-edge data science capabilities to provide this retailer with an automated solution. CCG utilized automated machine learning (AutoML) to rapidly develop a basic forecasting model that could be run on a daily basis without any oversight. Using the Databricks platform, CCG built the model which ran automatically via Python scripts. Every night, the scripts automatically update the forecasted traffic values for the next two weeks, and every weekend, the scripts automatically assess how well the forecasting model is performing and make adjustments as necessary for more accurate predictions.

In this way, the model learns to use the most effective methodology for the most reliable results. Additionally, CCG created visual dashboards in Power BI to graphically present the model’s predictions, as well as comparing the accuracy of the predictions with actual traffic. These data visualizations make these traffic forecasts easy to understand, assess and share. Beyond providing this solution, CCG presented this retailer with detailed documentation and heavily commented code, so they could easily replicate the processes used in future projects. Their team greatly benefited from learning how to to schedule a predictive model easily and put it into production themselves – saving them time, effort and money.

Now, this retailer can quickly, easily and automatically generate daily traffic predictions, and those predictions can take into consideration the most recent mall traffic values and trends. Where they had previously been developing and running manual models on their personal computers, the this retailer team can now take advantage of a robust Data Science platform that can be used for future predictive models and projects.

Want to get starte with Azure Machine Learning? Learn the basics in this blog: Introduction to Azure Machine Learning

Quick Facts

  • Industry -Retail
  • Solution - Staffing Optimization for Retail
  • Technology - AutoML

54%

Statistics show that 54 percent of business executives say that their adoption of AI within the workplace has led to a boost in productivity (PWC, 2018).

Join the hundreds of CCG customers.
Talk to Intelligent Enterprise Guide.