Image courtesy of rawpixel.com
The Problem
A B2B industrial supplier with a network of retail chains wanted to determine the state of their loyalty program. Like most loyalty programs, customers can sign up and receive points from every purchase. These points can be redeemed for various rewards online. While the program has been running for a few years, it’s success has never been analyzed in a quantitative fashion. Critically, the program had no connection back to the impact of sales or on actual customer loyalty.
The Process
I dove into this project with the desire to answer one question: Is the loyalty program bringing actual value to the company? To answer this, I began a project to build out a dataset to analyze the loyalty program’s connection to sales. Using Tableau Prep, their BigQuery environment, and Tableau Server data sources, I went to work. I had to tackle a lack of documentation to make the dataset possible. This required careful coordination between different teams to gather information and validate my work.
Once the dataset was built and validated, I moved to Tableau Desktop to analyze the data. I focused on a handful of main metrics to compare:
- Program Utilization: Those who are active in the program
- Average Sales per Customer: Those in and not in the program
- Average Orders per Customer: Those in and not in the program
- Average Revenue per Order: Those in and not in the program
- Sales Lift: Those in the Program
From the initial buildout in Tableau Desktop, it was clear that customers in the program were spending and ordering more on average than those not in the program. This was hopeful, but we had to drill deeper. Might only the existing big customers simply be taking advantage of the program?
Sales lift was the key metric here. For customers who had activity before and after signing up for the loyalty program, we compared their transaction activity in those two periods. Did they increase their spending and orders after joining the loyalty program? Yes, they did, on average! Even after controlling seasonal fluctuations with a smaller sample size, we saw an increase in sales after signing up for the loyalty program.
The Outcome
Given what was discovered, we had enough evidence to make a reasonable assumption that the program was driving revenue and loyalty. It was hard to put a specific figure on revenue. The average sales between those in and not in the program was the best estimate. But it left elements uncontrolled for. Regardless, I was able to create a statement of ROI to help marketing leadership advocate for this program.
My main recommendation was to continue to increase loyalty program participation. The analysis I did in Tableau Desktop became a tool for sales to understand their book of business. The sales field is the companies’ number one asset in increasing participation. So, the Tableau dashboard was critical.
Beyond this. I recommend that the company should invest in a project to piece together attributes on each customer. This will lay the foundation for customer segmentations to target promotional offers. Including those to join the loyalty program.