DOWNLOAD THE FULL WHITE PAPER
VIEW WHITE PAPEROUR METHODOLOGY
Our Five-Step Simulation Process
The Lenati Loyalty ROI Simulator uses a five-step process. These steps are based on best-in-class data science techniques that understand how customers will behave under different treatments.
NEXT
BACK
BACK
NEXT
1
Create Baseline Customer Behavior Profiles
The simulator uses historic or recent customer data to create a profile for how each customer is likely to behave without a new program. This profile includes information on the size of orders the customer tends to make, the frequency of ordering, and the expected date when the customer will stop shopping. Developed using data science models, these profiles provide unique values for each customer.
Customer Behavior Profile
The customer behavior profile is a combination of core mathematical components. A unique value is created for each customer profile.
QUESTION
Great! Leveraging real customer data provides more accuracy for how the model should behave and thus a more accurate ROI prediction.
That’s okay! Additional primary and secondary research can be fielded in this first step.
2
Simulate Future Behavior Based on Customer Profiles
Using the customer profiles from step (1), the simulation predicts what the future orders will be for each customer during the selected time horizon. These orders are based on conditions staying the same as they were historically; no new loyalty program features are added yet. These future orders serve as a baseline to compare new loyalty programs.
SAMPLE ORDERS OVER TIME
A sample of historic and simulated customers, where each row is a customer and each point is an order.
3
Define Loyalty Program Scenarios
A selection of possible loyalty programs are chosen, and each has its own set of drivers. A driver is a component of a loyalty program that affects customer behavior. For example, a driver may be an extra coupon on the customer’s birthday. Other drivers can include a customer’s desire to spend extra to move to the next tier, or a rewards coupon to redeem before the expiration date. The amount each driver influences the customer’s behavior is calculated.
EXAMPLE DRIVER
BEHAVIOR CHANGE
EXAMPLE DRIVER
Birthday coupon
BEHAVIOR CHANGE
Extra order around birthday
EXAMPLE DRIVER
Tier motivator
BEHAVIOR CHANGE
Order more if close to next tier
EXAMPLE DRIVER
Free shipping
BEHAVIOR CHANGE
Larger orders online
4
Re-run the Base Simulation with Each Loyalty Program
The baseline in step (2) is adjusted for each loyalty program scenario based on the loyalty drivers in step (3). These drivers are implemented using analytical rules applied to each customer order made in the baseline scenario. Potential adjustments for each customer include: larger orders, ordering sooner, or incurring additional expenses from using benefits of the rewards program. These adjusted orders are stored for comparison.
SAMPLE SIMULATION
Program drivers and motivators by customer segments are stored for comparison across programs.
SAMPLE SIMULATION
The revenue and expenses within a scenario can be broken down to gain a detailed understanding of the predicted ROI.
5
Compare Estimated ROI Under Various Scenarios
Once the simulations are complete, all the results can be compared side by side. At the aggregate level, we can see the overall ROI of each scenario to find the one with the highest ROI. We can also deep dive into particular segments of customers to understand what is driving revenue in a specific loyalty program. This simulation methodology provides detail into how each customer is incurring different program expenses, therefore enabling stakeholders to decide if an individual benefit is worth the incremental cost. The approach provides a detailed set of reporting and the ability to drill down into the data where necessary.
SHARE
