ATM Pricing Optimization: A Case Study
A large branch banker optimizes ATM pricing.
- The bank had employed a single fee for cash withdrawals by non-account holders in all of its ATMs.
- Using APT's Test and Learn for Sites™ solution the client designed and executed a number of different pricing tests in a cross section of ATMs across its network.
- Analysis using APT's Test and Learn for Sites™ showed significant differences in pricing impact by market and by ATM location type.
- The bank rolled out the price changes tailored by location and realized increased profits of over $10 million per year.
The Company:
A large branch bank with thousands of Automated Teller Machines (ATMs) across the United States
The Challenge:
The client had employed a single fee for cash withdrawals by non-account holders in all of its ATMs. Management was unsure whether its pricing structure was optimal in absolute terms (i.e., if the single price was the best price), and, more importantly, whether there might be different prices that would maximize profits at different machines: Did higher prices increase profits despite lower utilization? Did lower prices stimulate increased utilization that more than offset margin erosion? Were there significant differences by geography or ATM location type?
The company had attempted to analyze the question, but management was not confident in the accuracy and precision of the analysis, and no action had been taken. Specifically, the analysis had not successfully identified the attributable impact of price changes within the overall industry trend of declining fee-generating transactions.
The Solution:
Using APT's Test and Learn for Sites™ solution the bank designed and executed a number of different pricing tests in a cross section of ATMs across its network. APT's Test and Learn for Sites™ solution was used to isolate the impact of price changes on ATM performance, comparing performance of test ATMs to a scientifically selected set of ATMs matched on certain attributes, seasonality, and growth patterns. Using this approach, the client was able to produce a highly detailed and confident measurement of the actual impact of price changes.
The outputs of the analyses showed significant differences in pricing impact by market and by ATM location type. APT's Test and Learn for Sites™ solution developed a predictive model that identified the optimal price for every one of the bank's ATMs - some ATMs optimized profits by pricing up, some by pricing down, and some by remaining with the current pricing. These predictions were rolled up by market and location type to provide an implementation-friendly approach to execution with maximum positive profit impact.
The Results:
The client rolled out the price changes and realized increased profits of over $10 million per year. The client continues to selectively test new ATM pricing levels as the markets continue to evolve and has improved profits with fine-tuning adjustments to pricing when warranted.
Next Steps:
