P R O J E C T
Creating Gas Station Customer Insights
Utilizing transactions from a network of gas stations to create a method of monitoring customer behavior and preferences.
B A C K G R O U N D
The company manages a large network of gas stations across the main land United States including both fuel pumps and the surrounding convenience stores. They have a system in place to record the customer transactions at each point of sale.
They have been using a loyalty program to identify their customers who use their loyalty card alongside their purchases.
This project looked to expand their existing simplistic approach to be able to track the behaviour of customers who are not affiliated with the loyalty program. By associating purchases from the same bank card and also bank cards that have historically been associated with a loyalty card.
To do this we created a set of business rules that would switch which payment methods were associated with a hypothetical customer's wallet.
This allowed us to observe customer behaviours in loyalty card customers and apply that insight to estimate attributes of anonymous bank card users.
i.e. We don't have any provided information for this bank card but we can observe that its owner is behaving similar to our loyalty card users in the 18-24 age range.
T E C H N O L O G I E S
C H A L L E N G E S
Unlike many of the other projects, the toughest part of this project was scoping out the expectations and getting a clear picture of the acceptance criteria.
The next layer of challenge that this project posed is the sheer complexity of the business logic at hand, it was a gargantuan task to implement this in a way that has the ability to scale with the quantity of transactions being sourced.
There were a significant number of quirks and edge cases that came with the requirements that were baked into the design. One of these for example is that all cash transactions need to belong to a wallet, however since there is no way to tie any cash transactions together they all belong to the same wallet, implying to the analysts that there's one trucker with millions of dollar zipping around the continent.
P R O J E C T
O U T C O M E
We successfully created a "wallet" entity that made it easier and more accurate for analysts to monitor customer behaviour.
Analysts then produced new insights that were previously unattainable. Fuelling better business decisions driven by data.
The approach was built to be maintainable and flexible to any changes in how the business decides how wallets are handled.
P E R S O N A L
G R O W T H
This project was the teetering on the precipice of failure throughout. To achieve a project of this scale was not simple to estimate in advance, as a result I grew a lot in my ability to work under time constraints and to communicate when expectations need to change.
The complexity of the design is likely the most intricate and nuanced set of requirements that I have implemented into an existing warehouse. My ability to create complex logic in a way that is not too rigid or inflexible to maintain.
Documenting and demonstrating the efficacy of this approach helped me improve my ability to explain technically complicated work to business stakeholders and distill the useful information to the right people.
