How Fulton Ring unlocked a multi-million dollar opportunity in Additech’s retail data
Case Study
Understanding which new markets to sell fuel additives in has been a major challenge for Additech. With their product currently available in only a handful of regions, choosing the wrong locations could mean wasted investments and missed opportunities. To solve this, Fulton Ring partnered with the Additech technology and leadership team at the end of 2024 to extract sales data from one of their retail partners, uncovering key trends that revealed exactly where new dispensers should be placed. Now armed with data-driven insights, Additech can confidently target the right stores, optimize expansion, and drive significant revenue growth without guesswork.
The Challenge
For years, Additech had been collecting terabytes of data from retail stores their pumps were installed at, but lacked a data warehousing solution to finally analyze it. We both agreed that there was a treasure trove of information hidden in this data, so we took on the challenge of organizing it for the org to help uncover insights that would drive future revenue growth.
Our Solution
In order to uncover trends in their dataset, we would need to provide a real time reporting solution that would allow Additech decision makers to gain a birds-eye view of their data. The scale of the problem, while not at a Netflix or Google level, still represented a non-trivial amount of data. If we wanted to provide Additech with real time dashboards to visualize trends, we would need a highly performant time series data warehouse and an automated pipeline to ingest retailer data.
Initially, we looked into time series databases like TimescaleDB to store this data, but quickly ran into query time scaling issues as millions of transactions were being entered per day. To solve this, we adopted a data lake architecture, loading transaction data in Parquet format into S3 and querying it with Athena on AWS. This serverless architecture resulted in:
- Queries times for terabytes of data reduced from several minutes to under 8 seconds — fast enough for our real time dashboards
- Over a 92% reduction in expected costs compared to running databases on EC2
We highly recommend using data lakes to store large amounts of data at very little cost. While traditional data warehouses have the advantage of being more flexible, they require expensive infrastructure to maintain. Data lakes like the one we built are well suited to storing data that isn’t expected to change. Additionally, services like Athena are able to query structured data almost as quickly as a data warehouse without having to run 24/7.
We also recommend using workflow orchestrators like Dagster to automate transformation pipelines. While their infrastructure is not as native to AWS as something like Lambda and Step Functions, it provides a much better developer experience by reducing the amount of time spent writing DAGs, wrestling with permissions, and debugging data flow failures. Running Dagster on EC2 instead of a serverless solution also brought down the cost of doing the backfill with a large volume of data. Had we been billed for the amount of data processed (without any retries) instead of a fixed rate for a month of machine time, the cost would have been at least twice as much. To scale this solution, we’d recommend running Dagster on a containerized platform like ECS.
Once these datasets were available, we built dashboards for Additech using Amazon QuickSight to visualize trends about customers shopping at this retailer. These included:
- Loyalty account insights to better understand where and how often customers were shopping the most to generalize if recurring transaction volume correlated to higher Additech sales
- A dashboard to track average transaction volume and detect high volume days
- An adjustable projected revenue calculator to determine expected sales at locations where Additech is not installed to learn if a site would be profitable, and how reliable that estimate would be. This shows Additech exactly where to expand nationwide from the handful of states they’re currently in.
This solution showcases the power of modern cloud infrastructure in handling large-scale data challenges with speed, efficiency, and cost-effectiveness. By leveraging AWS services like Dagster, Athena, and QuickSight, we built a robust and reliable architecture that eliminates the need for costly database maintenance while delivering near-instant analytics. Additech now benefits from a reliable solution that meets them at their current scale while keeping infrastructure costs at a fraction of traditional data warehouse solutions.
Results
In conclusion, the Fulton Ring and Additech partnership successfully unlocked a multi-million dollar opportunity by leveraging retail data. Through the adoption of a data lake architecture and the implementation of Amazon Quicksight dashboards, Additech gained valuable insights into customer shopping trends, transaction volume, and potential revenue streams. The project not only resulted in improved decision-making and increased efficiency but also provided a foundation for future growth.
If you believe that you’re sitting on a gold mine of information and want to see how you can drill into it, get in touch with us. We’d love to see how we can help you grow.
