Financial Technology Firm Lending Case Study
In this study, we look at a financial technology firm that had been working on a business financing product, this product would offer access to fast and flexible financing for small and medium-sized businesses. At the same time, a team in a different division worked on extending their existing card issuing product to include credit options. A problem materialized: The two products have a mutual data dependency on the total lending position of a customer. ReJot was able to help unravel the dependencies by providing deep data integration between teams through its synchronization engine.
Key Results
- ReJot significantly reduced the time to market for the capital loan product. Development cycles were saved for both product managers and engineers.
- A single source of truth for the total lending position of a customer was created, reducing the risk of inconsistencies and errors. Minimal technical debt was created in the process.
- Product teams are empowered by staying in charge of their data, instead of becoming implicitly dependent on ETL pipelines. Any future product enhancements are not hindered.
Project
For the FinTech firm, it is crucial to determine the total lending position of a customer before approving a loan or any other financial product. For either of the two products to be able to go live successfully, data from both teams needs to be integrated:
- Before setting a credit limit, any currently open capital loan must be taken into account.
- Before approving a capital loan, the customer’s total credit limit and open credit must be known.
One key requirement is that this information needs to be considered in real-time, when the customer requests a loan or a credit limit increase. This request is usually made through a REST API.
In this case, the firm decided to go live with the credit card product first, as it is an addition to an existing product. Thus, the capital product had to be delayed. A decision had to be made on how data would be shared between the two teams. To still be first to market with the financing product, the teams had to improvise.
Options
In an alignment meeting, the teams discussed a number of options available to them. The main goals were to reduce time-to-market and minimize technical debt created, as both teams had their planning to consider. The following options were discussed before the introduction of ReJot.
Option 1: ETL - Reverse ETL
The company has an internal data platform capable of combining data from multiple sources. It seemed like a decent choice for the problem at hand. Yet, there were some clear downsides:
- Data Analysts own the ETL pipelines at the company, alignment between product and data teams would delay the go-to-market date.
- ETL pipelines are batched, rather than real-time. Either one of the teams would have stale data (by a day for a nightly batch job) from the other team.
- Reverse ETL was not a typical use case for the data team, there was no clear path forward without involving infrastructure teams.
Option 2: Event Streaming
An alternative considered was the event queuing/streaming platform for asynchronous communication between teams. Again, a number of downsides would cause problems:
- Integrating events from another team would mean setting up bespoke consumers, something that would require additional development time. Both teams would need to integrate events from the other team, meaning they would both have to implement similar, yet different, logic.
- New topic-partitions would have to be created, something that was not self-service available to product teams at this company.
Option 3: Intermediary Service
The last option would have both teams integrate with a to-be-created intermediary service. This service would become the source of truth for any credit given out to a single customer. Although the idea of having a single source of truth for credit limits is appealing, there would be several downsides:
- The new service would require a significant amount of development time.
- Ownership of the new service would be unclear, as it would be used by multiple teams.
- The new service would have to be maintained, which would add to the ongoing costs of the project(s).
Using ReJot
ReJot’s integration through synchronization approach changes the way the teams look at the problem.
To solve the problem at hand, both teams defined a ReJot contract which represents the data used by their respective products. This contract is automatically entered into the catalog of the company. From this data, a “total lending position” is derived. ReJot’s synchronization engine keeps the data up-to-date. Both teams then use this data in their products, and make decisions based on it. This elementary level of data modelling and integration is the simple strength of ReJot.
This solution is conceptually similar to the “Intermediary Service” option, but the downsides do not apply because of the automatic synchronization.
Insights
Jump back to key insights.