Personalizing the consumer lending experience
Oakbrook leverages the Databricks Data Intelligence Platform to simplify and personalize borrowing
Projected cost savings over the first 3 years
New financial products developed in 12 months
Revenue increase forecast over the next 12 months from new top-up loans
“We were looking for an open, flexible platform that would natively support our data science, advanced analytics and different languages. Furthermore, it would need to integrate with our platform hosted in Google Cloud and scale cost-effectively. As a regulated business, governance, security and auditability were also non-negotiable. In the end, Databricks Data Intelligence Platform was the only solution that ticked all the boxes.”
Stringent regulation and outdated approaches to consumer lending have made it more difficult for many consumers across the UK to obtain loans. However, fintech organizations such as Oakbrook Finance are revolutionizing the lending sector, bringing more agile decision-making driven by advanced data science models and analytics. With ever-growing data sources, storage requirements and governance challenges, Oakbrook realized they needed an open, scalable platform designed for big data processing and advanced analytics — one that would allow them to rapidly develop new products, such as Finio Loans. Oakbrook chose Databricks Data Intelligence Platform on Google Cloud to transform their lending process with data and AI, giving eligible consumers fairer access to affordable credit based on individual circumstances. Now the company has a modern infrastructure designed to simplify the management and analysis of large volumes of data to deliver better, faster lending insights for their customers.
Lack of scalability, storage and flexibility hindered growth
Oakbrook’s products are designed with flexibility, simplicity and positive customer experiences in mind, and are underpinned by their proprietary loan application and decisioning platform, known as “O6K.” Customers can access the company’s products directly through the Oakbrook website, or can apply through price comparison sites and aggregators. The organization has in excess of 100,000 customers and has offered £1.1B in loans since launch. A wealth of structured and unstructured data is generated across the company, and can come from many different sources, including consumer credit files, bank transactions, audio files from customer service calls, as well as up to 10 million events published by their O6K platform each day.
With data volumes rapidly increasing, Oakbrook faced critical challenges with their legacy SQL Server Integration Services (SSIS) environment. Optimized for simple SQL queries and reporting, the platform failed to support more advanced analytics or data science, such as predicting the risk of providing a loan to a specific individual. Data teams writing SQL were also using other languages like R and Python, which prevented them from collaborating. Processing was often taking place outside the main data environment — leading to increased storage demands and governance risks. Major issues included limited storage, the inability to scale for larger workloads, complex and unreliable data flows, and siloed data wrapped in third-party systems, all leading to a poor experience for their engineers and analysts. A new platform had become essential.
A unified environment for increased efficiency, collaboration and productivity
By moving their lakehouse into Google Cloud, Oakbrook’s ETL processes have been optimized, and are now quicker than before and require less code. With the Databricks Data Intelligence Platform, data scientists working in R or Python are able to work in a unified environment, improving efficiency and governance.
The lakehouse has also unlocked more resources. Previously, data scientists were running heavy SQL queries against data overnight, blocking the system for other users. On the Databricks Data Intelligence Platform, users can have their own clusters and run jobs without impacting other teams. Oakbrook’s analysts can also run SQL queries and schedule them in a familiar environment within the lakehouse. This, and the ability to access data easily and create reports and dashboards, has been critical to achieving engagement on the platform.
The lakehouse architecture has brought additional benefits, including security, resiliency and high performance as standard, with the ability to integrate with services on other clouds, where Oakbrook’s teams have existing expertise. Collaboration has also significantly improved, bringing software engineering and data engineering teams closer, allowing them to deliver more data even faster than before.
Adding further data onboarding functionality, Oakbrook has introduced Fivetran — an official Databricks technology partner — providing analysts with even more insights, quickly and efficiently, while saving valuable data engineering time. Instead of creating bespoke ETL processes, data engineers can now use Fivetran to ingest data into the lakehouse in a low-code way. This has enabled previously siloed data to be extracted from Oakbrook’s call center and marketing system, allowing analysts to explore data on customer queries and marketing preferences — ultimately generating deeper insights and improvements in customer experience. The teams can also onboard new data sources in half an hour, rather than a matter of days.
Driving innovation and personalization for better customer outcomes
Oakbrook now has a platform that is faster, scalable, more reliable, and secure, leading to considerable cost efficiencies and valuable new business use cases. Over the first three years, the company projects that they will realize cost savings of 14% through productivity gains and optimized compute.
The company also expects to deliver new innovations in their lending and collection strategies much more quickly than before. Changes that would have previously taken weeks or months can now be conducted in less than half the time. Indeed, the company has been able to launch two new products over the last 12 months, as well as a top-up loan and collections decision engine. Thanks to the Databricks Data Intelligence Platform, Oakbrook can build summaries of information about customers they already know, and provide even more personalized decisions based on detailed individual data. With this capability, Oakbrook has been able to offer 10,000 of their customers the opportunity of a top-up loan at potentially lower lending rates. They’re forecasting a 5% revenue increase over the next 12 months.
From a platform perspective, the environment for data engineers and data scientists is now greatly improved, giving them instant access to lower-level data as needed. The open nature of the technology also means they can easily explore other Google Cloud features such as Looker, BigQuery or ML tools in the future.
Ed Ball, Head of Data, Security and Service Management at Oakbrook, summarized: “The lakehouse provides a great experience for our data engineers and data scientists, and by constant development of our data and analytics platform, we’ll continue to work towards our purpose to simplify and personalize borrowing and ensure better, more responsible outcomes for our customers.”
