Want to know more?
Are you interested in this project? Or do you have one just like it? Get in touch. We’d love to tell you more about it.
Optimising credit card underwriting models
Global marketplace
Unlocking the potential of customer data for approvals increased approval rates for an online marketplace’s co-branded credit card by 50%, and drove revenue growth of over $23 million through incentives and rewards.
A major US-based online marketplace for home furnishings, with over $11 billion in annual revenue, recognised the potential value of launching its own credit card, co-branded with a major bank. At the same time as removing friction from the shopper journey, the co-branded card would enable it to offer incentives and rewards to drive repeat purchases among loyal customers. When acceptance rates for the card failed to match expectations, Equal Experts worked with the marketplace to unlock insights within its customer data that could accelerate risk analysis and approvals, and enable this multi-dimensional, revenue-driving role.
Outcomes
Growth
of customer base qualifying for card through pre-screening
50%
increase in credit card approval rate
$23 million
increase in co-branded card revenue
About the client
A global e-commerce leader in home goods, this company uses advanced technology to personalise shopping, manage logistics, and optimize inventory. Operating a vast supplier network through a dropship model, it delivers a seamless customer experience, showcasing how data-driven strategies and digital innovation can scale success in a competitive retail environment.
Industry
Retail and ecommerce
Organisation size
14,000+ employees
Location
Global, with headquarters in the US
Challenge
Low approval rates undermined card adoption and revenue impact
The co-branded credit card was designed to provide customers with flexible credit options while incentivising repeat purchases through rewards. However, low approval rates for the card limited its impact on customer adoption and revenue growth. The online marketplace worked with Equal Experts to develop its own predictive model, which could reliably identify high-quality, low-risk customers, and provide the banking partner with behavioural data to support applications. This significantly increased approval rates, enabled the bank to offer more competitive credit terms, and helped to unlock the card’s full potential for revenue growth.
Solution
Unlocking behavioural data to power smarter credit decisions
Equal Experts identified the key to growing approval rates and enabling more competitive financing in the online marketplace’s customer data. Analysis revealed that loyal customers of the store had significantly higher engagement and more favourable credit risk profiles than other customer segments. This established the potential for leveraging customer behaviour to qualify low-risk, high-quality customers.
In order to unlock this potential, Equal Experts needed to free data points such as length of time as a customer, order history and browsing history from the BigQuery data table that the marketplace used to aggregate customer data. BigQuery integrates seamlessly with the marketplace’s broader analytics infrastructure, but isn’t optimised for real-time data access, which made its approximately 600 million customer records inaccessible to the risk-profiling process.
The solution that Equal Experts adopted leveraged the speed and scale possible through Aerospike, the noSQL database that the marketplace already deployed for projects requiring low-latency data access. The flexibility of a non-relational database would allow the behavioural data to be analysed in real time, feeding into screening for the card and increasing approval rates.
Equal Experts built an Apache Airflow workflow on Google Cloud Platform Composer, which could synchronise daily customer data between BigQuery and Aerospike, making behavioural data accessible for the risk analysis process. This enabled the marketplace to access relevant customer data quickly, and include it in the credit application sent to its banking partner. The additional context and insight had a material impact on card approval rates, and therefore on the revenue contribution that the card could make.
Results
Approval rates up 50% – and $23M in new revenue
- The marketplace’s predictive model successfully identified high-quality, low-risk customers with a higher likelihood of credit approval
- Sharing relevant behavioural data with the banking partner increased approval rates and enabled more competitive financing terms
- This increased approval rates for the co-branded credit card by 50% and grew its revenue contribution by $23 million
Conclusion
By taking an innovative approach to unlocking the potential of its own data, the marketplace was able to significantly enhance the experience of its highest-value customers through the co-branded card. Besides flexible credit options and more competitive financing, it developed a platform for incentives and rewards that further increased customer loyalty among this segment, and multiplied the initiative’s revenue impact.