Credit Card Rewards

• 分析与建模 - 数据挖掘
• 分析与建模 - 预测分析
• 分析与建模 - 实时分析

• 金融与保险

• 商业运营
• 销售与市场营销

• 补货预测

• 数据科学服务
• 系统集成

The company provided a variety of different types of credit cards, and each card had different rewards programs designed to increase customer spend and engagement. While analysts did have access to advanced customer data, it was difficult for them to isolate the true causal relationship between different rewards offerings and customer spending. The company aimed to answer the following questions: Were rewards incentivizing customers to increase their card spending? Or were they merely subsidizing customers that would have reached various bonus thresholds of spending without incentives? Which cards were generating enough spend to justify the costs of reward redemptions? Which rewards were most effective with which customers? How could they target their offerings to maximize ROI? Amidst this uncertainty, the card provider wanted to introduce a new product to its portfolio. There was internal debate about which of two rewards programs would be more successful: a Double Rewards Points program, or a Cash Rewards offering. The programs were projected to have similar redemption costs, but the company needed to better understand the in-market impact of each offering on customer behavior before it could confidently choose between them.

The customer is a major card issuer that offers a variety of credit cards, each with different rewards programs aimed at increasing customer spend and engagement. The company has access to advanced customer data but faces challenges in isolating the true causal relationship between different rewards offerings and customer spending. The card issuer is looking to introduce a new product to its portfolio and is debating between two rewards programs: Double Rewards Points and Cash Rewards. The company needs to understand the in-market impact of each offering on customer behavior to make an informed decision.

The card provider used APT’s Test & Learn solution to build and run an in-market experiment on the two offers. The company selected a random group of test customers to offer each promotion to, and planned to assess their spending behavior compared to control customers that received neither offer. APT’s software, however, detected two sources of bias: Test customers were biased to have held their cards for shorter time periods, and test customers were biased to have higher sales in the prior year. APT’s software allowed company analysts to quickly adjust the control strategy, ensuring that test and control customers were well-matched in the period leading up to the test. The card provider was then able to analyze clean results of the test by comparing the spend behavior of test customers with that of control customers in the months after issuing the rewards offers. Test & Learn showed that the rewards offerings drove a 15% overall lift in customers actively using their cards. However, when segmenting performance, Double Rewards Points were proven to drive an 18% lift in sales per customer, while Cash Rewards only drove a 13% lift. Equipped with the finding that the Double Rewards Points were more effective, the company was then able to use APT’s software to dive deeper into the analysis and identify which customers were most receptive to the promotion. It found that the rewards program was most effective with the following customer segments: Customers that had spent more than $500 on the card in the past 6 months, customers that had opened their card more recently, and higher income customers. Using these insights, APT combined drivers of performance into a predictive model that identified the customers most likely to respond positively to Double Rewards Points.

• The test showed that the card provider’s originally planned rewards strategy would have been unprofitable.
• Using APT’s targeted strategy with Double Rewards Points, the client was able to generate positive ROI on the program.
• The company was able to identify which customer segments were most receptive to the Double Rewards Points promotion.

• The rewards offerings drove a 15% overall lift in customers actively using their cards.
• Double Rewards Points drove an 18% lift in sales per customer.
• Cash Rewards drove a 13% lift in sales per customer.