Insight, Impact, Innovation
HomeOur WorkCase StudiesHow can I boost campaign response?

How can I boost campaign response?

Download PDF


Campaign analysis improves response rate by 30%

The Business Challenge

Our client had been conducting several direct mail cross-sell campaigns for its personal loan products to pre-approved customers from its existing credit card customer base. These campaigns included a follow-up phone call. Due to limited budget allocations, the frequency and size of these campaigns were constrained. The challenge was in targeting the right customers who would most likely respond favorably.

The card issuer retained Fractal Analytics' consulting services to manage its personal loans cross-sell program and achieve at least a 30% improvement in response rates. We were also required to arrive at the optimal price point (interest rate) for the personal loan offers that would maximize both revenues and profitability.

The Solution

Our analysts began by integrating data from four past independent campaigns conducted by the bank, focusing on the seamless extraction and consolidation of the data for further analysis. It included personal offer details, customer demographics, general transaction and relationship characteristics prior to campaign, and campaign response details. We ensured the consistency, validity and integrity of the consolidated data.

The data was analyzed to discover underlying customer behavior patterns, detect seasonality fluctuations, and identify key factors contributing to higher response rates. Our analytics team then used advanced analytical techniques to create predictive response models that would identify which customers would be most likely to respond to a personal loan offer.

During this stage, we also investigated the "spam" effect of direct mail and the impact on credit card balances of a given customer following the loan offer.

The Results

After implementing Fractal Analytics' models, our client reported significant results. As part of the validation process, they verified that the top 30% of respondents (based on probability scores) captured up to 76% of actual respondents to the offer. This meant the bank could potentially focus on only 30% of the target market and achieve over 100% improvement in response rates.

The same models were used to generate response probability scores on a new set of credit card customers for cross-selling personal loans. But by focusing on the top 70% of this list for the target market, the response rate was improved by over 30%.

Our solution gave our client the confidence to make critical decisions based on reliable analytics powered by its own customer data. Additional cross-sell and up-sell campaigns can use similar data management, modeling, and scoring techniques to maximize revenues.

Response Model Lift