Finding the best card for you - nerdwallet
The Context:
In response to user research, we built a feature within the native app that allowed us to recommend a credit card to a user based on 1) their spending 2) their credit score, and 3) cards they already carry. The backend decided which card is the “best” card based on the algorithm’s criteria and then presented it to the user with details on why.
The Challenge:
How might we reveal the complex logic that powers the personalized credit card recommendation in a clear, concise way without overwhelming the user?
The STRATEGY:
Create a hierarchical flow of information that goes from general to more complex
Lead with the benefits and most important information to optimize for scannability
Leverage questions and user inputs to build trust and create a sense of discovery and control
what’s happening on the backend
For each card a user linked, we calculated the rewards points they can can receive based on their current spending. Then, we scanned our marketplace for credit cards with higher rewards rates within the approval odds based on the user’s credit score and found the card with the highest delta in rewards points.
In the end, we can summarize the best card per category based on what they have:
Then run the same calculation for what they don’t have:
and finally compare both tables and see what creates the biggest improvement.