Cash App Credit Score Pilot

Helping Cash App users build financial access beyond traditional credit scores



Summary

Role: Lead Product Designer

Timeline: 4 months

Team: PM, Data Science, ML Engineers, Lending Engineering, User Research

Responsibilities:

  • Product strategy

  • UX vision

  • Interaction design

  • Experiment planning

  • Cross-functional alignment

Challenge: Millions of Cash App customers had little or no credit history despite consistently repaying loans. Traditional credit bureaus underestimated these users, limiting access to lending products.

Outcome:

  • Shipped pilot to hundreds of users

  • Validated demand for alternative credit scoring

  • Generated learnings that informed the GA launch

  • Influenced how transparency and education were integrated into future iterations

background

Cash App offers a few lending products to eligible users, such as Afterpay. Research uncovered that a lot of Cash App users, who had thin files or no credit history, were unsurprisingly scored poorly by the traditional credit bureaus. Even so, these people showed some of the highest repayment rates in the Cash App ecosystem. We had the data to show that these users are financially responsible and trustworthy, but the traditional credit system was working against them.

We identified a strategic opportunity to unlock lending for a population underserved by traditional credit systems. The design challenge wasn't simply presenting a new score—it was helping users understand, trust, and improve a credit model they had never encountered before. Thus, the Cash App Score (CAS) was created.

Read more about the research in Block’s whitepaper.

Scope

The pilot was a stripped down version due to resourcing constraints, a short timeline, and the desire to test the concept on a few hundred users first. The important aspects were viewing your current score, seeing the weekly change, recommended actions to boost your score, and the lending amount unlocked based on pre-determined score tiers. As a result of these constraints, some important decisions were made to reduce scope and scope creep:

  • We intentionally surfaced weekly deltas rather than a live score because weekly updates better matched the underlying model, which had infrastructure limitations for the pilot, while reducing confusion when scores appeared unchanged day-to-day. As a result, it was vital that we communicated that through the experience when a user was actively boosting their score.

  • The recommended actions surfaced to each user to help boost their individual scores was limited to one lending product in the Cash App ecosystem due to legal constraints and the desire to test the score first before adding more lending products. Non-lending actions were not limited, but each week only presented three actions total to not overwhelm users and reduce the feeling of a to-do checklist.

  • The data visualization of the CAS was kept simple. We wanted to stay away from the typical visualization of scores (e.g. gauge, thermometer) to differentiate the CAS but also didn’t want to spend too much time creating an original visualization and animation for the pilot. This score visualization was another opportunity to differentiate the CAS from traditional credit scores, but we needed to balance effort with novelty.

Design Principles

Transparency over precision: Users needed confidence more than mathematical detail.

Actionability over reporting: Every score should suggest the next best action.

Progress over perfection: Celebrate improvements rather than emphasizing low starting scores.


success metrics

Primary

  • Weekly active users

  • Repeat visits

  • Recommended actions completed

Secondary

  • Lending product eligibility

  • Borrow conversion

  • User understanding

  • Support contacts

Because this was an early pilot, the team focused primarily on behavioral signals rather than long-term lending outcomes.

lessons learned

After the pilot concluded, the UX team conducted several user interviews. We discovered that:

  • the Cash App credit score was well received by users who are familiar with and consistently use Cash App’s lending products

    • this audience leaned slightly towards the younger crowd (under 30 years old)

  • 1/3 of participating users thought the Cash App credit score was the same as their FICO score

  • several users desired a deeper understanding of how their score was calculated

  • users with very high scores were less incentivized to interact with the new feature

This valuable feedback was incorporated into the next iteration of designs for the GA release the following year. What changed because of this pilot:

  • Added clearer education distinguishing the CAS from FICO

  • Increased transparency into score movement

  • Expanded explanation of contributing behaviors

  • Improved incentive model for high-score users

Personal reflection:

This project fundamentally changed how I think about trust in AI-driven financial products. Accuracy alone isn't enough. Users need to understand why a system behaves the way it does, and what they can do next. That insight influenced not only this pilot but also subsequent work on the general availability release.