2. Case Study - PM and UX Collaboration to Improve Loan Processing at Creditas

In this document, we will explore a real case from Creditas, a Brazilian fintech focused on lowering interest rates on loans by leveraging vehicle and real estate collateral. The case demonstrates how Product Managers (PM) and UX Designers worked together to address a specific challenge—increasing the loan processing rate. We'll break down the process, the interaction between teams, and the methods they used to optimize performance.

1. Understanding the Loan Request Process at Creditas

At Creditas, the loan request process involves multiple steps:

1. Customer Application: The customer visits the site, requests a loan, and completes a quick registration.
2. Pre-Qualification: An internal engine runs rules to determine if the customer qualifies for further analysis.
3. Offer Generation: Based on the customer profile and risk assessment, tailored loan offers are presented to the customer.
4. Data Collection: Complementary data needed for a deeper credit analysis is requested.
5. Authorization: The customer provides a digital signature allowing Creditas to check their data with the Central Bank of Brazil (Bacen).
6. Credit Analysis: A more robust analysis determines if the loan can move forward.
7. Formalization: This involves document submission, property or vehicle appraisal, contract signing, and other formal steps before loan disbursement.

2. Defining the Challenge: Improving Processing Rates

At Creditas, the processing rate measures how many customers proceed from pre-qualification to a full credit analysis. Two types of processing occur:

• Automatic Processing: The customer completes the steps independently through the site without manual intervention.
• Manual Processing: The customer needs assistance from the sales team, often due to drop-offs or confusion.

The objective was to increase the automatic processing rate, reducing the need for manual intervention.

3. Discovery Phase: Decomposing the Problem

In the discovery phase, the PM and UX teams worked together to identify opportunities for improving processing rates. Instead of focusing only on the user interface (UI), they took a broader approach, decomposing the processing funnel into smaller, manageable segments. They broke down the automatic processing rates into several conversion metrics:

1. Conversion from Offer Screen: Percentage of customers who choose an offer.
2. Conversion from Form Completion: Percentage of customers who fill out the required data after selecting an offer.
3. Conversion from Bacen Authorization: Percentage of customers who authorize data checks.
4. Profiles with Acceptable Risk: Customers who fit into the "low-risk" category and can be automatically processed for credit analysis.

Each of these segments presented potential areas for improvement.

4. Generating Solutions: Cross-Functional Brainstorming

The next step was brainstorming solutions across multiple dimensions—interface, technology, operations, credit modeling, and others. The PM and UX teams involved other stakeholders like credit and operations specialists to find holistic solutions. Below are some of the proposals for different stages of the funnel:

4.1. Offer Screen Conversion

• Hypothesis: The current offer screen may confuse users.
  • UX Solution: Test alternative designs, like adding calculators or changing the layout.
  • Tech Solution: Improve the performance of offer calculations.
  • Credit Modeling: Test if the displayed interest rates are too high for users.

4.2. Form Completion

• Hypothesis: Too many fields may deter users from completing the form.
  • UX Solution: Reduce the number of fields, add tooltips to explain fields, or break the form into multiple steps.
  • Tech Solution: Automate some of the form-filling processes.
  • Operations/Credit: Determine if all fields are essential or if some can be removed without compromising credit analysis.

4.3. Bacen Authorization

• Hypothesis: The digital signature may be causing drop-offs.
  • UX Solution: Test simpler authorization methods, such as checkboxes.
  • Credit Solution: Evaluate if a less complex authorization method would still comply with regulations.

4.4. Expanding Profiles with Acceptable Risk

• Hypothesis: Accepting a broader range of risk profiles could increase the number of customers reaching credit analysis.
  • Modeling: Consider adjusting interest rates for higher-risk customers.
  • Credit: Negotiate with external funding sources to determine if broader risk profiles could be accepted.

5. Prioritization: Mitigating Risks

Before moving forward with any solution, the team worked to mitigate risks. They identified four key risk areas:

1. Value Proposition Risk: Will users value this solution?
2. Usability Risk: Can users easily understand and use the solution?
3. Technical Feasibility: Can we build this solution efficiently?
4. Business Viability: Will this solution align with business goals, and is it allowed by regulations?

For each solution, they assessed the risk level in these categories. For example:

• Breaking the Form into Steps:
  • Value Proposition Risk: Low, as it simplifies the process for users.
  • Usability Risk: Low, based on previous successful tests.
  • Technical Feasibility: Low, as it's a simple modification.
  • Business Viability: Low, no regulatory issues.

On the other hand, offering a discount on one of the fees involved more risk, particularly technical and business viability risks. The technical challenge was that applying the discount required recalculating the entire fee structure, which could be a complex change.

6. Rapid Prototyping and Testing

Once risks were identified, the team focused on rapid prototyping and testing. Some key strategies included:

• Interface Prototypes: Testing various UI designs with users to gather feedback on usability.
• Technical Prototypes (PoCs): Building quick technical proofs of concept (PoCs) to validate feasibility without committing too many resources.
• A/B Testing: Running A/B tests on key solutions like offering discounts to gauge their impact on conversion rates.
• Collaboration with Tech: UX designers worked closely with developers, bringing them into user tests to create empathy and ownership over the product.

7. Hacking Solutions to Mitigate Risks

In situations where technical feasibility was a concern, such as with the discount solution, the team explored "hacks" to test user reactions without making major changes to the system. For example, they displayed the discount on the front-end but handled the actual application manually in the operations team, rather than changing the back-end immediately.

This allowed them to test the effectiveness of the discount without the complexity of rebuilding the loan calculator.

8. Roles and Responsibilities in Risk Mitigation

In each phase, different team members were responsible for mitigating specific risks:

• Value Proposition Risk: Both PM and UX were responsible, using tools like prototypes, MVPs, and user interviews.
• Usability Risk: Primarily UX-led, through testing and user research.
• Technical Feasibility: Led by the tech lead, using PoCs and performance tests.
• Business Viability: PM-led, involving legal and financial stakeholders to ensure compliance with regulations.

Conclusion

This case study from Creditas highlights the importance of collaboration between PMs and UX designers. By involving cross-functional teams in the discovery process and focusing on risk mitigation early, the company was able to identify multiple opportunities to improve the automatic processing rate. The approach of testing and validating solutions across various dimensions helped prevent wasted resources and ensured that the final product met both user and business needs.

This method of broad discovery, collaborative brainstorming, and incremental testing is a valuable approach for any company looking to optimize its processes and deliver better user experiences.