How Clip Transformed Its Lending Data Operations—From Snowflake to Databricks

Executive Summary

To support the rapid expansion of its lending services in partnership with Prestaclip, Clip faced the challenge of managing increasingly complex and fragmented loan data. Without a centralized platform, the finance, risk, and accounting teams were constrained by manual workflows and limited visibility into key metrics such as payment schedules and loan performance.

In response, Mutt Data partnered with Clip to transition from Snowflake to Databricks, designing and deploying a scalable modern data platform based on the Databricks Medallion Architecture. By structuring data into Bronze, Silver, and Gold layers, the solution delivered clean, unified, and business-ready insights. The platform included core data models for loan attributes, amortization schedules, and collections, and automated data ingestion and processing using Airflow and Databricks pipelines.

This modernization effort significantly reduced manual workloads and enabled real-time monitoring of loan performance and collections. With enhanced data granularity and automation, Clip is now equipped to scale its lending operations efficiently—empowering teams to make faster, data-driven decisions.

About the Company

Clip is the leading financial integral ecosystem of its country. It was born with the mission to help businesses have a platform for digital payments. It provides mobile payment services and allows companies and consumers to make transactions by turning their mobile devices into a card terminal. Clip serves customers worldwide. They promote the financial inclusion of people and companies through innovative & technologically trusted solutions, making it easy, accessible, and transparent.

Introduction

Clip had recently launched a loan service. This generated the need for a data platform to monitor loan performance and meet the needs of finance, risk, and accounting teams. The key outputs were data models designed to manage loan-related data, including attributes, collections, and amortization schedules.

The Challenge

Clip was growing. On their way to expanding their financial products portfolio to include loans, they faced significant challenges, in effectively managing and utilizing loan data.

Without a centralized, structured data platform, the company’s data was scattered across multiple sources, which made their monitoring of loan performance to be very time-consuming, and that actionable insights took time to reach internal teams.

Before partnering with Mutt Data, Clip managed its loan data using Snowflake. While Snowflake addressed basic reporting needs, it lacked the flexibility, orchestration capabilities, and real-time processing required to support a rapidly scaling lending operation. As loan volume increased, this fragmented environment led to delayed reporting cycles, manual reconciliations, and limited visibility into critical insights for the finance, risk, and accounting teams.

To address these challenges, Clip agreed to a discovery phase with Mutt Data , during which a comprehensive assessment of data sources and infrastructure was conducted to evaluate data quality and historical consistency. This deep dive revealed key limitations in Snowflake’s architecture—particularly in orchestration and scalability—as well as a notable rise in operational costs.

This analysis revealed the necessity of transitioning to a Data Lake architecture within a Databricks environment.

Based on these findings, Clip made the strategic decision to migrate to Databricks. The migration required careful planning, tool evaluation, and close collaboration between both teams. Over a four-month period, Mutt Data guided the implementation, ensuring the new platform was robust, scalable, and aligned with Clip’s long-term data strategy.

The benefits of Databrick’s lakehouse environments include:

• A single platform integrating data, analytics, and AI.
• Reliable, secure, and high-performance data handling.
• Open and secure data sharing without duplication.
• Robust support for data warehousing.
• Centralized governance for all data, analytics, and AI resources.
• Simplified real-time analytics, AI, and application development.
• Tools for building and deploying ML and Generative AI applications.

This endeavor required strategic guidance for planning and implementation, including the evaluation of optimal tools and components for each infrastructure segment. The migration spanned approximately four months and involved close collaboration between both teams.

An action plan was outlined, including the creation of a series of tables to capture the necessary information for reporting and analysis. This was collaboratively designed by members of Mutt Data and Clip’s teams. Mutt Data decided on a Medallion Architecture. A medallion architecture is a structured approach to organizing data in a lakehouse, designed to improve data quality and structure step by step. Data flows through three layers—Bronze, Silver, and Gold—each adding refinement and value. This method, also called a “multi-hop” architecture, ensures data is progressively prepared for analysis and business use.

The Solution

Mutt Data guided Clip through a strategic migration from Snowflake to Databricks, establishing a modern data platform. This transition enabled real-time data ingestion, transformation, and modeling at scale—consolidating these capabilities within a single, unified platform.

Leveraging tools like Airflow and Databricks for end-to-end automation across Bronze, Silver, and Gold layers, the new architecture unified previously fragmented datasets. It also unlocked a new level of data granularity and automation across finance, risk, and operations—empowering teams with faster, more actionable insights

Mutt Data designed a cutting-edge data platform using Databricks’ Medallion Architecture, which organizes data into three structured layers: The platform organizes loan data into three layers: Bronze (raw data), Silver (refined data), and Gold (business-ready insights). Mutt Data also built key models to capture essential loan details, expected and real payment schedules, and collection records.

Key Models Built

1. Loan Attributes: Captures all relevant details of loan creation and enrichment for analysis.
2. Original Amortization Schedule: unifies loans into a common system of automation to ensure consistent payment tracking. This also allows Clip to monitor Prestaclip success within clients.
3. Collections: Tracks payment histories, including refunds and chargebacks, with plans for bank reconciliation.
4. Operative Amortization Schedule: Provides a detailed financial history of loan amortization.

The entire data processing pipeline was also automated. This automation ensured reliable insights for Clip’s teams, enabling them to monitor loan performance and alignment with strategic goals. Moreover, the solution significantly reduced the time required by multiple teams to complete these tasks, effectively addressing the challenges of scaling analyses as the business expanded.

Avoiding recursivity

A notable challenge encountered was computing the necessary metrics for the operative amortization schedule model, as the formulas involved interdependencies (recursivity). This approach aims to optimize resource demands and processing times, ensuring that the system operates efficiently and meets performance expectations. Utilizing SQL window functions and reformulating certain metrics, the team successfully achieved the desired outcomes.

Testimonial

Impact

Despite being a new service, Clip’s loan offerings experienced rapid and exponential growth, reaching a point where tracking was unfeasible without a data platform providing real-time analytics and process automation.

From a process perspective, the accounting team previously calculated everything manually, which took days or weeks. With this solution, those numbers are now available in less than one day.

One of the most significant breakthroughs was refactoring their collections processes to maximize profits. Additionally, automating previously manual processes resulted in substantial time savings. The Modern Data Stack provided Clip with a level of granularity in loan information that was previously unattainable.

Results

Wrapping Up

Following the successful migration to Databricks, Clip’s financial team collaborated with Mutt Data to develop a data platform leveraging Databricks’ multi-layer architecture. This platform organizes loan data into three layers: Bronze (raw data), Silver (refined data), and Gold (business-ready insights), equipping Clip and Prestaclip with the tools to monitor loan performance, optimize collections, and support strategic growth in the lending market.

This solution not only centralized Clip’s data but also enabled them to monitor loan performance, optimize collections, and support strategic growth in the lending market.

Ready to unlock the full potential of your company? Schedule a call with us to discover how our expertise can elevate your capabilities.