en.Wedoany.com Reported - UK digital bank Monzo has redesigned its data warehouse, adopting a "mesh-style" approach to reduce warehouse costs by approximately 40% and increase data delivery speed by about 25%. The bank previously faced a complex situation with over 100 teams handling more than 12,000 dbt models.
Over the past year, Monzo rebuilt its data platform around defined modeling layers, explicitly declaring interface models for cross-team data dependencies, and enforcing validation of structure, naming, and access patterns through continuous integration (CI). This migration covered thousands of dbt models and introduced hundreds of managed interfaces, reducing redundant queries and duplicate computations, improving data landing times, and reversing the trend of rising warehouse costs. Each team owns and maintains its own data models, with Monzo supporting distributed ownership through automated guardrails and shared tools. Monzo's analytics engineers Antonia Badarau, Irina Mugford, and Massimo Frangiamore noted that while distributed ownership is powerful, it is difficult to implement correctly at scale, especially as AI-assisted coding becomes the norm, making it a challenge to ensure outputs remain efficient, consistent, and high-quality.
dbt models are SQL queries that transform raw data into structured datasets, designed as modular, reusable components for building and maintaining data pipelines. Monzo defined three principles for its data architecture: enforce clear standards, normalize data sharing through explicit interfaces, and rely on automation and CI checks to ensure quality rather than manual reviews. The bank divides data models into four layers: automated landing models (flattening raw events), generated normalized models (representing entities with complete history), logical models (combining entities with business logic), and presentation models customized for specific downstream uses.
Teams ensure consistency through the Modelgen command-line tool (which generates SQL and YAML models from object definitions) and CI-supported data standards (validating structure, conventions, and best practices). Luke Briscoe, Engineering Director at Monzo Bank, stated that scaling data in any rapidly growing organization is not easy, especially for a bank, and to his knowledge, few companies run similar tools. Mateusz Ulas, founder of Expeditious Software, commented that treating data interfaces as first-class code remains exceptionally rare, with most places relying on documentation and hoping for the best, while embedding standards into CI is what enables improvement.
Clear data layers, stable interfaces between datasets, and automated checks in CI maintain system consistency, allowing teams to work independently while reducing warehouse costs and processing times. Monzo enforces data quality and consistency by requiring each model to define a unique key, include freshness tests, run incrementally by default, declare its owning team, provide documentation, and adhere to strict naming and metadata conventions validated by CI.
Badarau, Mugford, and Frangiamore added that the company-wide migration is ongoing, with approximately 30% completed, and initial results are encouraging, showing around 40% cost reduction and about 25% faster data landing times in certain areas.
In another article, Monzo's engineering team described how they used a multi-task neural network to learn shared representations of fraud patterns, improving detection of rare and previously unseen behaviors beyond the capabilities of traditional models. At this year's QCon London conference, Suhail Patel demonstrated how Monzo built a developer platform capable of pushing hundreds of changes to production daily.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
