Reporting Solution Refactoring | Data Warehouse & ELT

How We Refactored an Inefficient Reporting Solution

The implementation of a data warehouse, ELT process, Customer Portal, and Reports for a US-based messaging platform, reducing development time by 58%.

Automation

United States

Data Engineering

Backend Development

[ Client ]

About the Client

The client is a US-based developer of a messaging platform. They had a reporting system in place, but it was implemented without a proper engineering approach: features were built in isolation, a mix of cloud and on-premise environments made the codebase impossible to navigate, and documentation was entirely missing.

400Gb

Data Warehouse Size Managed

40%

Reduction in Design Development

58%

Reduction in Total Dev Time

[ Challenge ]

A Reporting System Built Without Engineering Discipline

The client had a reporting system for their messenger platform, but the previous team implemented it without a proper engineering approach:

Each feature was built in isolation with no coherent architecture.
A mix of cloud and on-premise environments plus multiple languages (Python, Perl, SQL, Bash) made locating specific features impossible.
Documentation was entirely missing.

The data warehouse held 400Gb of data. The ETL ran once a day processing 1Gb across 150 tenants, generating 3,000 reports emailed daily.

[ Solution ]

Building the Solution

Our team analyzed the solution and improvement requests, designed a new architecture, and selected scalable, cost-effective tools. We developed the data warehouse, ETL process, Customer Portal, and Reports from scratch, then enabled direct access to data in the multi-tenant data warehouse.

Analysis: Documented the former solution's version and all improvement requirements.
Proposal: Prepared a new solution proposal covering architecture, mailing features, and software tooling.
Transition: Implemented the project and made a smooth transition from the old solution to the new one.

[ Results ]

Impact that Matters

The team delivered a simple, well-documented solution ready to scale. The new version only charges during active processing rather than when idle, making it cost-effective.

58% faster delivery

Total development time was reduced by 58% compared to the original system, thanks to a coherent architecture and clearly defined tooling from the start.

40% faster design dev

Web app design development was reduced by 40% through reusable components and a well-structured design system applied across the Customer Portal.

Cost-efficient and scalable

The new solution is well-documented and scalable, only charging during active processing rather than when idle, giving the client meaningful infrastructure cost savings.

[ Tech Stack ]