Challenge
The client aimed to build and scale a full-fledged marketing analytics platform around open-source Marketing Mix Modelling (MMM) frameworks. Their vision was to create a modern web application that would enable marketers and analysts to measure campaign performance, forecast ROI, and optimize budgets — all within a single, intuitive system.
The initial prototype, created by the client via ChatGPT, demonstrated the platform’s potential but required a complete architectural redesign, modularization, and performance optimization. The main challenges included:
Separating the existing monolithic system into independent front-end, back-end, and machine learning (ML) services;
Building a scalable data-processing pipeline capable of handling large datasets and multiple KPIs (sales, impressions, seasonality, spend);
Ensuring accurate and fast model fitting for marketing mix analysis;
Designing a simple yet professional user interface for data visualization and report generation;
Securing model management, authentication, and system administration for enterprise users.
The client aimed to build and scale a full-fledged marketing analytics platform around open-source Marketing Mix Modelling (MMM) frameworks. Their vision was to create a modern web application that would enable marketers and analysts to measure campaign performance, forecast ROI, and optimize budgets — all within a single, intuitive system.
The initial prototype, created by the client via ChatGPT, demonstrated the platform’s potential but required a complete architectural redesign, modularization, and performance optimization. The main challenges included:
Separating the existing monolithic system into independent front-end, back-end, and machine learning (ML) services;
Building a scalable data-processing pipeline capable of handling large datasets and multiple KPIs (sales, impressions, seasonality, spend);
Ensuring accurate and fast model fitting for marketing mix analysis;
Designing a simple yet professional user interface for data visualization and report generation;
Securing model management, authentication, and system administration for enterprise users.
Solution
The Jellyfish.tech team re-engineered the SIMBA platform using a modular architecture and the latest technologies in web development, machine learning, and data engineering.
Key steps and solutions implemented:
Data Aggregation: Utilized pandas for processing and merging data from multiple marketing and sales sources, combining KPIs, seasonality, and historical trends.
Architecture: Designed a modular system with separate front-end (React), back-end (Python/Flask), and ML components, ensuring scalability and maintainability.
Model Management: Integrated AWS S3 for secure storage and version control of models and datasets.
Budget Optimizer: Developed a forecasting function that predicts optimal marketing spend allocation across channels.
Asynchronous Model Fitting: Built a queue-based system for background model training, allowing multiple MMM calculations to run in parallel.
User Interface (UI): Delivered a clean, Tailwind CSS-based dashboard with easy data uploads, visualization panels, and real-time model feedback.
Authentication & Admin Panel: Added secure email authentication and admin functionality for user and project management.
APIs & Integrations: Developed custom REST APIs to facilitate seamless data flow between components.
Deployment: Deployed and maintained the solution in a production environment, ensuring high uptime and system reliability.
The next milestone involves transitioning to a microfrontend architecture for improved scalability and collaborative development.
The Jellyfish.tech team re-engineered the SIMBA platform using a modular architecture and the latest technologies in web development, machine learning, and data engineering.
Key steps and solutions implemented:
Data Aggregation: Utilized pandas for processing and merging data from multiple marketing and sales sources, combining KPIs, seasonality, and historical trends.
Architecture: Designed a modular system with separate front-end (React), back-end (Python/Flask), and ML components, ensuring scalability and maintainability.
Model Management: Integrated AWS S3 for secure storage and version control of models and datasets.
Budget Optimizer: Developed a forecasting function that predicts optimal marketing spend allocation across channels.
Asynchronous Model Fitting: Built a queue-based system for background model training, allowing multiple MMM calculations to run in parallel.
User Interface (UI): Delivered a clean, Tailwind CSS-based dashboard with easy data uploads, visualization panels, and real-time model feedback.
Authentication & Admin Panel: Added secure email authentication and admin functionality for user and project management.
APIs & Integrations: Developed custom REST APIs to facilitate seamless data flow between components.
Deployment: Deployed and maintained the solution in a production environment, ensuring high uptime and system reliability.
The next milestone involves transitioning to a microfrontend architecture for improved scalability and collaborative development.
Results
The outcome was a modern, flexible, and data-driven marketing analytics platform that simplifies complex MMM workflows and enhances decision-making.
Key achievements include:
Seamless data processing and visualization for marketing and sales teams;
Robust ML-driven insights enabling data-backed budget optimization;
Increased system reliability and reduced processing time through asynchronous model training;
A clean and intuitive UI that bridges data science and marketing usability;
Successful deployment of the platform used by analytics professionals for ongoing performance measurement.
The SIMBA platform, built and enhanced by Jellyfish.tech, positions 1749.io as a cutting-edge solution provider in probabilistic marketing analytics, combining Python’s analytical power with React’s interactivity and modern UX design.
The outcome was a modern, flexible, and data-driven marketing analytics platform that simplifies complex MMM workflows and enhances decision-making.
Key achievements include:
Seamless data processing and visualization for marketing and sales teams;
Robust ML-driven insights enabling data-backed budget optimization;
Increased system reliability and reduced processing time through asynchronous model training;
A clean and intuitive UI that bridges data science and marketing usability;
Successful deployment of the platform used by analytics professionals for ongoing performance measurement.
The SIMBA platform, built and enhanced by Jellyfish.tech, positions 1749.io as a cutting-edge solution provider in probabilistic marketing analytics, combining Python’s analytical power with React’s interactivity and modern UX design.
