Working title · Ongoing research and development
Why I’m building this
Marketing platforms generate vast amounts of data, but access to information does not guarantee understanding. Campaign performance can be measured and visualized in countless ways, yet turning those measurements into clear decisions remains difficult.
I’m building Marketing OS to explore that gap through software.
Rather than studying marketing technology only through existing tools, I’m developing a system that examines how advertising data can be structured, interpreted, and transformed into useful insight.
The project is both a software product and an applied research process.
Research through development
Building the system is a way to investigate how digital marketing tools work and where they fall short.
The development process raises practical questions:
- How should advertising data be structured?
- Which metrics are genuinely useful for evaluating performance?
- What parts of analysis can be automated?
- Where should human judgment remain central?
These questions shape the architecture, interface, and features.
Research and implementation inform each other. Technical constraints can reveal new perspectives on marketing problems, while marketing questions can lead to different technical decisions.
Building the system
The current technical direction combines React and TypeScript on the frontend with Python and Flask on the backend.
The system uses PostgreSQL, SQLAlchemy, and Alembic for data management, with integrations for advertising platforms such as Meta Ads.
I’m exploring how external advertising data can be collected, normalized, and presented in a way that supports meaningful analysis. This includes API integration, database design, account separation, reliability, and the relationship between backend processing and frontend experience.
I’m also investigating how AI can support the interpretation of structured marketing data.
AI-generated insights should not be treated as unquestionable conclusions. They should be connected to verifiable data, presented with context, and evaluated for usefulness, accuracy, and reliability.
The interface as part of the research
A dashboard does more than display metrics. Its design influences how people interpret performance and decide what to do next.
The choice of what to show, emphasize, or omit can shape the meaning of a campaign’s results.
I’m interested in designing interfaces that make complex information accessible without reducing it to misleading summaries. This brings software engineering, visual communication, and marketing research together.
The interface is not merely the final layer of the system. It is part of the investigation.
What I’m investigating
The project focuses on several connected questions:
- How can marketing data from different sources be organized into a consistent system?
- What makes automated analysis genuinely useful?
- How can AI-assisted recommendations remain transparent and evidence-based?
- How does interface design influence the interpretation of marketing performance?
- What technical decisions are required to support multiple users or organizations?
These questions will be explored through experimentation, implementation, and revision rather than treated as fixed assumptions.
An evolving project
Marketing OS does not yet have a fixed final form.
I have a technical direction and a set of problems I want to investigate, but the research should shape the product as it develops. Some ideas will become features; others may be revised or abandoned.
I will document the technical challenges, design decisions, experiments, and changes in direction along the way.
For me, this project is a way to understand marketing technology by building it, testing its assumptions, and learning where the work leads.