Cabonline Finland

The client was Cabonline, one of the largest transportation service providers in the Nordic region. The project is a good example of how rapid iteration and a client open to experimentation can deliver significant results in a short time.

Context

The background was a Swedish company that had expanded into the Finnish market by acquiring three long-established local operators. In Finland, its brands include Kovanen and Fixu Taxi.

Each of the acquired companies had its own way of operating. Individual companies also lacked the resources to carry out larger IT system initiatives on their own. In particular, financial administration processes had fallen behind in digitalization, and daily work was inefficient due to differences between companies and a lack of proper tools.

At the same time, changes in taxi regulation were increasing the pressure to further improve operational efficiency.

Role and responsibility

I was the sole developer of the system and responsible for the entire solution end-to-end: integrations, data model, user interfaces, and system operation. The work was carried out on-site in close collaboration with the client’s organization.

Some requirements had already been identified by management. Once the essential functionality was in place, the system could also be used for more exploratory and innovative use cases.

Technical implementation

The system was built as a custom solution on top of the JavaScript ecosystem with a PostgreSQL database.

The main challenge was that data was distributed across multiple independent systems, some of which did not provide any APIs. While some integrations were built against modern interfaces, others had to be implemented by reading directly from files. In some cases, the data produced by older devices was unreliable or even corrupted.

The system integrated sources such as taximeters, financial systems, registers, and even a machine handling physical cash.

Outcome and learnings

During the project, a total of nine integrations were implemented, nearly all using different protocols. Processes were significantly harmonized and automated, reducing manual work and improving data quality. The system automatically highlights anomalies that would otherwise be time-consuming to investigate — or in the worst case, go unnoticed.

The project progressed through rapid, iterative steps — with direction continuously refined based on real-world use.

The entire solution was built in approximately six months, demonstrating that even for a mid-sized company, a custom-built system can be a justified investment that delivers value quickly.