How we started mozaikoo.

Our research included complex geodata analysis, shortest-path calculations in multi-parametric dependent data, machine learning for signal improvement, and the combination of several types of neural networks to extract chronological context from video with semantic descriptions.

Shortly after ingesting huge amounts of data for analysis, we realised that we needed a monitoring, filtering and analysis system to identify deviations and clusters quickly. We also needed an administration and operating system with a user interface to organise data and analysis results.

At that point, we thought 'Why not bring both worlds together?' A visual analysis platform combined with a content management system, properly configured for different use cases and seamlessly interlinked. We wanted something dynamic enough that much of the recurring work required to build an appropriate software stack for our customers would already be done. Fortunately, we made the right decisions when choosing our software stack and shaping the architecture. Those decisions, made 15 years ago, still hold up today.

From the beginning, our architecture separated front ends, back end and API: a so-called headless system. We use triple-store graph databases and Elasticsearch indices. When we started working with the Elastic stack, it was still in beta at version 0.7. The databases accessed by our API also sit behind an abstraction layer, allowing us to work with document stores, graph databases or SQL engines.

During our research projects, when we showed the software to customers, we were repeatedly asked whether we could provide these features in ways that matched their business cases. In response, we continuously developed the operating system to cover almost every communication requirement and to become a true digital hub for communication. Our deep knowledge of data analysis, structural coherence, communication and marketing enabled us to build a system with a data foundation that still fits current needs almost out of the box.

Over the last 10 years, we released four major versions: mozaikoo 1, 2, 3 and 3.5. Strictly speaking, mozaikoo 3.5 was not a full major version, as the numbering suggests, because it was a rework and refactoring effort that modernised the stack and harmonised the data entities. But because so much of it was rebuilt from the ground up, it counts as a major version for us.

Our main focus throughout development was to remove pain points and build an interface that makes every part of communication easier. Our now two decades of experience in distributing information and building marketing reach helped us address this. Customer feedback continues to show that the efficiency and performance gains are enormous compared with the software used before.

With m4^®, we began a full refactoring process, changing the framework used for our front-end and administration apps and deeply integrating leading-edge autonomous and automated workflows to increase efficiency even further.

m4^® marks a new era in how content and communication can be delivered effectively to the addressee.