Solution

Based on requirements, a marketing planning platform comprising two components was developed:

• Firstly, standardized software is purchased and used as a frontend component.
• Secondly, a separate cloud project is being developed as a backend component which is responsible for technical automation of data flow and can thus be tailored very flexibly to OTTO's use cases.

With b.telligent's support, OTTO opted for the MRM software Marmind, in combination with a project in the Google Cloud for development and connection of interfaces. With BigQuery and dbt, data from Marmind are stored and prepared for further use.

1st component: Marmind

Marmind serves as a front-end component in which users enter campaign briefings via forms, coordinate their tasks and processes via checklists, control approval processes via technical workflows, and represent budget allocation. A plan view in Marmind provides a comprehensive marketing calendar, and bookmarks can be used to display various planning reports.

2nd component: Google Cloud project

A Google Cloud project was developed in parallel to ensure that Marmind does not exist as an isolated system but can display data and information from other systems and make them available to users. Marmind is integrated into OTTO's system landscape via this cloud project. Via the GCP project, campaign-planning data are extracted from Marmind using the GraphQL API and can thus be forwarded to campaign-management systems via API interfaces. At the same time, the project also serves as a gateway for OTTO's other internal and external systems, which make data and information available to Marmind. These are received in the GCP project, transformed, and sent to the Marmind GraphQL API.

Project Schedule

The two components, GCP and Marmind, were conceived and developed to ensure a seamless transition between the old system and the new marketing planning platform. For the configuration of Marmind, an inventory and analysis of existing forms, workflows, and budgeting dialogs were first carried out to eliminate outdated topics during the new implementation. This was followed by the development of the context structure in Marmind based on technical requirements, as well as the configuration of forms, checklists, workflows, and budget planning. On the technical side, migration scripts were written to map historical data to the new forms and structures in Marmind. A GCP project was set up to establish a connection to the Marmind GraphQL API and enable the transfer of historical data to Marmind.

With the help of services such as API Gateway, Cloud Function, PubSub, and BigQuery with dbt, processes were developed to retrieve data from Marmind, prepare it for further use, receive data from other systems, and send it to Marmind. Incoming data can be displayed in real time in Marmind via a serverless infrastructure. To supply interfaces such as campaign management systems with data, a separate small data warehouse was created for Marmind data, using BigQuery as a database and dbt for structuring and loading. Data retrieval from Marmind is only possible in batch processes.

Thanks to optimized process orchestration using Google Workflows, data can be updated every 15 minutes. Technical and operational errors in data processing are monitored via cloud logging and made visible to the product owner team in MS Teams. The launch of the new platform was planned as a hard launch with a short downtime, during which the final migration of historical data took place.