Project Turnaround (Mercedes)

A turnaround project is rarely easy. But our sustainable successes show that we can also solve complex issues.

A turnaround project is rarely easy. But our sustainable successes show that we can also solve complex issues. In this case, we were able to carry out an architecture audit at the Mercedes Car Group's digital factory:

Digital factories generate enormous amounts of data. This can certainly push the capacities of the infrastructure used to their limits. Despite comprehensive technical scaling, in this case neither nights nor weekends were enough to cope with the increasing system load.

We have experienced more than once that digital systems have reached their limits despite massive scaling, whether for technical or commercial reasons or a lack of user acceptance for response times.

Few users and high load are suspicious

All previous investigations and solutions were based on the assumption that the data volumes and computing processes were the cause of the enormous system load. Accordingly, technical scaling was carried out to the best of our knowledge and belief.

However, we had already seen similar scenarios in the TelCo industry - and identified completely different triggers there. After reviewing the infrastructure situation and consulting with IBM's technical experts, it was clear that the measures taken were all well engineered but ineffective.

Project turnaround through UX engineering

The project turnaround was achieved by pivoting the solution space. As the system operator had already pulled out all the stops technically, we analyzed user behavior instead of infrastructure.

Observing users during their typical operating steps quickly revealed the root of the problem: The user interface did not allow users to specifically filter the content relevant to them. The available filters were basically useful - but not for the working context of the users previously working on the system.

The cause therefore lay in the requirements analysis and inadequate UX engineering - with the result that users always had to request the entire data spectrum, even for the smallest data requirements, and only then laboriously isolate the relevant data section.

After the revision of the UX and filter concept, the system load was permanently reduced to values below 5%, so that the infrastructure with the associated economic aspects could be drastically reduced.

Outlook

The implemented solution has eliminated the cause of the scaling problems and allowed the necessary freedom to focus on the business case and functionally relevant content. 

Questions?

DE_DE