Are you getting too little ROI on investments in your data infrastructure?

Blog series | From Data to Business Value with Data Mesh | #1

Authors: Alexander Kern & Yu Li

That's great. Then you're in good company. Many companies have a very loose purse when it comes to data. This is understandable, since it is a strategic goal for many to develop into a "data-driven" company as quickly as possible. So far, however, most complain about not being able to generate the desired benefit [1][2]. The necessary end-to-end agility - from data generation to use in new, innovative solutions - is missing in order to anchor data in the company as desired.


Data is moving to the center of corporate and IT strategy. In many industries, they are seen as a beacon of hope for gaining unique selling points. Especially in saturated markets, where differentiation from the customer's point of view can hardly take place via price, but is significantly controlled via the customer experience (e.g. banking, insurance, etc.). From product design, to every touchpoint with the customer, to increasing efficiency in service delivery - the goal is to use analytics, machine learning and AI to gain a competitive advantage. And the basis for this is data.

The architecture of our data infrastructure has changed little in the last 60 years.

For decades, we built DWHs, which was sufficient for classic use cases. To give analytics, ML and AI access to raw data, we have been building data lakes for a decade. For better scaling, we've been doing it almost exclusively in the cloud in the meantime. Despite this evolution, we've stayed true to the basics: Data flows from data production (e.g., by business applications) through ingestion and preparation (in the DWH or data lake) to the consumer. To make our work more efficient, we organize ourselves by components. Thus, as a rule, know-how carriers of the specialist application, data preparation and consumerization are each in independent teams. In addition, central modeling creates a bottleneck in classic DWHs.

As a rule, we organize ourselves into silos.

So far, we have put up with these silos - despite the associated disadvantages such as:

  • the time-consuming hand-overs 
  • the disconnect and loss in know-how 
  • the local optimizations without an end-to-end view,
  • and as well as the slow development cycles - from the connection of new data, the preparation to the use in a new solution. 

Not to mention the resulting bottlenecks in the respective silos.

And where is the benefit?

As a reminder, companies hope to gain a competitive advantage through data and innovations built on it. It is essential to get from idea to solution as quickly as possible. With the current setup, this is hardly possible.

Customer examples.png
Figure 1: Customer examples

Don't worry. There is a concept to add agility to the data infrastructure.

It is called «Data Mesh» [3]. The good news: For once, you don't have to license a new product. The bad news is that it still involves effort. Because «Data Mesh» realigns the way data is handled, which has far-reaching consequences. But this effort pays off in every case.

We lead you to ROI!

In the following blogs, we will first take a closer look at the fundamental aspects of Data Mesh (Blog #2). Then we will take a trip into the technical implementation (Blog #3) and what experiences we have had with our customers (Blog #4). And always with the goal of showing you how to use data for your competitive advantage. And last but not least, how you can finally get ROI on your investments.


[1] Big Data and AI Executive Survey 2021 

[2] Market Guide for Analytics Query Accelerators (Gartner)

[3] «Data Mesh» by Zhamak Dehghani (O'Reilly Media)