Data Integration in 2022. What makes it so important and what do clients get out of it?

Mihai: We are living in a world where all industries have digitalized, everything resumes to businesses and processes. Taking this into account and having more and more functions to manage inside your business, all the stakeholders need to have all this data as aggregated as they can in order to take the right decision and steer the business on the right path of development and success.

Having these challenges, the concept of data integration comes in, trying to tackle all of these by bringing all the data together in one environment/tool/database ready to be read, transformed, and delivered to stakeholders for supporting them in their decision-making process.

When you think about data integration, how would you explain the kind of consulting services you deliver to clients?

Mihai: The subject “Data integration” is part of any implementation we are doing, whatever the scope of the project is. Starting from the design of the project we need to answer the question “What data we need to bring into the tool?” like for instance financial data and master data. Having an answer to this question our consulting services consist of analyzing any gaps we have between applications, the type of transformation that needs to be applied to the data, what the best approach for transforming the data is (at source or in the target application) and finding the best solution in creating an automatic process that keeps the data aligned between all databases part of the process.

What is the typical use case of data integration software at your clients?

Mihai: Depending on each project there are two main topics for data integration. We have the operational/financial data integration which I would call it first wave of integration and later during the development, we have wave two, which consists of integrating the master data.

In most of cases there is a data integration software that transfers the data required between two or more environments. Often the biggest part of the transformation that is covering the gaps between the two systems is performed during the transfer. We recommend this approach to our clients to try to make the file as clean as we can when arriving in the target system. We can still have transformation performed in the target system, but majority of these are process related such as rules and logics for the process, different table structure of the system and so forth.

Zooming in on a client case, what anecdotes can you share? What kind of challenges does Data Integration imply?

Mihai: In general, each project comes with different challenges but the characteristics of them are similar so in these cases you are delivering based on experience and the path to a successful result has not so many bumps.

In the past instead, we had one client for which the context of the project was a bit different, in the lifecycle of the project was included also a change in data taxonomy, meaning, the new tool was evolving very fast to this new architecture while at the same time we were trying to keep the old systems integrated in the process until the end of the digital transformation.

All this context translated in a cyclical data flow, we were transferring data from System A (old tool) to System B (new tool) and back to System A. In this flow the data needed to be transformed twice: from old to new, and back to old again The challenge we were trying to overcome was to not overload the new tool with all the obsolete master data that was not going to be used, so we were looking to keep it as clean as we could. In this particular case we needed to think outside the box and we achieved a good data flow by keeping relevant data (needed for reverting to the old data model) stored in the new system but to not interfere with the process (in order to be able to drop it anytime we were finished with it without affecting system tables and processes).

By using this approach, we were able to achieve a successful integration on both data flows (financial and master data) having all systems aligned between them by creating scheduled activities.

What is your general advice to clients who want to implement a data integration software?

Mihai: For a successful implementation of a data integration software the most important are people. You need that person that understands how to read the data and designs the outcome, so this person knows what the relevant data for a successful project is. At the same time, you always need to be part of the development someone from the Data Governance team who can support in understanding all the relationships and all relevant data model topics. In the end this person is in touch with most of the business functions and has a big picture of what is happening from a data model perspective. These people in my opinion are essential to achieve your goals.

From a technical perspective, there will always be challenges in aligning databases and tables from system to system. Having all these people working together we can always find a solution to make our client’s life easier.