Why CFOs can’t afford to ignore Data Engineering in 2025

Data Engineering in a nutshell

You can only extract meaningful insights from your data when that data is clean, consistent, and combined on a centralized platform. That is exactly what data engineering delivers: capturing, processing, integrating, and preparing data from various source systems in a structured way so analysts can work with it using a reliable reporting data model.

“Running a few experiments in Excel doesn’t qualify as data engineering,” says Ivo Merchiers, Data Engineering Manager at TriFinance. “It needs to be professional, scalable, and durable, so processes run smoothly—and keep running smoothly over time.”

Data engineering isn’t reserved for large organizations. “Any company that wants to operate in a data-driven way needs to approach this professionally,” adds Maarten Lauwaert, Expert Practice Leader Data & Analytics at TriFinance. “Think of it as a spectrum, ranging from highly advanced platforms to low-code solutions within a unified data architecture.”

Data Engineering as the foundation for analytics and Data Science

Data engineering is the backbone of both data analytics and data science. Without a solid data warehouse or modern alternative, analysts and data scientists cannot perform at the level required to deliver real business value.

• Data analytics: primarily backward-looking and reporting-focused, delivering descriptive insights.
• Data science: more forward-looking, using advanced statistical methods and machine learning to generate predictive and prescriptive insights.

From Data Warehouse and Data Lake to the modern Data Lakehouse

So what exactly is a data platform? In 2025, it is often a data lakehouse, a system that merges the strengths of the traditional database (the data warehouse) with those of its successor (the data lake).

Here’s a brief historical overview, explained by expert Ivo Merchiers:

1. Companies initially stored their reporting data in a data warehouse. It worked well: structured data, clear organization, reliable outputs. But as data volumes grew quickly, two issues emerged: costs increased in parallel with the volume, and system performance hit its limits.
2. The next phase introduced the data lake. These platforms operate differently and, above all, more cost-efficiently because storage and compute are decoupled. Regardless of how much data you store, you only activate compute resources when you need to transform or analyze data. Big data became far more accessible. The drawback: reliability was not always guaranteed.
3. Data lakehouses represent the latest generation of data platforms, combining the strengths of both predecessors. They enable scalable, affordable, reliable, and accessible data processing, though solid IT expertise remains essential in critical roles. Because lakehouses effortlessly handle diverse data types (structured, unstructured, and sourced from across the entire organization) they are also ideal foundations for machine learning and other AI-driven applications.

Many large companies are investing heavily in this technology today. Three of the leading platforms at the moment are Microsoft Fabric, the Databricks Lakehouse Platform, and Snowflake.

Maarten Lauwaert also sees a major shift on the horizon:
“All data agents and other tools running on a large language model need to be fed with data. And for these solutions, a data lakehouse is exceptionally well suited to collect and organize that data.”

Data Strategy and Data Management as the framework for Data Engineering

You never build a data platform in a vacuum. Because it is a means rather than an end, the first step is to define a data strategy tailored to your organization. Where do you want to go as a company, and what role should data and data analytics play in that evolution? These questions need to be revisited regularly, as no strategy should ever be carved in stone.

Data engineering is therefore more than a technical exercise. While IT skills are undeniably required, such as connecting software systems, setting up automated processes, embedding control mechanisms, the work starts elsewhere. Business needs must first be translated into technical requirements. Only then can data engineers turn those requirements into action in a unified data architecture.

Another essential theme in any strategic exercise is data management. Which processes, agreements, and roles are needed to ensure smooth data flows and consistent data interpretation? Who is responsible for maintaining the systems and platforms? And how do you safeguard data security at all times?
These are considerations that should ideally be addressed early in the process.

Finance in the driver’s seat

When it comes to bridging the gap between strategic and technical profiles (or even between a CEO and a CTO) the CFO is uniquely positioned. Finance naturally combines business insight with a strong command of numbers and analytics.

“The CFO holds the key to initiating that much-needed dialogue,” says Ivo Merchiers. “It matters, because organizations need not only vertical depth but also horizontal integration: breaking down silos and aligning both data strategy and data engineering to serve the entire organization.”

Breaking down those silos remains a challenge for many companies, especially in the data domain. And that is problematic: when departments develop their own data platforms and tools without coordination, efficiency suffers. “The risk of creating siloed data and multiple versions of the truth becomes very real,” warns Maarten Lauwaert. “It’s a waste of money and talent.”

One constant remains: the greater the complexity of overlapping tools, the harder and more expensive it becomes to drive any technical transformation.

“Finance needs to be in the driver’s seat,” Maarten adds. He therefore encourages financial professionals to develop at least a basic level of technical expertise in data (engineering) skills. “Not to build platforms themselves, but to lead. When you help define the strategic direction and speak the language of data, you can effectively guide your organization and your data engineers.”

The Data Engineering team

What a data engineering team looks like depends on the size, data maturity, and, most importantly, the strategic direction of the organization. 

Deciding how far you want to go in data analytics and data platform development is a key strategic question. One essential reminder: becoming data-driven is a never-ending story. Every tool requires ongoing maintenance and continuous updates. From a certain scale onwards, having dedicated data engineers becomes indispensable. These specialists focus full-time on building and maintaining the data platform.

Team structure should not be confined to rigid boxes either. Maarten explains: “Data analysts and data scientists, who, thanks to the platform architecture, can largely work with the data independently, typically sit within the business team. It can be highly valuable to place part of the data engineering capacity there as well, precisely to safeguard business relevance.”

Ivo agrees and sees strong potential in domain-specific teams. “For example, in the Finance domain you can create a project group that brings together engineers, analysts, and other relevant profiles. Such a multidisciplinary setup within a well-defined domain is an ideal foundation for a thorough dialogue about concrete business needs and technical requirements.”

Read more

• Why Finance can be the engine of data-driven organizations
• Empowering controllers with data engineering skills