Despite years of investment in analytics, many organizations still face a frustrating reality. Data is abundant, yet leaders often hesitate to act because the insights do not feel clear, trusted, or accessible. The gap between data and confident decision-making has not disappeared. In many ways, it has widened as data volumes, systems, and expectations continue to grow.

In this first installment of a three-part series, Mars leaders Ujjwal Sehgal, Global Head of People Analytics, and Rachel Belino, HR Data Officer, speak with Shachin Prabhat, Vice President and Business Unit Head at Tiger Analytics, about why this gap persists and what it takes to close it in practice. Drawing on experiences from leading people analytics and the People & Organization Data Office at Mars, the conversation explores how unclear problem framing, fragmented data ecosystems, and limited data reusability continue to hinder the path from data to decisions, and how Mars is working to address these challenges.

Mars operates in more than 80 countries and is home to iconic brands across pet care, food, and confectionery. With over 150,000 associates globally, people data plays a critical role in workforce planning, talent strategy, and organizational decision-making. Yet even at this scale, the challenge of turning data into trusted insight remains very real.

Why the data-to-insight gap still exists

Organizations today often assume that having more data automatically leads to better decisions. In practice, the opposite can happen. Sehgal explains that the root cause of the gap lies in both problem definition and data readiness. “I look at it from two different angles. One is getting the basics right. Have we clearly defined the business problem? Do we clearly know the hypothesis we are going to use to solve this problem? And then, do I have the relevant data?”

He points out that the widespread belief that enterprises already possess all the data they need is misleading. “People talk about data lakes. Many times, those data lakes are not lakes anymore; they’ve become swamps. There’s irrelevant data sitting there, which makes it very difficult to get the true data needed to solve the problem.”

The issue is not just technical. It is also human and organizational. “Some of it is user-led, where they are not able to clearly define what they’re looking for and what the business problem is. Also, the data may not exist, or it might take time to stitch that data together because it’s sitting in different systems.”

This combination creates what Sehgal calls a “mishmash of both” that slows the journey from data to action.

However, he sees progress emerging across the industry. “Data scientists and data engineers are becoming true business personas. They are able to understand and help business users frame the problem. It’s a journey. I don’t think we are there yet, but the right steps have been taken.”

The early lessons from scaling people analytics at Mars

For Mars, solving the data-to-insight challenge meant rethinking how data was created and managed across the people analytics function. Belino recalls an early phase when the team moved quickly to deliver insights but unintentionally created long-term complexity.

“We were trying to deliver value very quickly. We had data scientists bringing the data in and also playing the role of data engineers. The data was very specific for their product.”

This rapid innovation created unintended fragmentation. “Before we knew it, one product had this data, another product wanted to use that data, but it couldn’t be reused. Think of that 50 to 100 times over, and you’ve got data that is somewhat unusable.”

The realization led the organization to rethink the entire foundation of its analytics ecosystem. “It boils down to blueprinting what a scalable and sustainable ecosystem looks like for data as an asset and how it can be used along the entire analytics product lifecycle.”

Building a scalable data engineering foundation

To move forward, the organization established a more structured approach to data engineering and governance. “We set up our data engineering processes so they are followed by our teams. We aligned the right capabilities and roles to bring in data from ingestion and transform it all the way through,” says Belino.

This shift helped unlock data that had previously been difficult to reuse. However, structure alone was not enough. Governance became equally critical. “When many engineers are bringing in data, there have to be people looking at it and asking, ‘Is this data reusable? Are we creating redundancies that will lead to different interpretations?’” Belino notes.

Governance, in this context, is not just about compliance. It is about preventing fragmentation and enabling scale.

Designing a data ecosystem built for reuse

A core pillar of Mars’ transformation has been the idea of data reusability. Belino explains that the team deliberately moved toward shared datasets. “We started turning product-specific data into what we call reusable data sets. That’s a big component of a data ecosystem needed for analytics.”

She compares the concept to sustainability. “You can think of it like recycling. Reuse allows data to become standard and common across the organization.”

Governance technology plays a key role in making this work. “There are many tech enablers that help you govern data. If you have the people who can do data governance, technology applied on top of your data makes it easier to govern and easier for users.”

Making data discoverable and understandable

Well-governed data must be easy to find and understand to drive real value. Belino emphasizes the importance of data discoverability and literacy across teams. “Engineers and data scientists have to obtain knowledge of what the data is very quickly. We have many product and project teams working on data. They have to know what data exists easily and quickly.”

Understanding definitions and lineage is equally essential. “They also have to know the definitions and lineage of that data so they can understand if it will help them solve their problem.”

This focus on discoverability helps new teams become productive faster while reducing duplication and confusion.

*The next part of this series explores how Mars balances delivering immediate business value while building long-term data foundations, the role of user-centric design, and how emerging approaches like agentic AI are helping bring insights closer to decision-makers. Stay tuned.

CDO Magazine appreciates Ujjwal Sehgal and Rachel Belino for sharing their insights with our global community.