Henkel, the Düsseldorf-headquartered global company founded in 1876, operates across Adhesive Technologies and Consumer Brands, with a footprint that spans industrial and consumer markets and a workforce of around 47,000 people worldwide. Like many large enterprises, it has spent years building out its data and analytics foundations while navigating the tension between speed, control, and long-term scalability.

In the third installment of this interview series, Katrin Botzen, Corporate Director, Global Data and Analytics at Henkel, moves beyond platforms and operating models to focus on what actually works in practice. In conversation with Julian Schirmer of OAO, she distills Henkel’s experience into a set of clear, operational lessons, highlighting where early decisions created friction, where course corrections were needed, and how those insights now shape the company’s approach.

Across the discussion, Botzen outlines specific best practices that data and AI leaders can apply directly, from embedding governance and security into systems from the outset, to making deliberate build-versus-buy decisions, to balancing external innovation with internal architectural discipline. The conversation also surfaces the organizational realities behind these choices, particularly how leadership, collaboration, and decision-making evolve as data capabilities scale.

Part 1 of the series examined how Henkel applies AI to real business problems, while Part 2 explored how data products are operationalized.

Best practice 1: Build governance and security in from day one

Botzen begins with a lesson shaped by experience. Henkel has operated for years with a large data and analytics landscape and stable technology foundations. The company learned that early-stage flexibility still matters, especially when teams are trying to move quickly from business needs to usable use cases.

To move fast, Henkel started with a platform approach that gives broad freedom. “We started with a platform that was basically just a data lake, and everyone could do whatever they wanted, to be fast,” Botzen says.

The intent is right, she explains, because it aims to bring “faster value” to the people. But speed comes with a cost when guardrails are missing. “We lost oversight of governance and security,” she says.

Her first best practice is direct and operational: “Build governance and security in your technology processes and organization from day one.” The goal is to make technology easy to use while still having built-in security features.

Best practice 2: Treat “build vs buy” as a strategic choice, not a default habit

Botzen shares a second lesson that appears repeatedly across data and AI teams. The debate over whether to build tools internally or buy them is ongoing, and she says there is no universal answer.

But Henkel’s experience reveals a pitfall that becomes expensive over time. In one area, the company decided to build capabilities inside its platform, including tooling it could potentially buy. “The market back then was not ready,” Botzen explains. “We had great people, and they could do it. And they did it very well.”

The issue emerges later. “We are not a software company,” she says. The internal tool works at first, but the industry accelerates and begins to outperform. As the gap grows, the cost of maintaining internal tooling rises, especially when demands change quickly. “We had a lot of costs associated with dealing with this self-built tool, keeping up with the new demands,” she elaborates.

Her advice today leans toward buying if the market can meet the need. “My current advice would be that you can buy and be conscious on the make side,” she says, because keeping pace with technology companies is hard when your core business is not technology.

Making governance work in a fast-moving market

When asked how the company maintains governance while still adopting innovation, avoiding a fragmented tech stack, Botzen’s answer is a two-sided approach.

On one side, she emphasizes partner relationships and the role partners play in keeping Henkel close to market changes. The partners need to understand Henkel’s circumstances, stack, and organization well enough to recommend what fits.

She includes technology partners and consultancies and frames their value as being in proximity to what is changing. “They are at the heartbeat of the market when it comes to technology,” Botzen says, adding that strong relationships help because partners “bring in innovation.”

On the other side, she stresses internal architecture discipline. Botzen points to “very strong architecture principles,” naming enterprise architecture and data architecture as the mechanisms that ensure new capabilities fit into a coherent direction.

She mentions councils that help evaluate what the company wants to achieve, the technology that already exists, the gaps, and how new choices fit into the bigger picture. She is careful not to claim perfection, adding that Henkel still has to invest to become stronger in orchestration.

A career built on curiosity, not a fixed path

When asked about what it takes to move into a modern data leadership role, Botzen says her background does not follow a single expected template. She argues that the field increasingly rewards diverse routes into data and analytics. “It’s not that you have to have started computer science and this is the only way to get into data analytics. It’s the opposite,” Botzen says.

Her advice centers on flexibility and curiosity. “Do not stress yourself over one career path,” she says, and describes how she follows what she loves, moving from SPSS to Python to R, driven by interest in statistics and mathematics. The mindset matters more than the label on a degree.

Speaking about her daily challenges, Botzen frames leadership in data and AI as a human integration problem. She manages diverse teams across platform, science, engineering, governance, enablement, and strategy.

The central challenge is not only execution, but collaboration. She describes her role as bringing people together, facilitating mutual understanding, and translating across specialties so teams can align around a product and vision.

“Data is not a single team sport. It’s not one team that can solve data for a company. It’s actually everywhere,” Botzen concludes.

CDO magazine appreciates Dr. Katrin Botzen for sharing her insights with our global community.