Ahead of his session at the Data Innovation Summit, we caught up with Robert Børlum-Bach, Head of Data Architecture and Transformation at JP/Politiken Media Group, who advocates for a shift from reactive “data-as-a-service” to a “data-as-a-product” mindset. He emphasizes that by building internal “product muscles” rather than just expanding infrastructure, organizations can establish clear ownership and accountability, transforming data from a technical black box into a proactive driver of measurable business outcomes.
Many companies invest heavily in data platforms, pipelines, and tooling. Yet they often struggle to translate that investment into real decision-making impact. Why do you think that gap exists?
Robert Børlum-Bach: I think we often favour a belt and braces approach optimised for delivery, not value. Meaning, it feels safer to have a can-do-it-all platform and infrastructure that is under-utilised than the alternative. Especially when there’s so much FOMO regarding missing out on data and AI. Which platform vendors definitely also know.
The demand for data also grows exponentially, which we tend to solve technically first, organizationally second. And it is easy to mistake platform maturity for organisational maturity. Painting a skewed organisational picture of what we can actually do.
So the platform and the capabilities which support it often need to be strong and heavy due to criticality and complexity, but can tend to miss the agility and proactiveness that you get when the problem leads, and the tooling follows. Closing the gap between data reactiveness and proactivity is the million-dollar question.
You work closely with both platform and business intelligence teams. What patterns do you see in how data teams typically engage with the rest of the business?
Robert Børlum-Bach: Definitely the pattern of data-as-a-service, which, in a worst-case scenario, can cause a loss of a strategic voice. Requirements and tickets originate from “the business”, and the data team’s work and solution is often a black box. Which doesn’t help a business user’s frustration with why the question of “how many users did we have last month?” is apparently a complex query.
Fortunately, another pattern or trend is emerging: data-as-a-product. Ownership is defined end-to-end through a producer-consumer relationship. The data needs to have frontloaded clear use cases and criteria. Accountability doesn’t end at ‘shipping’, but when it has value. No more assets or dashboards that nobody is actually using.
This shift is, maybe most importantly, an organisational change. Data ownership moves closer to the domains that actually produce and consume the data. Each domain becomes responsible for owning, governing and exposing its data as products.
Which sounds great on paper, but there are multiple risks: siloed data, multiple sources of truth, a lack of technical skills in the domains, and no central governance function. Which is exactly why the platform and product layer need to be in sync. Decentralised ownership without a centre of gravity just creates another whirlwind of mess.
What does it actually take to build product thinking within a data organisation? What capabilities and practices make the difference?
Robert Børlum-Bach: Don’t skip leg day! Meaning, data organisations tend to train their infrastructure core and platform biceps, but tend to skip their product leg muscles. You need to deliberately train this product muscle, which is not an easy feat, and as with everything, you should start in small increments.
What we’re trying to do is to separate the platform from the products (solutions), which is uncomfortable as it creates visibility where there are gaps in technology, resources, and capabilities, but is needed to move from a cost centre to a strategic function.
Again, this is not primarily a technology problem. You can have the best platform in the world and still not solve it. It is an organisational adaptation.
When working with data products, a defined owner is needed. Someone accountable for its value and direction, who ensures it meets actual business needs, who prioritises what gets built and coordinates between the people producing the data and the people consuming it. That role is the bridge between the technical and the business layer.
This, together with an advisory/enablement team, is an important capability; they don’t create the dashboards or solution itself, but enable domain and product teams with discovery, problem framing and the translation between business need and data solution.
Building a product mindset also requires cultural change. How can data leaders train and develop this muscle within their teams?
Robert Børlum-Bach: To go back to the gym analogy, you don’t start with maxing out; you build gradually and make a habit. In my experience, you do not need a fully mature data product organisation, but you do need a north star answering the “why”: a product vision that clearly articulates what you’re trying to achieve. From there, you can scope and align around one data product with a real business problem, a clear owner and a clear consumer. A data product canvas is a useful tool here, as it forces the right questions: what is the problem, whose problem is it, what does the data need to look like, and how will we know if it worked?
Then, with the team, try to run a sprint, protecting the capacity and process from the as-is ad-hoc requests, getting important learnings about sizing, standards, governance and technical feasibility
Getting leadership buy-in is often the hardest part. How do you make the case for investing in data products with the same rigour as any other strategic priority?
Robert Børlum-Bach: You don’t lead by saying we need to think more like a product team, or even by mentioning data products. You start with the cost of the status quo. The capacity consumed by low-value requests, maintaining dusty dashboards that nobody opens, or the initiatives and OKRs where data hasn’t shown up.
When management sees these patterns and gaps, the data product framing becomes quite concrete.
But it IS a difficult conversation. Treating data as a product is an organisational commitment. It requires decentralised ownership, meaning someone accountable for its direction and value. The most mature organisations apply this to treating data quality issues as product issues, and data debt the same way as technical debt, something to actively manage.
And then show it, don’t tell it. The first imperfect data product, with a clear use case and data contract, is more valuable than 100 slides. Nobody gets strong from a single workout, but you do not get strong without starting either.
If you want to move beyond just training your “infrastructure biceps” and start building the “product muscles” that actually drive a high-impact data organization, don’t miss Robert Børlum-Bach live at the Data Innovation Summit.
In his session, “Beyond the Platform: How Data Products Turn Infrastructure into Impact,” Robert will provide a practical roadmap for aligning platforms and products to drive personalization, experimentation, and measurable business outcomes. Attendees will learn how to distinguish between platform scaling and true product value through real-world media industry examples and actionable strategies for data governance and organizational enablement.
