Leading with data: From access to action

🇫🇮 This interview is also available in Finnish (original version); please find it using the language switcher on the top right corner.

What does it really mean to lead with data and how to make it work in practice?

In this in-depth interview, we explore why data leadership is as much about culture as it is about technology. From the pitfalls of treating data as “just IT’s job” to the challenges of democratizing access across an organization, you’ll discover what separates companies that simply collect data from those that actually use it to steer decisions.

With real-world examples from finance, lessons learned from failed projects, and practical advice on building data literacy, this interview shows how leaders can move from availability to impact. If you care about making data everyone’s business, this conversation provides insights you can put into action.

What does “leading with data” mean from your perspective?

At its simplest, leading with data means making decisions based on concrete, measured information rather than gut feeling alone. Collected data is analyzed and refined into decisions that guide actions and business in the right direction.

In scientific research, it has long been a given that conclusions are based on measured data. For example, the effectiveness of medicines or animal behavior is evaluated using collected data and statistical methods. Conclusions are drawn from broad datasets, not single observations.

In business, this means decisions are based on analyzed data; whether it’s product comparisons, understanding customer behavior, or other business development areas. A key aspect is also weighing the benefits of decisions against their costs.

Leading with data isn’t just the privilege of giants like Amazon or Netflix. Small companies benefit just as much. That’s because it’s about systematically and proactively understanding your operations, customers, and market. Leading with data is not a technology investment, but a shift in mindset and culture where decisions are made openly and based on facts.

It’s important to remember, however, that data isn’t infallible. Most of the time it points toward better outcomes, but wrong or incomplete data can lead to mistakes. That’s what happened, for example, with the Google Flu Trends project, where search behavior didn’t reflect actual illness rates, or Amazon’s recruitment algorithm, which started discriminating against women. What these examples have in common is data bias, which led to flawed conclusions.

Photo from Recordly's latest hackathon 

Why do so many organizations still see data as IT’s responsibility, and what risks does this involve?

Many organizations still see data as IT’s responsibility because handling it involves many technical elements. That’s natural, but problems arise when technical implementation and business needs don’t align. IT may feel that all data is readily available, while business says it has no data at all. In practice, this means business has no way to find or use data without IT’s help.

The vocabulary and expertise between business and IT differ greatly. If you speak different languages, reaching a shared understanding is difficult. That’s why a common business glossary and mixed teams, including both technical and business expertise, are essential. Planning before doing also reduces misunderstandings and unnecessary work.

It’s also largely about culture: everyone should at least understand the basics of the other side’s work. Technical roles should know business, and business roles should know basic data skills. Product owners and business advisors play a key role as bridge-builders.

Without shared planning and dialogue, it’s easy to end up in situations where the work serves no one—hurting both business benefits and job satisfaction. Even small investments in communication and shared understanding can prevent wasted effort and increase data’s impact across the organization.

What makes the financial sector particularly challenging or interesting from a data perspective?

One of the main challenges is the long retention requirement for historical data. Data must be usable for years, even decades, and system changes make this harder. Information from different sources must also be aligned with history, which is laborious, costly, and drains resources from other development work.

Heavy regulation is another major factor. Tight deadlines and requirements consume resources and slow the adoption of new solutions. Regulation is nearly the same for all players, but smaller organizations have more limited capacity to implement it. Ensuring compliance consumes a significant portion of resources, leaving less room to use data for business development.

At the same time, the long-term data and sheer number of transactions make finance a uniquely fertile field for analytics and data use. Almost every event, whether a payment, loan, or investment, generates new data that can be leveraged for business development. This enables proactive management, where scenarios and forecasts help direct resources and strategic choices toward future needs.

The sector is also broad and diverse: different business areas have very different needs and operating models. Mastering one area doesn’t mean you master the whole field; there are always new domains and specifics to learn.

Photo from Recordly's latest hackathon and Finnish baseball game 

What are the most common pitfalls in data democratization?

The biggest challenges relate to lack of collaboration, absence of a shared vocabulary, and data not being easily discoverable. If data can’t be found, it can’t be democratized.

Another pitfall is unclear roles and responsibilities, or not having enough time to fulfill them. When “everyone is responsible,” it often means no one is.

How do you move from “data is available” to actually using it in decision-making?

Availability alone isn’t enough; data also needs to be understandable and relevant. If the technical side focuses only on building something that is technically correct but irrelevant to business, data will always be merely available, never used.

The solution is to bring data closer to business through conceptual models, catalogs, and glossaries. Data trainings and demo sessions are also essential for knowledge sharing. These give business the best chance to take ownership.

Which roles are critical to ensure data benefits more than just IT?

Product and project owners have a major role in ensuring the right data is made available and quickly put to use in the organization. A culture of collaboration also makes data use easier and more effective. As needs grow, product owners become even more important for clarifying business requirements quickly.

In large organizations, data steward roles may be needed to act as interpreters between business and IT.

Photo from Recordly's latest hackathon and Finnish baseball game 

What skills will leaders need in the future to leverage data?

Data literacy is essential. Above all, it means critical thinking and the ability to separate what’s relevant from the mass of analyses and reports.

Business behavior often fluctuates: in good times, companies are satisfied and don’t examine data deeply, while in bad times changes may be made hastily. In reality, it’s in good times that data should be examined more closely because when the market is strong, significant additional profits can be found by developing operations in the right direction. Conversely, in bad times metrics may look weak, but analysis might reveal that not all areas need changes. Adjustments should be targeted where they truly matter and that’s exactly what leading with data enables.

Another important skill is approaching data with healthy skepticism when results differ clearly from previous knowledge. The less data is available to support analysis and the less the sources are known, the higher the risk of error. But this is not a reason to avoid using data. Research makes it clear: intuition is neither reliable nor systematic for making major decisions.

Practical ways to improve your own data literacy

• Read literature: it helps you get familiar with core concepts and what they mean.

• Learn from successful examples: see how metrics and results are interpreted.

• Apply learning in everyday work: practice with your company’s data and explore it hands-on.

• Think critically about visualizations and reports: what’s understandable, and how do they support your work?

• Make changes gradually: small adjustments make it easier to track impacts. Changing too many things at once can confuse results and make interpretation harder.

What would it look like if data truly belonged to everyone?

If data really belonged to everyone, it wouldn’t just be about access rights. The data itself would need to be high-quality and available without intermediaries. Everyone could use it independently for decision-making, reducing the need for static reports.

At its best, this would speed up decision-making and bring new perspectives. When everyone shares the same understanding of key metrics, discussions are grounded in a common knowledge base.

The challenge is that open access can also increase misinterpretations. That’s why clear standards, a shared vocabulary, and agreed metrics are necessary.

The benefits could be significant: faster development of new products and services, more innovation from multiple perspectives, and improved customer experience when decisions are based on shared, up-to-date information.

In decision-making, data is ultimately a tool, not an automatic answer. By implementing predictive models and tracking their accuracy, organizations can move from merely reacting to events toward planning for the future. This doesn’t replace the need for business expertise. At its best, it combines experience and analytics, so decisions are based on both.

If leading with data and using analytics are left only to analysts and statisticians, the organization won’t capture their full potential. That’s why data use must extend across the whole company. Without the example and active role of top leadership, transforming into a data-driven organization is difficult to achieve.