Is your data an asset or a liability? Preparing your foundation for the future of AI and analytics 

In this blog post, our data strategy expert Laura Winkelbauer will outline five telltale signs of data friction, before presenting you with our tried-and-tested solutions. Laura’s career covers strategy consultancy, e-commerce and for the past six years at FELD M, she's advised decision-makers on data strategy and building data-driven organizations. So, in other words, she really knows her stuff. Let's dive in, shall we?

"When data products fail to deliver, the issue is rarely a lack of talent or technology. Usually, it points to a foundational gap in how data is managed."

- Laura Winkelbauer, FELD M data strategy expert

Data can be a highly valuable strategic resource, but it can also be a source of real friction. 

Ever been in a high-level meeting where two departments present two entirely different versions of the truth? Based on different numbers, or different dashboards, or different definitions of what a metric means? That's the kind of friction we're talking about. 

But when data products (the dashboards, recommendation engines, and personalization tools we rely on) fail to deliver, the issue is rarely a lack of talent or technology. Usually, it points to a foundational gap in how data is managed.

The cost of data friction: How to spot the problem

If your organization is struggling to scale its data efforts, you’ll probably recognize these five challenges:

1. Inconsistency across data products

If every analytics tool you look at tells you something different, it can be a significant barrier to scaling.

When dashboards, automated alerts, and decision-support tools (think e.g. next best action) provide conflicting insights, the result is operational paralysis.

This lack of uniformity forces high-value analysts to act as "data investigators" rather than strategic partners.

2. Delayed time-to-market

In an agile market, speed is a competitive advantage. However, many teams launching a new data product find it almost impossible to capitalize on that advantage.

If your engineers spend more than 30% of their time finding and cleaning data rather than building features, your innovation cycle is broken. 

3. The "Shisho" effect on innovation

You can’t build a high-performance engine with contaminated fuel. When the "starting point" for data is poor, adopting advanced technologies like Generative AI or predictive modeling becomes nearly impossible.

This "garbage in, garbage out" cycle prevents you from moving past basic reporting into delivering meaningful impact. 

4. The erosion of stakeholder trust

Perhaps the most significant damage is psychological.

When data products are consistently slow or inaccurate, business leaders lose confidence. They stop looking to analytics for guidance and return to making decisions based on intuition, effectively neutralizing your investment in data. Having stakeholder trust is crucial, as Jesse Rothenberg from Ableton points out in our recent interview.

5. Regulatory and compliance risks

With increasingly stringent global privacy standards and the EU AI Act adding further scrutiny to companies' data practices, you simply cannot afford not to know the exact lineage of your data and have clarity on the inner workings and outputs of data products.

Plus, handling customer data without a clear framework is a significant legal and financial risk that no modern company can afford to take.

Data governance: The strategic solution

"Data governance" is a term often clouded by consultant-speak and promises of "digital transformation." But stripping away the jargon, data governance is simply the operating system for your information.

The benefits of data governance 

The symptoms of data friction, represented as an iceberg. If you have struggles relating to trust, quality, efficiency, and compliance, it can be a sign of something going on under the surface.

Implementing a grounded governance framework isn't about adding bureaucracy; it’s about removing the hurdles that slow you down. 

“We want to be in a better place when it comes to data and the introduction of new, AI-driven technology and tools"

Sound familiar? 

If so, here’s why data governance deserves a seat at your strategic table: 

• Consistency. Dashboards, models, and experiments use the same definitions and logic, so that metrics align across marketing, sales, and analytics.

• Efficiency. Data engineers, data scientists, and analysts no longer need to rebuild pipelines or hunt down undocumented tables. They can focus on insights and features rather than repairs.

• Trust. When business users understand where numbers come from and what they represent, debates shift from “which number is right?” to “what does this tell us?”. With the trust in numbers comes the trust in the teams that provide them, which in turn fosters that crucial blending of business and analytics capabilities that brings maximum value.

• Compliance. With clear data ownership, lineage, and documentation, you meet EU AI Act requirements like transparency for AI systems and robust data quality checks. The same effect applies to compliance with privacy frameworks.

• Innovation readiness. With trusted, well-documented data, teams can adopt new technologies faster, experiment safely, and deliver new products to market with confidence.

What data governance actually looks like 

So what does this "operating system" look like in practice? Data governance isn't a single tool or a massive reorganization. It's a framework built on a few key pillars: 

Iceberg with the root causes revealed: By addressing these root causes you can create a brighter picture above the surface.

• Clear ownership. Every dataset has a designated owner. That person understands what the data represents, knows its limitations, and can answer questions about it. No more Slack posts asking "does anyone know if this is still relevant?" and mystery tables that no one dares to touch or delete.

• Shared definitions. Your organization agrees on what "customer," "conversion," or "active user" actually means. These definitions live in a central place where everyone can find them. When marketing and product discuss retention rates, they're speaking the same language.

• Documentation that travels with your data. Metadata isn't an afterthought buried in a forgotten wiki. It lives alongside your data: where it came from, how it was transformed, when it was last updated, and who's using it downstream.

• Quality standards with teeth. You establish what "good enough" looks like for different use cases and build checks to catch issues early. Not every dataset needs perfection, but every dataset needs appropriate quality for its purpose.

• Access that balances security and speed. People can find and use the data they need without waiting weeks for approvals, while sensitive information stays protected. It's about smart guardrails, not roadblocks (the same applies for our AI literacy training!)

Conclusion

We understand that many organizations are wary of "big-picture" solutions that fail to deliver practical results tomorrow.

However, data governance is not a one-time project, it’s a fundamental shift in how you treat your company's most valuable resource – it is one of the rare multiplier effects improving every future decision.

Change takes time, but the outcome is restored trust, better decisions, better awareness, the ability to truly innovate, and ensuring your company is ready for whatever comes next.

If you'd like to read more about what this looks like in practice, one example is our work with JobCloud, the leading digital recruitment platform in Switzerland.

Facing growing data complexity and increasing demand for new use cases, JobCloud partnered with FELD M to build a holistic data strategy aligned with business goals, covering governance and architecture as well as enablement and data literacy.

You can read the full case study here:

• Efficient use case implementation with FELD M’s data strategy