Analytic Drama - Why Business Intelligence Has Failed Users

The business intelligence (BI) space is massive—and growing rapidly. By 2032, companies are projected to spend even more on data and analytics, cementing BI as one of the top areas for corporate investment. This trend isn’t new; organizations have been pouring money into BI tools and analytics for years, aiming to gain a competitive edge through data-driven decision-making.

According to Forrester, spending on data and analytics is expected to skyrocket from $30 billion in 2022 to an astounding $390 billion over the next decade. Similarly, the global data warehousing market reflects this substantial growth trajectory. Valued at $21.18 billion in 2019, it is projected to reach $51.18 billion by 2028, growing at a compound annual growth rate (CAGR) of 10.7% from 2020 to 2028. These figures underscore the increasing reliance on data-driven solutions and highlight the monumental investments being funnelled into the tools and infrastructure that support advanced analytics.

Here’s where things get interesting. When we think about BI usage, we might imagine employees across the organization actively engaging with dashboards, charts, and reports to make better decisions. But in reality, the people who interact with these tools the most are often the data team itself—verifying the accuracy of dashboards—or interns tasked with creating reports for leadership meetings. True end-user engagement—the kind where non-technical employees use BI tools to drive their daily work—remains alarmingly low.

This disconnect raises questions about the effectiveness of BI tools in delivering value to organizations. Are these tools too complex for everyday users? Are they failing to provide actionable insights? Or is the problem rooted in how BI solutions are implemented and introduced to teams?
As we continue to invest billions into business intelligence, it’s time to critically assess why these tools have struggled to meet the needs of the very people they’re designed for. End users should be at the heart of BI adoption, and until this challenge is addressed, the promise of data-driven organizations will remain unfulfilled.

There are three main issues that are contributing to this. Let's dive into the first one.

Problem 1: Data itself

Data is hard, and building trust in data is even harder. The first issue we face is that data trust, as a foundational pillar, requires a significant amount of expertise and continuous investment. Without trust, data becomes unreliable, and unreliable data leads to poor decisions. Trust is built from the ground up—starting with accurate collection, proper storage, and consistent governance. It’s not enough to simply have data; it must be clean, validated, traceable, and properly named and described.

For many organizations, achieving this level of trust demands specialized skills. Data engineers, analysts, and governance teams are tasked with ensuring pipelines are functioning correctly, data is up-to-date, and any inconsistencies or errors are dealt with swiftly. But here’s the challenge: those skills are not only hard to find but also difficult to retain. As data complexity grows, so does the need for advanced tools and frameworks to manage it. Yet, even the best tools and smartest teams can only succeed if there’s ongoing investment in systems, training, and maintenance.

Another critical aspect of building trust is transparency. Stakeholders across an organization need to know where data comes from, how it’s processed, and whether it’s reliable. This requires clear documentation, robust monitoring, and proactive communication. When teams lack visibility into the data pipeline, trust erodes quickly. For example, if a sales team notices discrepancies between a dashboard and their own records, they’ll lose confidence in the insights being presented. Rebuilding that trust takes time and effort—far more than it would take to address the root causes of the issue in the first place.

As I’ll describe next, the second problem we face is very much related to the sheer volume and complexity of data. Organizations today are collecting vast amounts of information from countless sources—customer interactions, operational metrics, financial performance, and more. This deluge of data doesn’t just require attention, maintenance, and governance—it needs business context. Without the connections and semantics that give it meaning and value, the data remains just noise.

Problem 2: Definitions

Now, let’s dig deeper into issue 2: the definition of metrics. At its core, this is all about semantics, meaning, and ontology. These three elements form the foundation of how businesses create clarity and build trust around their data and metrics.

Semantics: The Language of Data

Semantics is the shared understanding of what a metric represents. It’s how we interpret raw data into something meaningful and trusted. Without clear semantics, different teams within an organization can interpret the same data in entirely different ways, leading to confusion and misalignment.

As Socrates famously said, “The beginning of wisdom is the definition of terms.” It's almost like the ancient philosopher foresaw the chaos of modern data management over 2,000 years ago. Who knew that the guy who spent his time asking deep questions in the agora would have such a spot-on take about metrics and trust in the digital age? Clearly, wisdom—and maybe even a bit of data insight—stands the test of time.

By establishing semantic consistency, businesses ensure that everyone is speaking the same data language. This isn’t just a technical task; it’s a collaborative effort between data teams and the business. Semantics bridges the gap between raw data and human understanding, creating a common vocabulary that’s crucial for decision-making.

Meaning: Getting Precise About Definitions

Meaning is where we get specific. It’s not enough to say, “We track revenue.” We need to ask deeper questions: What kind of revenue? Gross or net? Are we including discounts, refunds, or taxes? What’s the calculation?

Precise definitions create trust. Without them, even the most sophisticated analysis can lead to misinterpretation. This is why opinionated and explicit definitions matter. When someone sees a metric, there should be no ambiguity about what it represents. A metric should come with a clear explanation: its name, its purpose, its calculation, and its rules for aggregation.

Ontology: The Structure Behind Metrics

Ontology takes semantics and meaning one step further—it’s the framework that organizes how metrics relate to each other. Think of it as the blueprint that connects your business’s data ecosystem. Ontology defines the relationships between different data points: how metrics like revenue, profit margin, and customer lifetime value are interconnected. It also establishes the lineage of data, showing where it comes from and how it’s derived.

Why is ontology so important? Because it allows organizations to build scalable, reusable metrics. Instead of reinventing the wheel every time a new report is needed, businesses can rely on a well-structured ontology to pull consistent, reliable metrics that are grounded in the same rules and definitions.

Level up data-driven decision making

Make metric analysis easy for everyone

Get Started Now

Why It All Matters

When semantics, meaning, and ontology are ignored, businesses end up with what can only be described as a "metrics mess." You’ve probably seen this before—teams pulling data from different sources, using inconsistent definitions, and arguing over whose numbers are correct. It’s not just frustrating—it’s a threat to effective decision-making.

On the flip side, when businesses prioritize semantic clarity, precise meaning, and a solid ontology, they create a foundation of trust. Teams can spend less time debating data and more time acting on it. Metrics become the contract between data producers and consumers, ensuring everyone is aligned and working toward the same goals with confidence.

Moving Toward Opinionated Metrics
So, how do we address this issue and move toward a system that works? The answer lies in adopting opinionated metrics. Opinionated metrics are clearly defined, universally understood, and well-documented. They leave no room for ambiguity because they are built with purpose and precision.

To illustrate the concept of opinionated metrics, consider the following example:

• Name: Account Renewal Rate
• Description: Tracks the percentage of retention of renewed accounts down to the period of invoicing.
• Type: Ratio
• Aggregation: Mean
• Format: Percentage
• Expression: Count of Accounts Renewed / Total Accounts Up for Renewal
• Dimensions: Product, Country, Cohort 

This clear and structured definition ensures that everyone interprets and uses the metric consistently, eliminating ambiguity and fostering alignment across teams.

The Path Forward

In today’s data-driven world, metrics are the language of business. By focusing on semantics, meaning, and ontology, organizations can unlock the full potential of their data. They can move away from confusion and inconsistency and toward a culture of trust, clarity, and action.

As we continue to explore this topic, we’ll dive into some real-world examples of companies that have successfully tackled the challenge of metric definitions. Because at the end of the day, clear metrics aren’t just a technical necessity—they’re a competitive advantage.

Problem 3: Humans

This brings us back to a critical reality: no one person or group has all the answers. End-users are not data engineers or analysts, and engineers are not experts in sales, marketing, or finance. Everyone has their own area of expertise, but when it comes to data, collaboration is key. The challenge lies in bridging these gaps—combining technical knowledge with business acumen to ensure data is actionable, understandable, and ultimately useful.

Adding to the complexity is the fact that every individual brings biases to the table. Whether it’s how we interpret data, prioritize tasks, or make decisions, these biases shape our actions, often without us even realizing it. When biases and limited expertise collide, the end result can be decisions based on incomplete or misunderstood information. This underscores the importance of designing tools and workflows that cater to cross-functional teams, ensuring that no group is left behind due to lack of technical knowledge or access.

End-Users Are Still Not Using BI Tools

Let’s face it—most business users don’t spend their day in BI tools. They’re in email, Excel, PowerPoint, or other platforms where they’re already comfortable. Even if the business invests in state-of-the-art BI solutions, adoption often lags because these tools aren’t seamlessly integrated into their workflows. For many end-users, BI dashboards remain intimidating, overly complex, or just not seen as a priority. Instead, they rely on analysts or data teams to interpret the data for them.

This cycle perpetuates a critical problem: the disconnect between data insights and decision-making. If BI tools aren’t accessible to end-users, and data teams are overburdened interpreting reports, the organization’s ability to act quickly and strategically diminishes. It’s clear that BI tools, as they stand today, are not fully solving this problem.

Are Visualizations Helping or Hindering?

The visualizations that come out of modern BI tools are supposed to help us make sense of data. But as we saw with the examples earlier, many of them do the opposite. Overly complicated dashboards, confusing visual elements, or excessive details can render even the most insightful data useless. If the average end-user can’t glean actionable insights within seconds, the visualization has failed its purpose.

           
Pie charts, tree maps, and other flashy visualizations might look impressive, but they often prioritize aesthetics over function. Worse, they can mislead users by oversimplifying or obfuscating key information. When faced with visuals that are too complex or unclear, users often revert to their default behavior: asking an analyst for help or bypassing the BI tool altogether in favor of tools like Excel.

Building a Future that is Trusted

The future of BI isn’t just more powerful tools or flashier visualizations—it’s about creating systems that truly empower users to make data-driven decisions.

I've outlined some of the key challenges and persistent obstacles to data adoption, along with three critical takeaways to help organizations move forward. By addressing these areas, you can build better systems, foster collaboration, and ensure your data truly drives your business forward.

First, prioritize the quality and completeness of your data. This is the foundation of everything else. Trust begins here, so work closely with your data team to understand what investments are needed and ensure this groundwork is solid.

Second, focus on semantics. Clearly define your metrics, dimensions, and terms to create a shared understanding across teams. This alignment serves as a "contract" between your data and how it’s consumed, ensuring everyone is on the same page about what the data represents and how it should be used.

Finally, tie your data efforts to business objectives. Align your analytics with your company’s goals and strategies, ensuring that the insights you generate are directly tied to what matters most. The saying goes, “what gets measured gets improved,” but the reverse is also true—if it matters, people will want to measure it. When data is connected to critical business outcomes, it becomes indispensable.

Building trust is the thread that ties all of these elements together. It’s a process that takes time and requires collaboration between end-users and data teams to ensure confidence across the entire system. While challenges remain, there’s a lot of exciting potential for analytics to drive meaningful progress in the future.