Business "Un"-Intelligence

Why does Business Intelligence still fall short of delivering impactful decision supporting insights and what we can do about it?

Business intelligence (BI) is often seen as a critical functional area in modern enterprises where it is expected to be the key to democratizing business impacting insights and empowering decision makers. With a team dedicated to maximizing the value of the enterprises' data, opportunities to create operational efficiencies, mitigate risks, and identifying opportunities for revenue enhancements or cost reductions should be the bread and butter of the organization. However, in practice, many BI functions fall short of delivering real, business-changing value. Instead of acting as a dynamic decision-support system, identifying anomalies, patterns and trends, they often operate as operational reporting organizations. BI teams frequently find themselves mired in operational tasks, such as generating dashboards, preparing reports, or responding to ad hoc queries, rather than driving transformative insights that influence decisions that will enable risk mitigation, revenue enhancement and cost reduction.

Year over year, month over month, and ranking metrics are ubiquitous on most business intelligence dashboards or reports that flood emails on a daily basis across a host of companies. Though interesting, and necessary to understand the status quo and to keep you relevant when interacting with your leadership, do they really provide "intelligence"? Can you act on any of these metrics? Do they provide enough information to influence your decisions? Probably not.

Dashboards with their lines, pies and bars reflecting top line operational measures are informative but provide little insight into the variables that really impact your business. If they did, metrics around topics like the below would be at our fingertips:

• How are efficiency gains or disruptions in your supply chain contributing to your sales or margins changes?

• Has pricing competition intensified in your market and how is that impacting your sales strategy?

• Have there been changes in customer preferences?

• How are economic factors impacting demand?

The answers to questions like these will influence your planning and decision making.

So, if year over year sales, top 5 products, top regions, can't inform on what actions to take next, is Business Intelligence in the form that it exists in many organizations really a differentiator or force multiplier or is it just a weathervane that keeps us informed on which direction the wind is blowing, but nothing about what lies beyond the horizon. How do we make business intelligence "intelligent"? Capable of delivering actionable insights that minimize the risks and maximized the benefits when we make decisions.

Having spent the majority of my career in the data technology or data analytics space, this question has always surfaced in mind as a topic of extreme interest. However, before attempting to suggest a resolution to the challenge of intelligent BI, I think it would be worthwhile to explore some of the reasons why the impact of BI implementations tends to be underwhelming with its impact on organizations. While researching the topic and reflecting upon my past experiences and those of some of my peers, I came upon a host of possibilities. I would doubt that there is anything novel about the list below, that being said, I think it misses the mark as I think the list below can be considered symptoms of an underachieving BI function, not the reason for it.

Descriptive Analytics Domination: Most BI outputs are descriptive, focusing on what happened rather than why it happened or what will happen next. Metrics like year-over-year trends, top products, or regions highlight performance but do not provide actionable guidance for decision-making, but are a quick win. Without context or causation, decision-makers cannot bridge the gap between data and action.

Lack of Contextual Integration: Dashboards often operate in isolation, presenting data without linking it to external or internal factors such as market trends, customer behavior, competitive dynamics, or macroeconomic conditions. This disconnection leads to insights that are superficial and uninformative for strategy formulation.

Over-Reliance on Top-Line Metrics: Metrics like total sales, profit margins, or product rankings provide a high-level view but fail to dissect the underlying variables driving performance. This results in blind spots around efficiency gains, supply chain impacts, or shifts in customer preferences.

Operational vs. Strategic Disconnect: BI often focuses on operational reporting rather than strategic decision-making. Reports are designed to track KPIs, not to predict disruptions, recommend actions, or optimize future outcomes.

Limited Analytical Depth: Organizations frequently lack advanced analytics capabilities, such as predictive modeling, machine learning, or decision intelligence frameworks, that can uncover patterns, trends, and causation within the data.

Cultural and Communication Gaps: BI teams often fail to effectively communicate insights in a business-relevant narrative. Technical jargon, complex visualizations, or data overload can alienate decision-makers and reduce the actionable value of BI outputs.

I've attempted to break down the above symptoms to their core and have found that the underlying causal themes tend to be around:

1. The analyst's understanding of the business. Its strategy and objectives and quite possibly, the industry itself and their ability to communicate effectively.

2. A lack of data analysis skills. Not tool specific skills, but mathematics, statistics specifically. There is no getting around it, an understanding of statistics is critical to being a data analyst. Something as basic as understanding whether or not changes/deltas are within or outside historically observed ranges, as in, within one or two standard deviations adds significance to what is being observed.

3. A lack of data engineering skills. Data engineering does imply tools like SQL and/or Python. In its basic form, it is understanding that multiple data elements when combined, are better descriptors of what you may observe in the data. A simple, yet powerful example is creating a RFM field during your analysis that can segment your customers based upon customer behavior. More on RFM can be found here, https://www.actioniq.com/blog/what-is-rfm-analysis/.

4. Data architecture alignment with business use cases enabling an accurate view of the current state of operations. Architecture is beyond the scope of this discussion as to the topic's complexity. The first phases of architecture are more focused on future proofing the data environment than on being able to deliver immediate insights. By that I mean, the initial focus is on storage, compute, data quality and governance with the initial stories offering a proof of concept, but not necessarily enabling earth-shattering insights. Those capabilities usually unfold in future phases of the architecture's evolution as the data architect should expand their focus from infrastructure and data to people and processes.

I believe the above can be remedied with a renewed focus on the mission and purpose of BI. However, it will require for some, significant changes to the BI Analyst's job description and skills requirements. Business Intelligence analysts are often hired for their expertise in a specific tool (Tableau, PowerBI, etc.). In my career, I have witnessed 3 migrations from one BI tool to another in the hope that the new tool would solve the woes of the last one only to see a new version of the old problem. The problem was never the tool, it was the vision and as importantly, the data analyst's expertise and business understanding. On a recent engagement, I realized that the team had great BI tool skills but lacked terribly in their understanding of data (statistics) and core data skills (SQL). This resulted in beautiful, but underwhelming dashboards if you were seeking insights into the business' operations or state of health. Even though they worked very closely with the business units that they supported, their conversations with the business and their deliverables were constrained by their expertise in the BI tool.

My tendency to not place as much importance on tool competence as I do on data and business competence stems from an experience early in my data journey when I was hired as a statistical programmer in JPMorgan Chase's Behavioral Analytics and Insights Delivery team without any knowledge of SAS, their tool of choice. The vetting process was focused on statistics, problem solving and explaining my thought process clearly. Upon being hired, my manager told me something that I still live by today and it applies in this case, "You can always learn the tool, but I can't teach you critical thinking. That's why I hired you."

However, this only gets you partly there. The next step involved formal training in the business' functional areas and processes prior to being turned loose on the challenge of problem solving. In this case, a program that involved job rotation or cross training during the candidates first 4-6 months of employment was implemented. The result was of course, delayed productivity, but far more impactful productivity when the candidate was reintegrated into the team.

By shifting the purpose of BI to supporting the business' growth strategies, we should attempt to drive as much operational reporting to automated solutions, self-service and if necessary, a team of Operations Analysts, specifically trained to generate what has historically been called BI. With a mission to identify the most actionable opportunities to grow revenue, reduce risk or achieve other business objectives, the team can focus on developing trusted partnerships with experienced business partners and senior leaders and utilizing their unique toolset to delivering the actionable insights we all desire. By complementing this change in leaderships perspective and expectations of BI and changes to what we look for when hiring, we get closer to creating an "intelligence" providing organization. When hiring, I would emphasize "interests", "learner mindset", and "problem solving techniques" above BI tool expertise.

Business intelligence, as it exists in many organizations, often functions as a rear-view mirror, informing stakeholders about past performance without offering guidance on what lies ahead or how to navigate it. To make BI truly intelligent, it must evolve into a forward-looking, decision-support entity that goes beyond surface-level metrics to uncover actionable insights. With the right team in place, an approach to Business Intelligence that enables an ability to remedy the aforementioned symptoms outlined at the beginning of this document becomes possible. Once those shortcomings are addressed, activities like integrating more advanced analytics, focusing on cause-and-effect relationships, and embedding BI into strategic decision-making are possible. Only then can organizations transform BI into a force multiplier—one that minimizes risks, maximizes benefits, and drives sustainable growth. The key is simple: stop reporting the "what" and start uncovering the "why" and "what next."