In today’s business world, the term “data-driven” is often used and overused, but what does it really mean? Many organizations claim to be data-driven, yet in practice, most are actually data-informed. This distinction is crucial for understanding how data is used in decision-making processes. While companies may believe they’re leveraging data to drive their strategies, they’re often still relying on human intuition and past experiences, with data serving as one of many inputs rather than the guiding force. Furthermore, even fewer are aware of the concept of decision-driven analytics—a more evolved, prescriptive approach that should be the ultimate goal for any organization striving to use data in a truly impactful way.
According to Stylianos Kampakis, author of The Decision Maker’s Handbook to Data Science, being data-informed means using data as contextual support for decision-making. Dashboards, reports, and KPIs guide discussions, but ultimately, human judgment and intuition dominate the final decision.
Moving beyond merely being data-informed, organizations that are truly data-driven rely heavily on intelligent algorithms and models. These organizations actively use data to influence or even automate decisions. For instance, a sales team using predictive lead scoring isn’t just guided by data—it’s driven by it, as machine learning models dynamically prioritize leads and adjust outreach strategies in real-time.
However, becoming fully data-driven is challenging. According to the 2025 MIT Sloan Management Review survey, only 33% of organizations describe themselves as genuinely data- and AI-driven. Therefore, approximately two-thirds of data scientists still find themselves in environments where data isn’t fully embraced, making it difficult to translate insights into meaningful actions.
This gap emphasizes a key insight: simply having access to data isn’t enough. Organizations need the right tools, the right mindset, and—most importantly—the right roles and skillsets to evolve from data-informed to data-driven, as a first step.
Introducing Decision Scientists
The role of Decision Scientist is crucial for bridging this gap. Unlike traditional data scientists who focus primarily on modelling and analytics (the science of inputs), Decision Scientists combine the deep technical expertise of a data scientist with strong business acumen and problem-solving creativity. Their focus is on decision-making and providing actionable recommendations (the science of outputs), ensuring organizations successfully navigate the critical “last-mile problem” in analytics.
Comparison: Data-Informed vs. Data-Driven vs. Decision-Driven
| Key Attribute | Data-Informed (Topic-Based) | Data-Driven (Mixed Analysis) | Decision-Driven (Decision-Based) |
|---|---|---|---|
| Decision Style | Data provides contextual support for decisions, but intuition and experience frequently override deeper analysis. | Algorithms & insights from models actively influence or automate decisions, reducing reliance on gut instinct. | Data dictates or prescribes decisions, ensuring alignment with specific business goals and measurable outcomes. |
| Example | A marketing team notices open rates have dropped but relies on past playbooks to fix it, with no rigorous testing. | A sales team uses predictive lead scoring to rank leads but may not explicitly ask “How do we best balance acquisition costs vs. conversion rates?” | A telecom co. asks “How do we best retain customers while controlling costs?” and works backward to define data, models, and interventions, validating causal drivers before prescribing tailored solutions. |
| Analytics Focus | Descriptive & Diagnostic – focus on “topics” that emerge from the data (e.g., lower open rates). | Predictive & Some Prescriptive – advanced models guide decisions but may not be systematically tied to broader strategy. | Full Prescriptive – begins with a defined business decision and works backward, using descriptive, predictive, and prescriptive analytics in tandem. |
| Analysis Approach | Topic-based—issues or trends surface from the data, but not necessarily linked to specific decisions. | Mixed—combines topic-based and some decision-based approaches; predictive models optimize certain processes but broader questions may remain unanswered. | Decision-based—every analysis is framed around “Which decision are we making?” ensuring analytics outputs directly inform strategic objectives. |
| Causality | Minimal—often correlation-based or anecdotal. | Partial—models can highlight cause-effect, but it’s not always systematically tested or confirmed. | Prioritized—experiments or causal inference confirm which levers truly drive outcomes. |
Why Decision-Driven Analytics Matters
While data-driven decision-making is important, it can still fall short by generating insights without clear paths to action. Experts Bart de Langhe and Stefano Puntoni advocate for a decision-driven approach.1 Rather than starting with available data or broad topics for exploration, decision-driven analytics begins by identifying the specific critical business decisions at hand and then determines the data and analysis required. This ensures that analytics are directly aligned with strategic objectives, leading to insights that are not only relevant but also actionable.
Data-driven decision-making can fail to meet expectations. Decision-driven data analytics may fare better.
MIT Sloan Management Review
Decision-driven analytics follows three essential steps:
- Define the Decision –> clearly articulate the business problem and corresponding decision(s) to be addressed.
- Identify Necessary Data & Analytics –> determine which data, models, and analyses will effectively support the decision.
- Measure Outcomes & Close the Loop –> evaluate the impact of decisions made and adjust strategies based on measurable outcomes.
At its core, decision-driven analytics ensures that every analysis is framed around answering a key business decision. It begins by defining the problem—What decision are we trying to make?—and uses that clarity to guide the entire process of data collection, modelling, and analysis, to running “what-if” scenarios to quantify uncertainty and risk, exploring expected-value frameworks to bridge the gap between predictions and profitability, and embracing the relentless pursuit of optimizing incrementality and lift. Unlike a data-driven approach that may produce insights without necessarily aligning with strategic goals, decision-driven analytics works backward from the decision to ensure that the right data, models, and analyses inform the most critical business objectives.
For example, a telecom company might ask, How do we best retain customers while controlling costs? This decision drives the entire process. From there, the company would define the data needed (e.g., churn risk factors, customer behavior, cost structures), apply models to predict which customers are most likely to churn, and then — a very critical next step — use prescriptive analytics to test interventions and optimize retention strategies. The goal is not just to understand the problem, but to prescribe actionable solutions that directly impact business outcomes (note: in the coming weeks I will introducing the The Decision Science Pre-Analysis Framework in a six-part series, based on a decision-driven framework).
Transitioning to decision-driven analytics helps organizations avoid the common pitfall of “water-is-wet” analysis—surface-level insights that confirm the obvious without providing actionable direction. For instance, a Marketing team might report declining engagement rates but fail to offer actionable guidance on what to do next. In such a case, the team remains data-informed rather than data-driven or decision-driven.
A Call to Action for Executives
For executives, understanding the distinctions between data-informed, data-driven, and decision-driven is not just about semantics—it represents a fundamental shift in strategic decision-making. According to McKinsey, a leading authority on C-suite strategy, “data culture is decision culture.” Organizations don’t merely need more data—they require a culture prioritizing decision-driven analysis at every organizational level.2
Executives relying solely on a data-informed approach risk stagnation, trapped in cycles of retrospective reports and dashboards without meaningful transformation. In contrast, organizations that embrace decision-driven analytics embed data science and predictive analytics deeply into their strategic decisions, optimizing processes and empowering decision-makers with actionable insights.
As suggested by Erik Larson in this Forbes article, it is important to consider that decision-making is both “the most important and most poorly managed business activity”. These two numbers tell the story:3
- Decisions drive 95% of company performance
- Decision-makers fail to use best practices 98% of the time
A McKinsey survey revealed that for a typical Fortune 500 company poor decision-making resulted in approximately 530,000 wasted days of managerial time each year. Furthermore, this inefficiency cost these enterprises around $250 million in annual wages. Consequently, the average employee spent 47 days each year dealing with the repercussions of ineffective decisions, which included rework, delayed product launches, and missed business opportunities.4 Ugh!
Furthermore, according to a 2017 Gartner report, 85% of big data projects fail5, while VentureBeat in 2019 highlighted that a staggering 87% of data science projects never reach production. 6 Furthermore, Gartner’s 2019 prediction indicated that through 2022, only 20% of analytic insights are expected to deliver business outcomes.7 All the above statistics underscore the significant challenges faced by organizations in harnessing the full potential of data science initiatives, and the fall-out for ineffective data-driven decision-making.
As McKinsey further notes, many organizations struggle due to Executives lacking clarity between traditional analytics (business intelligence and reporting) and advanced analytics (predictive and prescriptive tools like machine learning). This clarity gap prevents full leverage of data for strategic outcomes.
Executives must actively lead this transformation—from data-informed to decision-driven—to unlock the full potential of analytics and secure lasting competitive advantage.
Will your organization lead or lag behind?
- Decisions, not data, should drive analytics programs ↩︎
- Why data culture matters ↩︎
- Can Market Researchers Save The Day By Becoming Decision Scientists? ↩︎
- Three Keys to Faster, Better Decisions ↩︎
- Why Big Data Science & Data Analytics Projects Fail ↩︎
- Why do 87% of data science projects never make it into production ↩︎
- CIO: Transforming analytics into business impact ↩︎
Leave a Reply