Data Stories vs. Stories With Data: Choosing the Right Framework for Data-Driven Decision Making

Published: October 2, 2024

Introduction: The Two Faces of Data Storytelling

Organizations now invest heavily in data communication. A 2023 Gartner survey indicated that 80% of business intelligence initiatives fail due to poor data narrative execution, not technical limitations. The critical distinction lies not in the data itself, but in the structural relationship between data and narrative.

Brent Dykes, author of Effective Data Storytelling, introduced a culinary analogy that clarifies this structural distinction: the sandwich model versus the pizza model. A data story functions like a sandwich—data serves as the essential foundation (the bread), without which the structure collapses. Conversely, a story with data resembles a pizza—the narrative forms the crust, and data functions as a removable topping.

This article argues that neither model holds inherent superiority. Each originates from different communicative requirements and serves distinct audience segments. The strategic skill lies in matching the framework to the context, not in privileging one approach over the other.

What Is a Data Story? (The Sandwich Model)

In a data story, the data precedes the narrative. The data itself reveals the story; narratives and visualizations exist solely to contextualize and explain what the data has already demonstrated.

Structural characteristics:

• Data as the foundation: Without the data, no story exists. The narrative cannot be constructed until the data has been analyzed and conclusions drawn.
• Rigid sequencing: The story follows a logical, data-determined path. The author cannot arbitrarily rearrange insights without compromising integrity.
• Heavy visualization dependency: Charts, graphs, and statistical representations constitute the primary communication tools. Text serves as annotation.

Applicable contexts:

Data stories perform optimally in environments where audiences possess data literacy and require defensible, actionable insights. Typical use cases include quarterly business reviews, KPI dashboard presentations, financial audit findings, and scientific research publications.

Limitations: The data story structure exhibits low flexibility. If the audience lacks statistical fluency, the message fails. Additionally, the rigid structure limits emotional engagement—data stories inform but rarely inspire.

What Is a Story With Data? (The Pizza Model)

In a story with data, the narrative precedes the data. The author first determines the story arc, theme, and emotional trajectory; data is then inserted to strengthen or validate specific points.

Structural characteristics:

• Narrative as the crust: The story exists independently of the data. If the data were removed, the narrative would remain coherent, though less persuasive.
• Flexible sequencing: Data points can be rearranged, emphasized, or omitted based on narrative needs. The story dictates data placement, not the reverse.
• Sparse visualization: Charts are used sparingly, typically as illustrative support rather than primary evidence.

Applicable contexts:

Stories with data appeal to general audiences, including consumers, voters, and non-specialist stakeholders. Common applications include marketing campaigns, TED-style talks, news features, and corporate brand narratives.

Limitations: This approach carries higher risks of data manipulation. Without rigorous oversight, authors may selectively present data that supports the narrative while suppressing contradictory evidence. Additionally, stories with data rarely drive direct action—they generate awareness rather than decisions.

Core Differences At a Glance

| Dimension | Data Story (Sandwich) | Story With Data (Pizza) |
|-----------|----------------------|------------------------|
| Origin | Data-led | Narrative-led |
| Primary audience | Business decision-makers, analysts | General public, stakeholders |
| Structural rigidity | High (fixed sequence) | Low (flexible arrangement) |
| Visualization density | High (charts are central) | Low (charts are peripheral) |
| Action orientation | Strong (defensible decisions) | Weak (idea reinforcement) |
| Complexity | High (detailed, technical) | Lower (thematic, accessible) |
| Economic cost per unit | High (data prep + visualization) | Moderate (narrative design + selective data) |

Prevalence data: Stories with data appear approximately four times more frequently in corporate communications than data stories (Source 1: Analysis of 500 corporate presentations, 2023). This asymmetry reflects the broader audience reach and lower production costs of the pizza model.

When to Use Which: A Practical Decision Framework

The selection between frameworks depends on three variables: audience sophistication, decision proximity, and information complexity.

Use data stories when:

• The audience possesses statistical literacy (e.g., analysts, executives with quantitative backgrounds).
• The objective requires defensible, data-driven decisions (e.g., resource allocation, investment justification).
• The information complexity demands systematic deconstruction (e.g., multi-variable regression analysis, financial audits).

Use stories with data when:

• The audience lacks specialized data skills (e.g., consumers, employees from non-technical departments).
• The objective involves persuasion or awareness-building (e.g., brand positioning, cultural change initiatives).
• The information is inherently thematic and benefits from narrative framing (e.g., customer success stories, product origin narratives).

Case comparison:

A technology firm experiencing a 15% quarterly sales decline should use a data story for the board of directors: the data reveals the decline, analysis identifies the root causes (pricing elasticity, competitor entry, channel saturation), and the narrative follows the data to recommend specific corrective actions.

The same firm, when communicating with customers about product value, should use a story with data: the narrative centers on customer success, supported by selective data points (satisfaction scores, retention rates, feature adoption percentages). The data enhances the story but does not dictate it.

Mixing the frameworks carries risks. A data story injected with narrative elements that contradict the data undermines credibility. Conversely, a story with data overloaded with technical visualizations confuses the general audience. The boundary must remain clear.

Hidden Economic Logic: Efficiency vs. Scale

Beneath the structural differences lies an economic trade-off. Data stories require higher investment per unit—more time in data cleaning, visualization design, and logical verification. However, they yield higher decision quality per consumer, making them economically efficient for high-stakes, low-audience contexts.

Stories with data require lower investment per unit—narrative templates exist, and data can be sourced from existing dashboards. They achieve broader audience scale, making them economically efficient for low-stakes, high-audience contexts.

Empirical evidence: A 2022 McKinsey study found that organizations using data stories for internal decision-making improved decision accuracy by 34% compared to those using stories with data. Conversely, organizations using stories with data for external communications achieved 28% higher audience retention rates (Source 2: McKinsey Global Institute, "The Data-Driven Decision Advantage," 2022).

Conclusion: Structural Selection as Strategic Competence

The distinction between data stories and stories with data represents a fundamental structural choice in business communication. Neither framework holds universal superiority. Data stories optimize for precision and actionability; stories with data optimize for reach and engagement.

The strategic skill lies in recognizing that the choice is context-dependent. Organizations that develop clear criteria for framework selection—based on audience, objective, and complexity—will outperform those that default to one model or accidentally mix the two.

Industry prediction: Over the next three to five years, a third hybrid framework will likely emerge: the "layered narrative." This model will present a data story to expert decision-makers while simultaneously generating a parallel story with data for broader stakeholders, using AI-driven narrative generation to maintain structural integrity across both outputs. Organizations that invest now in understanding the fundamental distinction will be positioned to leverage this evolution.

The question is no longer whether to use data in storytelling—the question is which structural relationship between data and narrative serves the specific communicative objective. The answer determines credibility, clarity, and commercial outcome.