Storytelling With Data

Cole Nussbaumer Knaflic, a data visualization specialist and former member of Google's People Analytics team, argues that ineffective data visualizations are pervasive across industries because people are never taught how to combine numbers with narrative. Technology has made it easy for anyone to create a graph in applications like Microsoft Excel, but tools cannot identify the story in a dataset; that responsibility falls to the communicator. The book presents six core lessons, developed through Knaflic's workshops across a wide range of organizations, to help readers move from simply showing data to storytelling with data.

Knaflic traces the origins of her approach to her early career in credit risk management, where she discovered that investing time in visual presentation attracted more attention from leadership. She later joined Google's People Analytics team, which used data to inform human resources decisions around hiring, engagement, and retention. There, she helped communicate findings from Project Oxygen, a research initiative studying what makes a great manager, to audiences ranging from detail-oriented engineers to big-picture executives. When asked to develop internal training on data visualization, she researched the underlying principles and built a course rolled out across Google. Demand for her workshops beyond Google confirmed that the need for these skills is universal.

The first lesson concerns understanding context. Knaflic distinguishes between exploratory analysis, which she likens to opening a hundred oysters to find two pearls, and explanatory analysis, which focuses on communicating those specific pearls to an audience. Before creating any visual, communicators must answer three questions: Who is the audience? What should they know or do? How can data support the case? She stresses narrowing the audience as much as possible and urges communicators to make explicit recommendations rather than passively presenting data, arguing that even a wrong recommendation sparks productive conversation. She introduces the "Big Idea," a concept borrowed from communication expert Nancy Duarte's book Resonate, as a single sentence that articulates a unique point of view, conveys what is at stake, and proposes action. She also recommends storyboarding with Post-it notes to plan narrative structure without forming premature attachment to slides.

The second lesson addresses choosing an appropriate visual display. Knaflic argues that roughly a dozen visual types meet the majority of everyday needs. Simple text works best for one or two numbers; placing them in a graph dilutes their impact. Tables suit audiences scanning for specific values but should be avoided in live presentations because audiences read instead of listen. Heatmaps layer color saturation onto tables to speed identification of high and low values. Graphs fall into four categories: points (scatterplots for relationships between variables), lines (line graphs and slopegraphs, which compare values across two time periods), bars (vertical, horizontal, stacked, and waterfall charts for categorical data), and area charts for numbers of vastly different magnitudes. She champions the horizontal bar chart as the best default for categorical data because its left-to-right labels match natural reading order. She establishes a critical rule: Bar charts must always have a zero baseline, since omitting zero can visually inflate differences far beyond their actual magnitude. She warns against pie charts, donut charts, 3D effects, and secondary y-axes, all of which distort perception or confuse audiences.

The third lesson targets clutter, defined as visual elements that consume space without adding informative value. Knaflic introduces the Gestalt Principles of Visual Perception, a set of concepts from early 20th-century psychology describing how people organize visual stimuli. Six principles guide decluttering: proximity, similarity, enclosure, closure, continuity, and connection. Closure, for example, means people perceive incomplete shapes as whole, making chart borders unnecessary, while continuity means eyes follow the smoothest path, allowing axis lines to be removed. She advocates left-justified text over center-aligned for clean visual lines and argues that white space should be preserved rather than filled. A step-by-step decluttering demonstration removes chart borders, gridlines, and data markers; cleans up axis labels; and replaces legends with direct data labels.

The fourth lesson explains how preattentive attributes, visual properties processed by the brain's iconic memory before conscious thought, can direct audience attention and create a visual hierarchy. Knaflic demonstrates this with a "count the 3s" exercise: Without visual cues, scanning a block of numbers is slow, but rendering the target digits in bold dark color against light grey makes them instantly visible. She catalogs attributes including size, color hue, intensity, and spatial position, noting that some encode quantitative information while others serve as categorical differentiators. She recommends designing in shades of grey and using a single bold color for emphasis, offering five rules: Use color sparingly, use it consistently, design with colorblind users in mind, be thoughtful about the tone color conveys, and consider brand guidelines. She also notes that audiences scan pages in zigzag patterns starting at the top left, so the most important information belongs there.

The fifth lesson asks readers to think like designers, organized around three concepts: affordances, accessibility, and aesthetics. Affordances are aspects of a design that make its intended use obvious; in data visualization, this means highlighting at most 10% of a visual, eliminating distractions, and creating a clear hierarchy. Accessibility means making designs understandable to people of varying technical skills. Knaflic insists that every chart and axis needs a title and recommends action titles on slides that state the takeaway explicitly, such as "Estimated 2015 spending is above budget" rather than a generic label. She argues that aesthetics matter because research shows more attractive designs are perceived as easier to use and promote creative thinking. Three aesthetic priorities emerge: smart use of color, careful alignment, and intentional white space. She also offers strategies for gaining acceptance when audiences resist unfamiliar visual approaches, including side-by-side comparisons and enlisting vocal supporters.

The sixth lesson concerns storytelling. Knaflic draws on Aristotle's three-act structure and Robert McKee's framework for screenwriting to argue that stories unite ideas with emotions and that conflict is the critical ingredient. She contrasts narrative with conventional rhetoric such as bulleted slides, contending that facts alone do not inspire action. Her framework for data stories includes a beginning that introduces the setting and an imbalance requiring resolution, a middle that builds the case with data and articulates consequences, and an end that delivers a clear call to action. She introduces the "Bing, Bang, Bongo" approach to leverage repetition: Tell the audience what you will tell them, tell them, then tell them what you told them. Four tactics check story clarity: horizontal logic (slide titles alone tell the overarching story), vertical logic (all content on a slide reinforces its title), reverse storyboarding (writing down the main point of each finished page to verify structure), and seeking a fresh perspective from someone unfamiliar with the material.

Two chapters demonstrate cumulative application. One walks through recommending a price range for a new consumer product based on competitors' historical pricing. Knaflic transitions from a cluttered bar chart to an annotated line graph, constructs a nine-slide narrative for a live presentation, and ends with a specific pricing recommendation. Another chapter presents case studies on adapting color for dark backgrounds, using animation to reveal data progressively, establishing logical data ordering, untangling spaghetti graphs with too many overlapping lines, and replacing pie charts with clearer alternatives such as simple text, bar graphs, and slopegraphs.

Knaflic closes by positioning data visualization at the intersection of science and art. She offers five tips for continued growth: Master your tools, iterate and seek feedback, devote adequate time to the communication step, seek inspiration from good examples, and develop a personal style. For building organizational competency, she recommends foundational training for everyone, investment in one or two internal experts, and selective use of external consultants. She recaps all six lessons and encourages readers to find the story in their data and tell it clearly.