A Data Visualization Guide for Business Professionals
by Cole Nussbaumer Knaflic
The 60-Second Take
Storytelling with Data is Cole Nussbaumer Knaflic's practical guide to communicating with charts, drawn from her time at Google and in consulting. Its argument is that showing data and making a point are different tasks, and most business visuals default to the first. She organizes the work into six lessons covering context, chart selection, decluttering, directing attention, design thinking, and narrative, grounded in how visual perception actually operates. The book is software-agnostic and heavily worked through before-and-after examples rather than theory.
Your Software Doesn't Know What Your Point Is
Every charting tool has defaults, and those defaults are designed to display data accurately. None of them knows what you're trying to say, which is why a chart built by accepting the defaults tends to present twelve things with equal emphasis and communicate none of them.
Cole Nussbaumer Knaflic spent years on Google's people analytics team and in consulting before writing Storytelling with Data in 2015, and her central distinction is between showing data and communicating with it. The first is a display problem your software solves. The second is a communication problem that requires you to decide what matters, direct attention to it, and build a case around it.
The book is deliberately software-agnostic, organized into six lessons, and taught largely through worked examples where she rebuilds a real chart step by step. This summary covers the lessons and what makes them stick.
What You'll Learn
The context questions to answer before you build anything
Which chart types Knaflic recommends and which she avoids
How to identify and eliminate clutter, and why it costs your audience
What preattentive attributes are and how they direct attention
Why narrative structure applies to an analytical presentation
Lesson One: Understand the Context
Everything starts with two questions you should be able to answer in a sentence each: who is your audience, and what do you need them to know or do?
Knaflic is direct that this is where most analytical communication fails. If you cannot state the action you want, you're producing a report rather than making a case, and the audience will treat it accordingly. She pushes for specificity about who is actually in the room, what they care about, what they're skeptical of, and what authority they have to act.
Two tools operationalize it. The three-minute story: if you had three minutes in a corridor, what would you say? The Big Idea: a single declarative sentence that states your unique point of view, conveys what's at stake, and is a complete sentence. Both force the analytical work to resolve into a position rather than a summary.
She also distinguishes explanatory from exploratory analysis. Exploratory work is what you do to find the insight, and it involves hundreds of views of the data. Explanatory work is what you present, and it should contain the small number that support your point. Presenting your exploration is one of the most common causes of an unconvincing meeting.
Lessons Two and Three: The Right Visual, Without the Clutter
On chart selection, Knaflic favors a small set she finds reliably effective: simple text for one or two numbers, tables when the audience needs to look up specific values, heatmaps for tables with a magnitude dimension, scatterplots for relationships, line graphs for continuous data over time, slopegraphs for before-and-after comparisons across categories, bar charts, stacked bars, waterfalls, and a few others.
Her list of things to avoid is equally specific: pie charts and donuts, because humans judge angles and areas poorly; 3D effects, which distort without adding anything; and secondary y-axes, which invite misreading. Bar charts must have a zero baseline, because bar length is the encoding and truncating it misrepresents the comparison.
Then decluttering, which is where the biggest visible improvements come from. Clutter is anything occupying space without increasing understanding: gridlines, chart borders, redundant labels, data markers, decorative fills, legends that could be direct labels. Each element imposes cognitive load, and Knaflic draws the lineage explicitly from Edward Tufte's work on chartjunk.
She explains why this works through the Gestalt principles of visual perception, meaning the rules by which the brain automatically groups elements: proximity, similarity, enclosure, closure, continuity, and connection. These operate whether or not you design for them, so a chart that ignores them creates groupings you didn't intend. Alignment, white space, and consistent visual order do the rest, and her worked decluttering sequences, where a busy default chart becomes a clean one in six steps without changing a single number, are the most persuasive pages in the book.
Lesson Four: Direct the Attention
Once the noise is gone, you have to point at something.
Preattentive attributes are visual properties the brain processes before conscious attention: size, color, position, intensity, orientation. Something differing on one of these dimensions is noticed in milliseconds, without the viewer reading anything.
The practical rule is restraint. Color works as a signal only when used sparingly, so the effective technique is a chart mostly in gray with one element highlighted. Color everywhere signals nothing. Knaflic's guidance is to highlight a small fraction of the visual, and to layer attributes deliberately: bold text for the important label, size for the important series, position for the thing you want seen first.
Underneath this is visual hierarchy: an audience should be able to tell within seconds what to look at first, second, and third. If everything has the same weight, they choose their own order, which is not the one you wanted.
Lesson five, thinking like a designer, extends this into affordances, meaning making the intended use obvious; accessibility, meaning designing for the least expert person in the room; aesthetics, since well-designed things are perceived as easier to use; and acceptance, since audiences accustomed to the old chart may need to be brought along.
Lesson Six: Tell a Story
The final lesson is where the book's title comes from, and it's the least technical.
Knaflic argues that narrative structure applies to analytical communication as much as to any other kind. A presentation with a beginning that establishes the situation, a middle that develops the tension between what's happening and what should be happening, and an end that resolves into a call to action holds attention that a sequence of findings does not.
The practical techniques are storyboarding before you build anything, ideally with sticky notes rather than in software so that reordering stays cheap; deliberate repetition of the central point, since audiences need it more than once; and thinking about the order of revelation, including whether to lead with your conclusion or build to it depending on the audience and the stakes.
Her point about tension is the one analysts most often skip. A story without conflict has no reason to be listened to, and in a business context the conflict is usually the gap between the current trajectory and the desired one. Stating that gap explicitly is what turns data into an argument.
Storytelling with Data at a Glance
The six lessons. Understand context, choose an appropriate display, eliminate clutter, focus attention, think like a designer, tell a story.
The Big Idea. A single declarative sentence stating your point of view and what's at stake.
Exploratory versus explanatory. The many views you built to find the insight versus the few you present to make the case.
Clutter. Anything occupying space without increasing understanding, which imposes cognitive load on your audience.
Gestalt principles. The automatic grouping rules the brain applies, which operate whether or not you design for them.
Preattentive attributes. Size, color, and position, processed before conscious attention and therefore usable to direct the eye.
A Quick Start Guide to Better Data Communication
Write the Big Idea first. One sentence stating your point and what's at stake, before you open any charting tool.
Name the action. State what you want the audience to do; if you can't, you're reporting rather than communicating.
Declutter aggressively. Remove gridlines, borders, legends, and markers, and replace legends with direct labels.
Gray everything, then highlight one thing. Color used once directs attention; color used everywhere directs none.
Storyboard on paper. Sequence your argument before building slides, while reordering still costs nothing.
Who Should Read Storytelling with Data (and Who Can Skip It)
Read it if you present analysis to decision-makers and suspect your charts aren't landing, which is the exact audience it's written for.
Read it if you work in finance, analytics, consulting, or research and were never taught anything about visual communication, which is most people in those roles.
Read it if you want tool-agnostic principles rather than software instructions, since nothing here expires when your stack changes.
Skip it if you're a trained designer or a data visualization specialist. The principles are foundational and you'll know them.
Skip it if you need advanced or interactive visualization guidance. The book targets the ordinary business chart, not dashboards, geospatial work, or complex multivariate displays.
Skip it if you want statistical rigor. This is about communicating a finding, and it assumes the analysis behind it is already sound.
Final Reflections
The reason this became the standard reference is that it addresses a gap almost nobody's education covers. Analysts are trained extensively in producing correct numbers and not at all in communicating them, and the resulting default, which is to show everything and let the audience decide, undermines a great deal of otherwise good work.
The decluttering material produces the fastest visible improvement, and the before-and-after sequences are genuinely convincing in a way that description isn't. The context chapter is the more important one, though, because it operates before any chart exists and forces the question of what you're actually asking for.
The Bottom Line
Your software displays data; making a point is your job. Decide what you want the audience to do, strip out everything that doesn't help them see it, and use color once to show them where to look.
Frequently Asked Questions
What are the six lessons in Storytelling with Data?
Understand the context, choose an appropriate visual display, eliminate clutter, focus your audience's attention, think like a designer, and tell a story. Each builds on the previous one, starting before any chart is created.
What are preattentive attributes?
Visual properties such as size, color, and position that the brain processes before conscious attention. Because an element differing on one of these is noticed almost instantly, they're the mechanism for directing an audience's eye to the point you want made.
Why does Knaflic advise against pie charts?
Because people judge angles and areas poorly compared with lengths, which makes pie charts hard to read accurately, especially when segments are close in size. She generally recommends a bar chart or simple text instead.
Business Floss is reader-supported. When you use our links we may earn an affiliate commission that helps us keep the site running. Thank you for your support!