Chapter 30: Slide Design: Visual Communication for Presentations

"Tell me the facts and I'll learn. Tell me the truth and I'll believe. But tell me a story and it will live in my heart forever." — proverb of uncertain origin, widely quoted in presentation training; treat it as folklore, not a sourced attribution.

Chapter Overview

Dana Whitfield has the best result of her year. Her team's analysis shows that customer churn at the company is driven almost entirely by a slow first week — customers who take more than seven days to reach their first "aha" moment churn at seven times the rate of customers who get there in two — and the fix is cheap: a guided onboarding flow the product team could ship in a sprint. She has the data, the chart, and the recommendation. Next Tuesday she presents it to the leadership team, including Renée Okafor, the VP of Marketing who controls the budget that would pay for the work. So Dana does what almost everyone does: she opens her slide software, picks the corporate template, and starts pasting. Twenty-two slides later she has a deck that opens with an agenda slide, walks through her methodology, shows six dense charts with the company logo crowding each corner, and lists her findings as bullet points in 18-point gray text. She plans to talk through every bullet. On Tuesday she will lose the room before slide four, and the cheap, high-impact fix she discovered will die in a meeting — not because the analysis was wrong, but because the presentation of it failed.

This chapter opens Part VI, where the book leaves the page entirely for the room. You have spent twenty-nine chapters learning to write for a reader who controls the pace — someone who can reread a sentence, study a figure, flip back, set the document down and return. A slide audience can do none of that. They move at your speed, they cannot rewind you, and the instant a slide asks them to read a paragraph while your voice says something different, their attention splits and they lose both. If that sounds familiar, it should: Chapter 18 taught exactly this discipline for the research talk — the assertion–evidence slide, one point per slide, the death-by-PowerPoint failure mode. This chapter takes those ideas off the conference stage and makes them general. The same principles that save a fifteen-minute conference talk save a sales pitch, a board update, a design review, a thesis defense, a lunch-and-learn, and Dana's Tuesday meeting. Slide design is not a research skill or a business skill; it is a communication skill, and it rests on the same foundation this book has built since Chapter 2 — know your audience, lead with what matters, and make every element earn its place.

By the end of this chapter you will be able to build an assertion–evidence slide for any talk — a full-sentence claim as the title, a single visual as the body — and you will know the concrete limits that keep a slide legible: one point per slide, a roughly six-word ceiling on a topic title, a font floor you can read from the back of the room, and a contrast standard borrowed straight from Chapter 10. You will learn when animation earns its place (building a complex diagram one piece at a time) and when it is pure noise (bullets flying in from the left). You will learn to use a template without producing a slide that screams "template," and you will meet Guy Kawasaki's blunt, useful 10/20/30 rule for keeping a deck honest. Then you will redesign five genuinely bad slides, because the only way to learn this is to fix slides — yours and other people's.

In this chapter, you will learn to:

• Treat a slide as a visual aid, not a document — your voice carries the argument; the slide shows the one thing words can't.
• Design assertion–evidence slides for any talk — a sentence claim over a single visual — and recognize why topic-and-bullets slides cause death by PowerPoint.
• Apply concrete legibility limits — one point per slide, a ~6-word title ceiling, a readable font floor, and a 4.5:1 contrast minimum from the back row.
• Use animation only when it clarifies — building a diagram step by step helps; flying bullets and spinning text distract.
• Audit a deck with the 10/20/30 rule, redesign weak slides, and adapt a template so it doesn't look like one.

📘📕📗 All tracks: This chapter serves everyone, because everyone presents. Business and professional readers (📘): Dana's leadership deck is your home case — assertion–evidence is how you stop boring executives and start moving budgets. Engineering and science readers (📕): you met this in Chapter 18 for conference talks; here it generalizes to design reviews, all-hands, and demos, and Chapter 32 will extend it to architecture diagrams. Software readers (📗): the same ideas drive a good tech talk or a brown-bag. Wherever you present, the threshold concept below is the whole chapter — everything else is tactics.

30.1 A Slide Is a Visual Aid, Not a Document

Start where Chapter 18 started, because the failure is universal and it explains everything that follows.

You have suffered through it. A presenter stands at the front and projects a slide packed with text — full sentences, sub-bullets, a chart with a legend you can't read, the company logo in the corner — and then reads it to you. You read faster than they talk, so you finish the slide and wait, mildly bored. Meanwhile the words coming out of their mouth don't quite match the words on the screen, so part of your brain tries to reconcile the two and quietly gives up. Six slides in, you're answering email. The presenter might be brilliant and the content might be exactly what you needed; it doesn't matter, because the talk has already failed. This is death by PowerPoint, and it is not caused by the software or by a boring speaker. It is caused by a single misunderstanding about what a slide is for.

Here is the misunderstanding, and correcting it is the threshold of this entire chapter. People cannot read and listen at the same time. Reading text and listening to speech both draw on the same verbal channel in working memory; force them to compete and both degrade. When your slide shows one sentence and your voice says a different sentence, you have not given the audience two chances to absorb your point — you have given them two half-chances that interfere with each other. Reflexively, the audience reads the text in front of them, so the slide wins and your voice loses, and the thing you actually wanted to say evaporates. A text-dense slide does not support your talk. It competes with it, and it wins.

🚪 Threshold Concept: a slide is a visual aid, not a document. Before you cross this, a slide feels like a page you happen to project — a container for your content, complete and self-contained, so that anyone who zoned out could just read the slide and catch up. You build the slide to survive on its own. After you cross it, you see the slide as a visual aid for a talk you give with your voice. The spoken words carry the argument; the slide shows the single thing words can't — the graph, the photo, the diagram, the one big number. The audience reads OR listens, never both at once, so you strip the slide down to one claim and one visual and let your voice do the explaining. Writers who never cross this threshold build decks that work fine as handouts and fail completely as talks, because a document performed aloud is the recipe for death by PowerPoint. (If Chapter 18 already walked you across this line for conference talks, good — this is the same line, redrawn for every other kind of presentation you'll ever give.)

The fix follows directly from the diagnosis. If the audience can attend to your voice or your slide but not both, then the slide must show what your voice cannot — a picture, a relationship, a quantity made visible — and your voice must carry everything else. A slide is at its best when it puts up a single image the audience grasps in about two seconds, frees their verbal channel to listen to you explain it, and then gets out of the way. Notice that this is not a new idea: it is Chapter 10's thesis ("design is not decoration; it is part of the meaning") and Chapter 9's thesis ("a figure does not speak for itself; the surrounding interpretation makes it mean something") meeting a live audience and a clock. The slide is the figure; you are the caption.

Watch it at the level of one slide. Here is the opening "Findings" slide as Dana first built it, pasted straight from her written memo.

Slide (described): "Key Findings." Title in the corner, small: "Key Findings." The body is a six-item bulleted list in gray text: "• Overall monthly churn rate is 4.2%, up from 3.6% a year ago • Churn is concentrated in the first 30 days of the customer lifecycle • Customers who reach first value within 2 days churn at 3.1% • Customers who take more than 7 days to reach first value churn at 22% • Price sensitivity analysis shows weak correlation with churn (r = 0.08) • Onboarding completion rate is 61% across all cohorts." Beneath the bullets, the company logo and a slide number. Every line is a complete thought in eight-or-nine-point body text.

A listener facing that slide reads all six bullets in about ten seconds — faster than Dana can speak them — and then sits idle, having extracted the topic ("findings") but no message. The one finding that matters (7× churn for slow-onboarding customers) sits as the fourth bullet among six equals, no more emphasized than the onboarding completion rate. The slide is a paragraph wearing a list's clothing. It commits Dana to reading aloud, which guarantees the audience stops listening. Now the assertion–evidence version, which we'll build properly in the next section:

Slide (described), assertion–evidence: The title, a full readable sentence across the top, reads "Customers who onboard slowly churn 7× as often — and it's not about price." The body is a single bar chart filling the slide: two bars, "Reached value in 2 days" (3%) and "Took more than 7 days" (22%), the tall bar drawn in a strong color, the gap between them impossible to miss, each bar labeled directly with its value. No bullets. The other numbers (the 4.2% overall rate, the weak price correlation, the 61% completion) Dana says as she stands beside the one chart that carries the headline.

Now the audience reads one sentence — the claim — in two seconds, then looks at one picture that proves it while Dana's voice supplies the supporting detail. Even someone who glanced up from their laptop for a single moment leaves with the point. Same data. One slide competes with the speaker and buries the lede; the other serves the speaker and makes the lede unmissable. That contrast is the whole chapter in one pair of slides.

🔄 Check Your Understanding. A colleague defends his text-heavy deck: "I put all the detail on the slides so that people who can't attend the meeting can read the deck afterward and get everything." What's the flaw, and what's the right fix for his goal?

Answer

The flaw: he's optimizing one artifact for two incompatible jobs. A deck built so absent people can "read everything" is a document — and a document performed aloud causes death by PowerPoint, because the live audience reads the dense slides instead of listening to him. He's sacrificing the people in the room (who matter most) to serve the people who aren't. The right fix is to separate the artifacts: build clean assertion–evidence slides for the live talk, and produce a separate written document — a memo, a one-pager, or speaker notes exported as a leave-behind — for people who weren't there. One slide deck can't be both a teleprompter, a handout, and a report. (This is the same separation Chapter 18 reached: the slide aids the talk; the paper or memo stands alone.)

[📍 Good stopping point — from here on, the chapter is tactics for one principle: the audience listens at your pace and can't read and listen at once, so the slide shows what words can't and your voice carries the rest.]

30.2 The Assertion–Evidence Slide, Generalized

The single most useful technique in slide design comes from work by Michael Alley and colleagues on the design of scientific presentations, and it is called the assertion–evidence approach. Chapter 18 introduced it for research talks; here is the general statement, good for any presentation you will ever give: the title of each slide is a full-sentence assertion — the one claim you want the audience to take away — and the body is the visual evidence for that claim. Not a topic. Not a noun phrase. A sentence that says something.

Set it against the format almost everyone absorbs by osmosis from corporate templates: the topic-and-bullets slide. That slide has a short noun-phrase title ("Q3 Results," "Background," "Next Steps," "Methodology") and a body of bulleted text fragments. It is the format that produces death by PowerPoint, and it fails for three compounding reasons. First, the title names a topic, not a point — "Q3 Results" tells the audience the subject but not what to conclude, so they must extract the message themselves while also trying to listen, which they can't do. Second, the bullets are text, so they trigger reading and switch off listening. Third, fragments are ambiguous: "• churn up 4.2%" could mean a dozen things until the speaker explains it, so the slide can neither stand on its own nor stay out of the way. It is the worst of both worlds — too much text to ignore, too little meaning to be useful.

The assertion–evidence slide fixes all three at once. The sentence title states the point, so even an audience member who reads nothing but the title still gets the message — this is the inverted pyramid from Chapter 4, applied to a slide: the bottom line, on top. The body is a visual, so it doesn't fight your voice for the verbal channel; the brain processes images and speech through different channels, so the audience can study a chart and listen to you at the same time. And the slide now has exactly one job — back one claim — which is why it ends up clean.

Figure 30.1 (described): the assertion–evidence slide template. The slide divides into two zones. The top band (roughly the top fifth of the slide) holds the assertion: one complete sentence, left-aligned, in a large readable font, stating the claim — for example, "A guided first week would cut churn by an estimated 40%." The lower four-fifths is the evidence zone: a single visual — a chart, a labeled photo, a simple diagram, or one large number — that demonstrates the assertion, with labels placed directly on the visual rather than in a separate legend. There is little or no body text. The reader's eye lands on the sentence, drops to the picture that proves it, and is done. A whole talk is just a sequence of these: claim, proof; claim, proof; claim, proof.

The discipline this format imposes is itself the point, and it ties straight back to this book's first thesis. To write a sentence title, you have to know what the slide's point is. "Background" is easy to type because it commits to nothing. "Our customers stall in the first week, and that's where they churn" forces you to decide what this slide is for and what you want remembered. Presenters who convert their topic slides to assertions routinely discover that several slides had no point at all — they were just buckets of stuff — and those slides get cut or merged. Writing the assertion is how you find out whether the slide deserves to exist. That is Chapter 1 again, on a screen: phrasing the claim clearly is how you discover whether you have one.

Let me show the conversion on the slide presenters most often default to bullets for — the "agenda" or "overview" slide, and then a methods slide.

❌ Before (topic-and-bullets, an overview slide):

Slide (described): "Agenda." Title: "Agenda." Bullets: "• Background and context • Methodology • Key findings • Recommendations • Next steps • Q&A." Six fragments, centered, in the template's default body font.

A listener reads this in three seconds, learns that the talk has six parts (which they could have guessed), and gains nothing they'll remember. Worse, it spends the most valuable real estate in any talk — the opening — on an administrative table of contents instead of a hook.

✅ After (assertion–evidence):

Slide (described): Title sentence: "We can cut churn 40% with a one-sprint change — here's the evidence." Body: one large number, "7×," with a one-line label beneath it: "how much faster slow-onboarding customers leave." No bullet list. Dana opens on the point, not the procedure; the structure of the talk emerges as she gives it.

Why it's better: the opening slide now makes the promise of the whole talk — a big, cheap win is available, and the evidence is coming — instead of reciting its own table of contents. The audience knows in five seconds why they should keep listening. (If you genuinely need a roadmap slide for a long, complex talk, fine — but make even it earn its place, and never let it be slide one. The hook comes first.)

Now a methodology slide, where the bullet reflex is strongest:

❌ Before:

Slide (described): "Methodology." Title: "Methodology." Bullets: "• Analyzed 18 months of customer data (n = 47,000 accounts) • Defined churn as no activity for 30 consecutive days • Segmented by time-to-first-value • Controlled for plan tier, acquisition channel, and company size • Ran logistic regression on churn predictors • Validated on a held-out sample." Six dense fragments.

A listener reads it, learns the topic is "methodology," and gets no message — just a procedure list competing with whatever Dana is saying.

✅ After:

Slide (described): Title sentence: "We tested 47,000 accounts over 18 months — the pattern is real, not a fluke." Body: a simple funnel diagram showing 47,000 accounts narrowing through the analysis, with "n = 47,000" and "18 months" labeled directly on it. No bullet list. Dana says the rest — the churn definition, the controls, the regression, the held-out validation — because a listener can absorb those by ear while looking at the diagram.

Why it's better: the title now makes the only point a methods slide should make in a talk — we did this rigorously enough that you should trust the result — instead of naming the topic. The procedural detail that doesn't need to be read (definitions, controls, model choice) moves into Dana's narration, where a listener handles it comfortably. The slide and the speaker cooperate.

🧩 Productive Struggle. Before reading on, convert this topic slide to an assertion yourself. It's Dana's "Recommendations" slide, currently bulleted: "• Build a guided onboarding flow • Add in-app progress indicators • Trigger a check-in email on day 3 • Estimated engineering effort: one sprint • Estimated impact: 40% churn reduction in the target segment." Don't cram all five bullets into the headline. Pick the one point this slide must land, write it as a full sentence, and decide what single visual goes in the body. Then read on.

One strong answer

Headline: "One sprint of onboarding work could cut churn 40% — the cheapest win on the table." Body: a simple before/after bar showing projected churn in the target segment with and without the change, the gap labeled "40%," and a small "1 sprint" tag on the change. Notice what you had to do: choose the message (it's not "we have five recommendations"; it's "this is cheap and high-impact") and let the rest — the specific tactics, the progress indicators, the day-3 email — live in your spoken words or on a follow-up slide. A recommendations slide that lists five things equally makes the audience remember none of them. A recommendations slide that makes one point, backed by one picture, moves a budget.

The assertion–evidence approach carries a handful of corollaries worth stating plainly, because they're the rules you'll actually reach for:

• One point per slide. If your title needs an "and" joining two unrelated claims, that's two slides. A slide that makes two points makes neither memorable. (This is Chapter 6's one-idea-per-sentence discipline and Chapter 8's one-idea-per-paragraph discipline, applied to the slide as a unit.)
• Keep the assertion to about one line — a dozen-or-so words, readable in a glance. If it wraps to three lines, it's a paragraph, and you're back to making people read.
• Let the visual be the evidence, and label it directly. A separate legend forces the eye to ping-pong between key and chart. Put the labels on the bars, the arrows on the diagram, the one number in the center of the slide. (Chapter 9's "label directly, kill the legend" rule.)
• If a slide has no visual evidence, ask whether it should be a slide at all. Some claims are better spoken with a blank or near-blank slide than illustrated with a stock photo of a handshake. A decorative image is not evidence; it's chartjunk with better lighting.

🔍 Why Does This Work? Why does putting the claim in the title — rather than letting the audience infer it from the chart — make such a large difference? Think about the cognitive work you're asking of someone watching a talk.

Answer

Two reasons, both about reducing load on a listener who can't pause you. First, interpretation is work, and you're already using the audience's verbal channel. A bare chart titled "Q3 Churn" requires the viewer to study it, figure out what's notable, and decide what it means — all while also listening to you. That's a lot of simultaneous processing, and under that load people default to not doing the interpretive work, so they leave with shapes, not conclusions (exactly Chapter 9's "a figure does not speak for itself"). Putting the claim in the title does the interpretive work for them: they read one sentence, then look at the chart already knowing what to see in it, so the chart confirms rather than puzzles. Second, the title is what survives. Audiences forget almost everything in a talk, but a sentence headline is a complete, memorable thought; a topic word is not. If the only thing someone retains from a slide is its title, an assertion title leaves them with your point and a topic title leaves them with nothing.

30.3 The Six-Word Title and the One-Point Slide

Two limits do most of the work of keeping slides clean, and both are blunt enough to apply without judgment calls in the moment. The first is about titles; the second is about scope.

The six-word ceiling on a topic. Here is a tension worth resolving directly, because it confuses people who've just learned assertion–evidence. An assertion title is a full sentence and may run a dozen words — that's fine, because it's carrying the claim. But many slides, especially in business and engineering decks, still get topic titles by reflex ("Market Overview," "System Architecture and Integration Points and Future Considerations"). When you do use a topic phrase, cap it at roughly six words. Why six? Because a topic that runs longer has almost always crammed two ideas into one slide — "System Architecture and Integration Points and Future Considerations" is three slides pretending to be one. The six-word ceiling isn't about aesthetics; it's a tripwire. When your title won't fit in six words, that's the slide telling you it contains more than one point, and the right move is to split it, not to shrink the font. Treat the ceiling as a diagnostic: a topic title that overflows is a slide that overflows.

Compare:

❌ Before: Title: "Current Onboarding Process Challenges and Proposed Improvements with Projected Impact" (10 words). The slide beneath it is a wall of bullets covering three distinct things — what's broken, what to do, and what it would buy.

✅ After: Three slides, each with a one-line assertion title: "Our onboarding takes too long — and that's where customers leave." / "A guided first week fixes the slow start." / "The fix pays for itself in one quarter." The title that wouldn't fit in six words was hiding three slides.

Why it's better: the overflowing title was a symptom of an overflowing slide. Splitting on the seams the title revealed gives each idea its own air, its own visual, and its own moment in the talk — and it gives the audience three clean points to remember instead of one muddy one.

One point per slide. This is the corollary from the last section, promoted to a rule because it's the most violated rule in all of slide design. Every slide should make exactly one point — advance one claim, show one relationship, answer one question. The test is the same as the title test: if you can't say what the slide's single point is in one sentence, the slide doesn't have one point; it has zero or several. There is no penalty for having more slides. Slide count is free; audience attention is not. A common and false belief is that fewer slides means a shorter or tighter talk — but a deck of 10 dense slides that each make three points is far harder to follow than 25 clean slides that each make one, because the bottleneck is not slide count, it's points-per-minute the audience must absorb. Kawasaki's 10/20/30 rule (next section but one) caps slide count for a specific reason; do not confuse "few slides" with "good slides."

Here's the move in practice. Dana's original deck had a single slide titled "Findings" carrying six bullets. The fix is not to make that one slide prettier. It's to ask, how many points are actually here? — and the honest answer is about three: (1) churn is concentrated in the first week, (2) slow onboarding multiplies it 7×, (3) price isn't the driver. So that one overloaded slide becomes three assertion–evidence slides, each making one of those points with one visual, and the two minor numbers (overall rate, completion rate) get spoken, not slotted into a bullet. The deck gets longer and clearer at the same time. That is almost always the right trade.

⚠️ Warning — the "I only have 20 minutes so I'll use fewer, denser slides" trap. This reasoning is backwards and it's responsible for a huge share of bad talks. Cramming your content into fewer slides doesn't save time — you still have to say all of it — it just guarantees that each slide is too dense to follow, so the audience reads instead of listening and you lose them. Time is governed by what you say, not by how many slides you advance. A clean slide you spend ten seconds on costs the same ten seconds whether it's slide 5 or slide 25. Pace your talk by your script, not your slide count, and let the slides be as many and as simple as the points require.

🔄 Check Your Understanding. A slide's title reads "Performance and Cost." Without seeing the body, what's the most likely problem, and what should the presenter check?

Answer

The "and" is the tell: the title joins two distinct topics, which almost certainly means the slide is trying to make two points (a performance point and a cost point) on one slide — the most common one-point-per-slide violation. The presenter should check whether the body is making two separate claims; if so, split it into two assertion slides ("Throughput doubled after the rewrite" and "And it costs 30% less to run"). If the two genuinely are one point — say, "We got faster and cheaper, which is rare" — then write that as the assertion title and show one visual that captures the trade-off. Either way, "Performance and Cost" is a topic, not a point, and it should not survive as written.

30.4 Type, Color, and Contrast: The Back-Row Test

Everything so far has been about what goes on a slide. This section is about whether the people in the room can actually see it. The rules are nearly identical to Chapter 10's document-design rules, with one twist: a slide is read from across a room, often a large one, sometimes through a dim projector, by people who get exactly the seconds you spend on it. So the legibility bar is higher than for a document, and the diagnostic is simple.

The back-row test. Stand at the back of the room you'll present in — or, failing that, shrink your slide to about a third of your screen and step back six feet — and look at each slide. If you can't read the title and the labels on the visual without squinting, neither can the person in the back row, and you've just lost everyone past the third row. Most slide legibility failures are caught instantly by this one test, and almost nobody runs it.

The concrete defaults that pass the back-row test:

• Type size: big, then bigger. A slide title wants to be around 28–40 point; body or label text rarely belongs below about 24 point, and 18-point body text — the corporate-template default Dana inherited — is unreadable past the front row. The honest function of a large font floor is the same as the six-word title ceiling: it's a constraint that forces good behavior. You physically cannot fit a paragraph on a slide in 28-point type, which is exactly the point. If your text doesn't fit at a readable size, the answer is less text, never smaller type. (This is Chapter 10's "bigger type, shorter lines" rule, made stricter for distance.)
• Font: don't agonize. As Chapter 10 said, the serif-vs-sans question barely matters for legibility. A clean sans-serif (the kind your template ships with) is a fine default for projection. Pick one or two fonts for the whole deck and stop thinking about it. What ruins slides is not the wrong font; it's too much text in any font.
• Contrast: 4.5:1, checked, not eyeballed. This is Chapter 10's standard, and projectors make it more important, not less, because a projector washes out contrast that looked fine on your laptop. Dark text on a light background or light text on a dark background — either works, but the contrast must be high. The light-gray-text-on-white "modern minimalist" look that survives on a sharp monitor disappears entirely on a dim projector. Check the ratio with a tool (Chapter 10 named them); don't trust your bright laptop screen in a dark room.
• Color: meaning never rides on hue alone. Straight from Chapter 10, and just as binding here: roughly one in twelve men is red-green color-blind, and a projector can shift colors unpredictably. So never encode meaning with color alone — pair it with a label, a shape, a position, or direct text. A chart whose only distinction between "good" and "bad" is red versus green is illegible to a color-blind viewer and ambiguous to everyone in a badly color-calibrated room. Use a color-blind-safe palette (blue/orange, not red/green), and label the thing that matters directly. Use color sparingly and purposefully: one accent color to make the one bar that matters jump out is worth more than a rainbow that means nothing.
• Data-ink, on a slide. Chapter 9's Tufte principle applies with extra force on a projected chart: maximize the ink that represents data, delete the chartjunk. Heavy gridlines, 3-D effects, gradient fills, redundant legends, and the company logo crowding every corner all steal attention from the one shape you want seen and shrink the space the data gets. A slide chart should be cleaner than the same chart in a document, because the viewer has less time and worse resolution. Strip it to the data, the labels, and one accent color.

Watch the difference on Dana's churn chart:

❌ Before (described): A 3-D bar chart with six bars in six different rainbow colors, heavy gridlines, a legend in the bottom-right mapping colors to segment names, the y-axis starting at 3% (so the difference looks small), 12-point axis labels, and the company logo in the top-right corner. The viewer in row five sees a colorful blur and can't tell which bar matters.

✅ After (described): A flat bar chart with two bars — "2 days" (short, gray) and "7+ days" (tall, in the deck's one accent color, orange). No gridlines, no legend (each bar labeled directly with its value, "3%" and "22%"), y-axis starting at zero so the 7× difference is honest and dramatic, 28-point labels, no logo. The viewer in the back row sees one tall orange bar dwarfing a short gray one and reads "22% vs 3%" instantly.

Why it's better: the redesign removed everything that wasn't the comparison — colors, gridlines, the legend, the logo, the truncated axis — and emphasized the one thing that was (the gap between fast and slow onboarding), with a single accent color and direct labels readable from the back. The chart now passes the back-row test and tells its one truth at a glance. (Note that starting the bars at zero isn't just cleaner — it's honest; the truncated axis in the "before" understated a real, huge effect, which is the Chapter 9 / Chapter 38 ethics point in miniature.)

💡 Tip — the grayscale glance. Before you finalize a slide, view it in grayscale (your slide software can do this, or just imagine the color drained out). If two things that need to be distinguishable — "good" vs. "bad," "us" vs. "them," series A vs. series B — become indistinguishable in gray, you're encoding meaning with color alone, and a color-blind viewer or a grayscale handout sees exactly that gray ambiguity. Fix it with labels, shapes, or position before you add the color back.

30.5 Animation: When It Clarifies, When It Distracts

Animation has a bad reputation it half deserves. Used by reflex, it is the purest form of chartjunk in motion — bullets that fly in from the left one at a time, titles that spin into place, text that bounces, slides that dissolve through a checkerboard. None of it adds meaning; all of it steals attention from your voice and signals that the presenter is more interested in the software than the message. The default for most slides is no animation at all. A slide should simply appear, whole, and let the audience take it in while you talk.

But there is one job animation does better than anything else, and it's worth getting right: building a complex visual one piece at a time so the audience follows the construction instead of being hit with the finished tangle. This is the build animation, and its logic is the same concrete-to-abstract logic this book has used since Chapter 1 — show one idea, let it land, then add the next.

The clearest case is a diagram with several parts and a flow between them. Imagine Dana explaining how the proposed onboarding system works: a user signs up, hits a guided flow, sees progress indicators, gets a day-3 email, and reaches first value. If she puts the whole diagram — five boxes and four arrows — on screen at once, the audience's eyes dart around trying to find the entry point while she's still saying "so first, the user signs up," and they're lost. If instead each box and arrow appears as she narrates it — box one as she says "the user signs up," the arrow and box two as she says "then they enter the guided flow," and so on — the audience's attention is led exactly where her words are, and the diagram assembles in their understanding at the speed of her explanation. By the final click, they've watched the system get built and they get the whole thing. That is animation earning its place: it synchronizes what the audience sees with what they hear, which is the entire goal of slide design.

The same logic justifies a small set of other builds:

• Revealing a chart in layers — show the baseline trend, let it land, then reveal the second series that tells the real story ("and here's what happened after the change"). The reveal creates a beat that the all-at-once version loses.
• Walking down a list you genuinely must show (rare, but it happens — say, a numbered procedure) one item at a time, so the audience attends to the current item instead of reading ahead. Note this is the exception that proves the rule: if you've already moved your content out of bullets and into visuals, you rarely have a list to walk down.
• Highlighting — dimming everything on a busy slide except the one element you're discussing, then moving the highlight as you move through the parts. This focuses attention without changing the slide.

Here is the simple decision rule, and it's worth memorizing:

The test for any animation is one question: does the motion help the audience understand, or is it there to look interesting? If it synchronizes seeing with hearing — building a diagram as you explain it, revealing a result at the moment you say it — keep it. If it's there because the software offered it, cut it. Motion that doesn't carry meaning is the most distracting chartjunk there is, because the eye is wired to track movement and will helplessly follow a flying bullet right off the point you're trying to make.

🔄 Check Your Understanding. A presenter sets every bullet on every slide to fly in from the left, one click at a time, "so the audience doesn't read ahead." Separately, another presenter makes a five-box architecture diagram assemble one box per click as she explains each component. Both are using click-by-click animation — why is one wrong and one right?

Answer

The difference is whether the motion carries meaning. The first presenter is animating text — but the real problem isn't the animation, it's that the slide is bullets at all; flying them in one at a time just adds distracting motion on top of a format that already causes death by PowerPoint. The fix is to get rid of the bullets, not to animate them politely. The second presenter is animating a visual in sync with her narration: each box appears exactly when her words introduce that component, so the motion leads the audience's attention to the right place and the diagram builds in their understanding at the speed of her explanation. Same mechanism (click-by-click reveal), opposite effect — because one is decorating text the audience shouldn't be reading, and the other is constructing meaning the audience needs to see assembled.

📐 Project Checkpoint: the Presentation (Portfolio Piece 6, part 1). Your portfolio's sixth piece is a presentation — a slide deck plus speaker notes — and this chapter builds the deck; Chapter 31 will add the delivery and the notes. Choose the technical work you know best from an earlier portfolio piece: the strongest candidate is your data-analysis memo (Piece 4), because, like Dana, you already have a finding and a chart and now need to present it — but the technical report (Piece 1) or the proposal (Piece 2) work just as well. This chapter's concrete addition: build a 10-to-12-slide assertion–evidence deck for that work. For each slide, write a full-sentence assertion title (not a topic), choose the single visual that serves as its evidence, and run the back-row test on type size and contrast. Open on your point, not an agenda. Make exactly one point per slide; if a title needs an "and," split the slide. Reuse the cleaned-up charts from your Chapter 9 work — but rebuild any chart that doesn't pass the back-row test (kill the legend, label directly, strip to data-ink, one accent color). Add a build animation to at most one slide, and only if it has a multi-part diagram that genuinely benefits. Save the deck; in Chapter 31 you'll write speaker notes, rehearse it to time, and plan for the live room and Q&A. Preview: the deck you just designed is only half the piece — a great deck delivered badly still fails, and a modest deck delivered well can succeed, which is why delivery is its own chapter.

30.6 Templates That Don't Look Like Templates, and the 10/20/30 Rule

Two practical matters close out the craft: how to use a template without producing a slide that screams "template," and a famous rule of thumb for keeping a whole deck honest.

Templates are a gift — use them the Chapter 10 way

Chapter 10 said it about documents and it's just as true for slides: a good template is a gift. It has already made the decisions you'd otherwise make badly — the font pairing, the color palette, the title placement, the safe margins, often the accessibility basics like contrast. You should let the template own those pixels and spend your energy on the part only you can supply: the structure of meaning. "Generic" doesn't come from using a template; it comes from filling a template thoughtlessly — accepting its default agenda slide, its bulleted content layout, its stock-photo placeholders, and its "Click to add text" boxes without ever asking what your point is.

The distinction worth internalizing is template, not theme. Use the template for the surface (type, color, spacing, layout grid). Do not use it as a theme that dictates your content structure — the prefab "agenda / content / content / summary" skeleton, the bullet-list body layout, the decorative imagery. Override all of that with assertion–evidence:

• Throw out the bulleted-content layout and use a title-plus-one-visual layout (most templates have one, often called "title and content" or "picture"; if not, just delete the body box and place your visual).
• Replace topic titles with assertion sentences, even though the template's title box invites a short topic phrase.
• Delete decorative stock photos that aren't evidence; a clean slide with one real chart beats a "professional" slide with a smiling-team photo behind your text.
• Keep the logo small and in one corner, or drop it from interior slides entirely — it's chrome, not content, and Chapter 9's data-ink principle says chrome that crowds the data has to go.

Done this way, two presenters using the same corporate template produce completely different decks, because the template governed only the surface and each presenter supplied her own meaning. That's the goal: let the template make you consistent and legible, and make yourself clear.

The 10/20/30 rule

Guy Kawasaki, a venture capitalist who has sat through thousands of startup pitches, proposed a deliberately blunt rule for presentations, and while he framed it for pitches, it's a useful sanity check for almost any talk. The 10/20/30 rule says: a presentation should have about 10 slides, last no more than 20 minutes, and use a font no smaller than 30 point.

Take the three numbers as what they are — rules of thumb, not laws — but understand why each exists, because the reasoning is sound even when the exact number bends:

• 10 slides is a forcing function for focus, not a hard cap. Kawasaki's point is that no audience can hold more than about ten genuine ideas in a single sitting, so a pitch that needs forty slides is a pitch that hasn't decided what it's about. The number fights the instinct to include everything. (Caveat: a visual-heavy assertion–evidence talk may run more than ten slides while making only ten points, because some points need a build or a sequence — recall §30.3, slide count is free, points are not. Read "10 slides" as "about 10 points.")
• 20 minutes respects the audience's attention and, in a meeting, leaves room for the discussion that actually moves decisions. Even when your slot is an hour, the content should be tight enough to deliver in a fraction of it; the rest is questions and conversation. The number fights the instinct to fill the whole slot with talking.
• 30-point font is the same idea as our back-row test and our large-font floor: if you can't fit your text at 30 point, you have too much text. Kawasaki's blunt version of "less text, bigger type" — and the most quotable of the three because the font size forces the text discipline. (Our own floor was a touch lower for dense labels, but the spirit is identical: big type, little text.)

The rule's real value isn't the precise numbers; it's that it bundles this whole chapter into three memorable constraints. Few points, tight time, big type — chase those three and you cannot produce death by PowerPoint. When you finish a deck, run it against 10/20/30 as a final audit: roughly ten points or fewer? Deliverable in twenty tight minutes? Every word readable at 30 point from the back? If a deck fails all three, it's not a presentation yet; it's a document you're planning to read aloud.

🔍 Why Does This Work? The 10/20/30 rule is arbitrary on its face — why 10 and not 12, why 30 and not 28? Why does such a crude rule of thumb improve so many decks anyway?

Answer

Because the failure it's fighting is almost always too much, never too little. The universal slide-design disease is overload — too many points, too much time, too much text — and a crude rule that pushes hard in the "less" direction corrects the actual error even when its specific numbers are wrong for your case. Nobody ever ruined a talk by having too few ideas or type that was too large; the rule is calibrated against the real-world distribution of mistakes, which is lopsided toward excess. The precise numbers don't matter because they're not meant to be optimal — they're meant to be blunt enough to override your instinct to cram. That's also why "less text, bigger font" works as advice generally: the constraint physically prevents the overload, the way a small plate helps you eat less. A good rule of thumb isn't accurate; it's usefully biased against the common error.

30.7 Common Mistakes and Practical Considerations

The principles are simple; the failure modes are predictable. Here are the ones that recur most, with the fix for each.

The deck that's secretly a document. The deepest mistake, and the source of most others: building slides meant to be read (by people who missed the meeting, by your boss later, by your future self) instead of projected behind a speaker. The tell is dense slides and the excuse "but people need to be able to read it later." The fix is to separate the artifacts — clean slides for the room, a written leave-behind for everyone else. If you truly need one artifact that does both, build the document first and the slides second; never the reverse.

Reading your slides aloud. Even with clean slides, presenters often read the title sentence verbatim and then say nothing more. The slide's title states the claim; your job is to say what the slide can't — the why, the how, the caveat, the story behind the number. If your spoken words duplicate the slide, you've turned yourself into a narrator of your own teleprompter. Say more than the slide, not the same as it.

Charts pasted, not rebuilt. The chart you made in your analysis tool to find something is almost never the chart to show on a slide (Chapter 9's exploratory-vs-explanatory distinction). Pasted-in charts carry their legends, their default colors, their tiny labels, their gridlines, and their truncated axes onto a projector where none of it is legible. Rebuild every chart for the slide: one comparison, direct labels, one accent color, big type, axis at zero. Budget time for this; it's where slide quality actually lives.

Death by template. Accepting the template's content structure — the agenda slide, the bulleted layouts, the stock imagery — instead of just its surface. Override the structure with assertion–evidence; keep the surface (font, color, spacing).

The font-size death spiral. It starts innocently: the text doesn't fit, so you shrink it a point. Then it still doesn't fit, so you shrink it again. Three iterations later you're at 14-point and unreadable. Every time you're tempted to shrink the font to make text fit, do the opposite thing instead — cut the text. The font floor is non-negotiable; the text is always negotiable.

Meaning that rides on color alone. Red-for-bad, green-for-good, with no other distinction. Invisible to color-blind viewers, ambiguous on a miscalibrated projector, gone on a grayscale handout. Always pair color with a label, shape, or position. Run the grayscale glance.

Animation as entertainment. Flying bullets, spinning titles, checkerboard transitions. The eye chases motion, so gratuitous animation actively pulls attention off your point. Cut all of it except genuine build animation that synchronizes a visual with your narration.

Slide count anxiety. Cramming content into fewer, denser slides because "fewer slides = shorter talk." It doesn't; talk length is governed by what you say. More clean slides are easier to follow than fewer dense ones. Don't ration slides; ration points.

"It depends" — when the rules bend. Assertion–evidence is the right default for talks given to an audience that listens. A few situations legitimately differ. A pre-read deck that genuinely will be circulated and read silently (some executives and some all-remote teams work this way — Amazon famously uses written narratives instead of slides) is a document, and should be written as one, denser and self-contained — but then it isn't really a slide deck, and you shouldn't perform it aloud. A reference slide held in reserve for Q&A (a detailed table you flip to only if asked) can be dense, because it's not part of the narrated flow — it's an appendix. A regulated or templated context (some clinical, legal, or safety presentations have mandated formats) constrains your choices, and you comply, then make the best slides the format allows. The rule isn't "never put text on a slide"; it's "a slide you narrate should aid listening, not compete with it." Know which kind of slide you're making.

🪞 Learning Check-In. Pause and take stock — Part VI is about a skill many people fear. Think back to the last presentation you gave or sat through. Can you now name why it worked or failed in this chapter's terms — was it a document performed aloud, or a set of visual aids? Could you, right now, take one slide from a deck on your hard drive and convert it to assertion–evidence? If yes, the threshold concept has landed and the tactics are within reach. If the assertion–evidence idea still feels abstract, that's the signal to do the redesign exercises before you read Chapter 31 — designing slides is learned by redesigning slides, the same way writing is learned by rewriting. One honest question: when you build your next deck, will you start by opening the template and pasting, or by writing the one-sentence claim of each slide? The order you choose is the whole difference.

Frequently Asked Questions

How many words should be on a slide?

As few as the slide's one point requires — and the right framing is not a word count but a format. Replace the bulleted text with a full-sentence assertion title (a dozen words or so) plus a single visual, and the question mostly dissolves: a clean assertion–evidence slide has one sentence of text and almost no body text at all. If you find yourself counting words to see whether a bulleted slide is "too dense," that's the signal to abandon bullets and switch formats. The audience can't read and listen at once, so every word they read is a moment they don't hear you.

What is the assertion–evidence approach?

It's a slide design method (developed by Michael Alley and colleagues for scientific presentations, and general to any talk) where each slide's title is a full-sentence claim — the one thing you want remembered — and the body is a single visual that serves as evidence for that claim, with little or no bullet text. It replaces the default "topic-and-bullets" slide (a noun-phrase title over a list of fragments) that causes death by PowerPoint. The payoff: the title states your point so even a distracted viewer gets it, and the visual doesn't compete with your voice the way text does.

What is the 10/20/30 rule?

It's a rule of thumb from venture capitalist Guy Kawasaki: a presentation should use about 10 slides, run no more than 20 minutes, and use no font smaller than 30 point. It's deliberately blunt and aimed at pitches, but the spirit applies broadly — few points, tight time, big type. Take the exact numbers loosely (a visual-heavy assertion–evidence talk may use more than 10 slides while making only ~10 points), but treat the rule as a final audit against the universal slide disease, which is always too much.

Should I use animation in my slides?

Use it only when the motion helps the audience understand — almost always that means building a complex diagram or chart one piece at a time, in sync with your narration, so attention follows your words. Cut every other animation: flying bullets, spinning titles, bouncing text, and fancy slide transitions (cube, dissolve, checkerboard) add no meaning and actively steal attention, because the eye chases motion. The test is one question: does the motion carry meaning, or is it there to look interesting?

How do I make my slides not look generic when I have to use the company template?

Use the template for the surface (font, color, spacing, layout grid) and override its content structure with assertion–evidence. Throw out the prefab agenda and bulleted-content layouts, replace topic titles with sentence claims, delete decorative stock photos, shrink or drop the logo on interior slides, and put one real visual on each slide. "Generic" comes from filling a template thoughtlessly, not from using one — the template's job is to make you consistent and legible; your job is to supply the meaning. (Chapter 10 makes the same case for documents.)

Chapter Summary

Key Takeaways

• A slide is a visual aid, not a document. The audience reads OR listens, never both — so the slide carries the picture and your voice carries the argument. Building slides to be read causes death by PowerPoint.
• The assertion–evidence slide is the central technique, for any talk. Title = a full-sentence claim; body = one visual that proves it; almost no bullet text. It states your point and stops the slide from competing with your voice.
• One point per slide; a ~6-word ceiling on topic titles. If a title needs an "and" or won't fit in six words, the slide holds more than one point — split it. Slide count is free; audience attention is not.
• Pass the back-row test. Big type (titles ~28–40 pt, nothing important below ~24), 4.5:1 contrast checked not eyeballed, no meaning on color alone, charts stripped to data-ink with one accent color and direct labels.
• Animate only to clarify. A build that assembles a diagram in sync with your narration helps; flying bullets, spinning text, and fancy transitions distract — cut them.
• Templates are a gift; use the surface, override the structure. And run the deck against the 10/20/30 rule — few points, tight time, big type — as a final audit against overload.

Action Items

1. Open your next deck by writing the one-sentence assertion of each slide before you touch the layout. If you can't state the point, the slide has none.
2. Convert every topic-and-bullets slide to assertion–evidence: sentence title, one visual, kill the bullets.
3. Rebuild, don't paste, every chart for the slide: one comparison, direct labels, axis at zero, one accent color, big type.
4. Run the back-row test and the grayscale glance on every slide before you finalize.
5. Audit the finished deck against 10/20/30 and the one-point-per-slide rule; split any slide that fails.

Common Mistakes

• Building a deck that's secretly a document (and reading it aloud).
• Pasting charts instead of rebuilding them for projection.
• Shrinking the font to make text fit, instead of cutting the text.
• Encoding meaning with color alone.
• Gratuitous animation; accepting the template's bulleted content structure.

Decision Framework — what kind of slide am I making?

Spaced Review

A few questions reaching back, to strengthen retention.

1. (From Chapter 18) In the research-talk context, you learned the assertion–evidence slide and the "one point per slide" rule. State the threshold concept from that chapter in one sentence — the thing that explains why a text-dense slide fails — and explain how this chapter has done anything more than repeat it.
2. (From Chapter 9) Captions should climb to "Level 3 — interpretation": what the figure means, not just what it is. How is an assertion–evidence slide title the same move as a Level-3 caption, and what does that tell you about where a slide's "caption" lives?
3. (From Chapter 10, bridging) Chapter 10 gave you a 4.5:1 contrast standard, the "never color alone" rule, and "templates are a gift." All three reappear in this chapter. For each, name the one way the slide context makes the rule stricter or more important than it was for a printed document.

What's Next

You now have a deck that works on the screen — but a deck is only half of a presentation. Chapter 31, Delivering Technical Presentations, takes you from the slides to the room: how to structure the talk's opening hook and roadmap, how to rehearse until you land at ~90% of your time, how to speak from notes without reading them, how to manage nerves (and relabel them as energy), how to run a live demo without disaster, how to present convincingly over video, and how to handle the question period — including the hostile question and the one you can't answer. The assertion–evidence deck you designed here is the visual half; Chapter 31 is the human half. A great deck delivered badly still fails, and a modest deck delivered well can carry the day — which is exactly why delivery gets its own chapter, and why Dana's Tuesday meeting isn't won until she's rehearsed it out loud.

Practice: Exercises · Quiz Go deeper: Case Study · Case Study 2 Review: Key Takeaways · Further Reading