Chapter 18: Conference Presentations and Posters

"The slides are not the presentation. You are the presentation. The slides are visual aids." — a maxim repeated in countless presentation-skills workshops; the sentiment is folklore among professional speakers, not a single sourced quotation.

Chapter Overview

Dr. Lena Foss has a paper. After three years on the electrochemistry of solid-state batteries, she has a result she is genuinely proud of — a new electrolyte interphase that holds 94% capacity after 800 cycles, where the prior best held 80% after 500 — and a manuscript working its way through peer review. Now she has been accepted to give a fifteen-minute talk at the field's flagship conference, and she will also stand beside a poster for two hours in a crowded hall. She has done what most researchers do: she opened her slide software, started pasting figures and paragraphs from the paper, and produced thirty slides dense with text, equations, and eight-point captions. She plans to read them. In fifteen minutes she will lose the room, run over time, and watch the session chair hold up a card that says 2 minutes. The science is excellent. The communication of the science is about to fail — and, unlike a rejected paper, this failure happens live, in front of the exact people whose attention she most wants.

This is the chapter where everything you learned about visuals (Chapter 9) and structure (Chapter 4) leaves the page and meets a clock and a crowd. A conference is a different medium with a different physics. A reader of your paper controls the pace: they can reread a sentence, study a figure, flip back, set the paper down and return. A listener at your talk can do none of that. They move at your speed, not theirs; they cannot rewind you; and the moment your slide asks them to read a paragraph while you say something different, you have split their attention and they have lost both. This chapter previews the deeper presentation craft of Part VI — slide design in Chapter 30, delivery in Chapter 31 — but here we stay inside the research context: the conference talk, the research poster, the lightning talk, and the question period that follows. The discipline is the same one this book has taught since Chapter 2 — know your audience, lead with what matters, make every element earn its place — applied to an audience that owes you nothing and will not wait.

By the end of this chapter you will be able to build an assertion–evidence slide — the single most important design idea here, where a slide's headline is a full-sentence claim and its body is one supporting visual, not a list of bullets. You will know how to structure and time a talk so you finish on the clock, how to lay out a poster so a stranger reads it in the right order and leaves with one idea, how to deliver the sixty-second elevator pitch that turns a passerby into a conversation, how to compress a talk into a five-minute lightning version, and how to field a hostile or confused question without coming apart. This is intermediate material; it assumes you can already interpret a figure (Chapter 9) and lead with your conclusion (Chapter 4). Here those skills meet a room.

In this chapter, you will learn to:

• Treat the live audience as your real reader — they listen at your pace, can't rewind, and split their attention the instant a slide makes them read.
• Design assertion–evidence slides — a sentence-headline claim over a single visual — and recognize why bulleted "topic" slides cause death by PowerPoint.
• Structure and time a talk (and a 5-minute lightning talk) so you land on the clock and a listener can follow without notes.
• Lay out a research poster with one takeaway, a clear reading path, and an elevator pitch that works in sixty seconds.
• Handle Q&A — including the hostile, the confused, and the question you can't answer — with composure.

📕 Engineering/Science Track: This is a core chapter for you, and it closes Part III's arc from page to podium. The Specific Aims logic you built in Chapter 17 — hook, gap, what I did, why it matters — is exactly the spine of your talk's first minute and your poster's elevator pitch. If you'll later take the Presentations track (Part VI), read this as the research-specific foundation; Chapter 30 generalizes assertion–evidence to any deck, and Chapter 31 goes deep on delivery and nerves. Here, learn the conference versions first.

18.1 A Talk Is Not a Paper Read Aloud

Start with the failure, because the failure explains the whole chapter.

The single most common mistake in research presentation has a name in workshops: death by PowerPoint. You know it from the receiving end. A speaker stands at the front and projects a slide crammed with text — full sentences, sub-bullets, a dense figure, a caption in eight-point font — and then reads it to you. You can read faster than they can talk, so you race ahead, finish the slide, and now you're bored and slightly ahead of the speaker, waiting. Meanwhile they're saying words that don't quite match the words on screen, so part of your brain tries to reconcile the two streams and gives up. Twelve slides in, you've stopped listening entirely and you're checking your phone. The speaker may be brilliant. The talk has failed, and it failed for a reason rooted in how attention works, not in how good the science is.

Here is that reason, and it is the threshold of this entire chapter. People cannot read and listen at the same time. Reading and listening both consume the same verbal channel in working memory; force them to compete and both degrade. When your slide shows a sentence and your mouth says a different sentence, you have not given the audience two chances to get your point — you've given them two half-chances that interfere. The slide wins (people reflexively read text in front of them), your voice loses, and the thing you actually wanted to say evaporates. A slide dense with text doesn't support your talk; it competes with it.

🚪 Threshold Concept: a slide is not a document. Before you cross this, a slide feels like a page you happen to project — a place to put your content, complete and self-contained, so that if someone missed your words they could read the slide instead. After you cross it, you see a slide as a visual aid for a talk that you give with your voice: the spoken words carry the argument, and the slide shows the one thing words can't — the picture, the graph, the diagram, the number. The audience reads OR listens, never both at once. Once you internalize this, you stop writing slides and start designing them: you strip the text down to a single headline claim, hand the body over to a visual, and let your voice do the explaining. Writers who never cross this threshold produce slide decks that work fine as handouts and fail completely as talks — because they built a document and then tried to perform it.

The fix follows directly. If the audience can attend to your voice or your slide but not both, then the slide must show what your voice cannot — and your voice must carry everything else. A slide is at its best when it displays a single image the audience grasps in two seconds (a graph, a photo, a diagram, a big number), frees their verbal channel to listen to you explain it, and then gets out of the way. That is the design principle this whole chapter elaborates.

Watch the difference at the level of one slide. Here is Lena's opening "results" slide as she first built it — pasted from her paper.

Slide (described): "Results." Title in the corner reads "Results." The body is a bulleted list: "• Novel sulfide-based electrolyte interphase synthesized via solution coating • Capacity retention of 94% measured after 800 charge-discharge cycles at 0.5C • Comparison to baseline (LiPON) which retained 80% after 500 cycles • Coulombic efficiency improved from 99.1% to 99.7% • Interfacial resistance reduced by approximately 40% • Cross-sectional SEM (inset, bottom right) shows uniform 50 nm interphase layer." Six lines of small text plus a thumbnail electron-microscope image too small to read.

A listener facing that slide reads all six bullets in about eight seconds, faster than Lena can speak them, and then waits — half-bored, the SEM image too small to see, the one number that matters (94% vs. 80%) buried as the second bullet among equals. The slide is a paragraph in disguise. Now the assertion–evidence version, which we'll build properly in the next section:

Slide (described), assertion–evidence: The headline, in a full readable sentence across the top, reads "Our electrolyte holds 94% capacity after 800 cycles — the previous best held 80% after 500." The body is a single line chart filling the slide: two curves of capacity retention vs. cycle number, Lena's curve staying high and flat far to the right, the baseline curve dropping off early, each curve labeled directly on the plot, the gap between them shaded. No bullets. The supporting numbers (coulombic efficiency, resistance) are spoken, not shown.

Now the audience reads one sentence — the claim — in two seconds, then looks at one picture that proves it while Lena's voice supplies the detail. The point is unmissable even to someone who glanced up from their phone for a moment. Same data. One competes with the speaker; the other serves her.

🔄 Check Your Understanding. A presenter defends a text-heavy slide: "I put everything on the slide so that people who get distracted can catch up by reading it." What's wrong with this reasoning?

Answer

It guarantees the very distraction it's trying to fix. A slide full of text forces the audience into reading, which shuts off their ability to listen — so the moment they start reading the "catch-up" slide, they stop hearing the speaker, and the speaker's voice (which carries the actual argument) is lost. You can't give people a reading fallback and a spoken argument at the same time; the reading wins and the talk loses. If catching up matters, the fix is a handout or the paper itself (a document, distributed separately), not a slide that doubles as one. The slide's job is to aid the live talk, not to survive without it.

[📍 Good stopping point — the rest of 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.]

18.2 The Assertion–Evidence Slide

The most useful single technique in research presentation comes from work by Michael Alley and colleagues on the design of scientific slides, and it is called the assertion–evidence approach. The idea is small and changes everything: the headline of each slide is a full-sentence assertion — the claim you want the audience to take away — and the body of the slide is the visual evidence for that claim. Not a topic. Not a noun phrase. A sentence that says something.

Contrast it with the default most people learn by osmosis from corporate templates: the topic-and-bullets slide. That slide has a short noun-phrase title ("Results," "Methods," "Background," "Discussion") 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 states a topic, not a point — "Results" tells the audience the subject but not what to conclude, so they have to extract the message themselves while also listening, which they can't. Second, the bullets are text, so they trigger reading and shut off listening. Third, fragments are ambiguous: "• 94% retention, 800 cycles" could mean almost anything until the speaker explains it, so the slide can't stand on its own and it interferes with the explanation. The format is the worst of both worlds.

The assertion–evidence slide fixes all three. The sentence headline states the point, so a listener who only reads the headline still gets the message — this is the inverted pyramid (Chapter 4) applied to a slide: the bottom line, up top. The body is a visual, so it doesn't compete with your voice for the verbal channel; the audience can look at a graph and listen to you simultaneously, because images and speech use different channels. And the slide now has exactly one job — support one claim — so it's clean.

Here is the anatomy, described as a figure:

Figure 18.1 (described): the assertion–evidence slide template. The slide is divided into two zones. The top ~15% is a single horizontal band containing the assertion: one complete sentence, left-aligned, in a large readable font (roughly 24–28 point), stating the claim — e.g., "Adding the sulfide interphase doubles cycle life at no cost to energy density." The bottom ~85% is the evidence zone: one visual — a graph, a labeled photo, a diagram, or a single large number — that demonstrates the assertion, with direct labels on the visual itself and no separate legend where one can be avoided. There is little or no body text. The eye lands on the sentence, drops to the picture that proves it, and moves on. A whole talk is a sequence of these: claim, proof; claim, proof.

The discipline the format imposes is itself valuable, and it connects to this book's first thesis. To write a sentence headline, you have to know what the slide's point is. "Results" is easy to type because it commits to nothing. "Our electrolyte holds 94% capacity after 800 cycles — the previous best held 80% after 500" forces you to decide what this slide is for and what you want remembered. Many presenters discover, while converting topic slides to assertions, that several of their 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's idea again: the act of phrasing the claim clearly is how you discover whether you have one.

Let me show the conversion on a methods slide, where presenters most often default to bullets.

❌ Before (topic-and-bullets):

Slide (described): "Methods." Title: "Methods." Bullets: "• Solid-state cells assembled in argon glovebox • Sulfide electrolyte deposited by solution coating at 60 °C • Cycled at 0.5C between 2.5 and 4.2 V • Capacity measured by galvanostatic cycling • SEM and XPS used for interphase characterization • n = 12 cells per condition." Six fragments of dense text.

A listener reads this in seconds, learns the topic is "methods," and gleans no message — there's no claim here, just a procedure list competing with whatever Lena is saying.

✅ After (assertion–evidence):

Slide (described): Headline sentence: "We built and stress-tested 12 solid-state cells per condition — a sample size big enough to trust the difference." Body: a clean schematic diagram of the cell stack (anode / electrolyte-interphase / cathode), with a small inset photo of the coin cells and the cycling window (2.5–4.2 V, 0.5C) labeled directly on the diagram. No bullet list. Lena says the rest — the glovebox, the deposition temperature, the characterization tools — because those are details a listener can absorb by ear while looking at the picture.

Why it's better: the headline now makes a point a methods slide can make — we did this rigorously enough that you should believe the comparison — instead of just naming the topic. The body is a diagram the audience grasps at a glance, not text that fights Lena's voice. And the procedural details that don't need to be read (temperatures, voltages, tool names) move into her spoken narration, where a listener handles them comfortably. The slide and the speaker now cooperate instead of competing.

🧩 Productive Struggle. Before reading on, take this topic slide and write its assertion headline yourself: a "Background" slide whose bullets are "• Solid-state batteries promise higher energy density and better safety • Main barrier is poor interface stability between electrolyte and electrodes • Existing interphase materials degrade after a few hundred cycles." Don't summarize all three bullets — pick the one point this slide should make to set up Lena's work, and write it as a full sentence. Then read the next paragraph.

One strong answer

"Solid-state batteries stall on one problem: the electrolyte–electrode interface falls apart after a few hundred cycles." That single sentence names the field's promise implicitly (we wouldn't care otherwise), states the specific barrier, and sets up Lena's contribution (a better interphase) — all as one claim the audience takes away. The body becomes a simple diagram or micrograph showing interface degradation. Notice what you had to do: choose. A background slide can't make three points in a talk; it makes one, the one that makes the next slide land.

The assertion–evidence approach has a few corollaries worth stating plainly:

• One point per slide. If your headline needs an "and" joining two unrelated claims, you have two slides. A slide that makes two points makes neither memorable.
• Keep the assertion short enough to read in one glance — roughly a single line, a dozen-or-so words. If it wraps to three lines, it's a paragraph, and you're back to reading.
• Let the visual be the evidence, and label it directly. A separate legend forces the eye to ping-pong; put the labels on the curves, the arrows on the diagram, the one number in the middle of the slide.
• Almost no body text. A short axis label, a data callout, a key word — fine. A bulleted list — no. If you find yourself writing sentences in the body, you're writing a document.

🔍 Why Does This Work? Why does a sentence headline beat a topic headline, when the audience will hear you explain the slide anyway? Because of how memory encodes a talk. A listener can't hold your whole argument in working memory; they retain a sequence of points. If each slide's headline is already the point — a complete claim — then the slide does the encoding work for them: glance up, read one sentence, store one claim, move on. If the headline is a topic ("Results"), the listener has to derive the point from your speech in real time and store that — extra cognitive work, done while also trying to follow your next sentence, and often failed. The sentence headline is a gift to the audience's memory: you've pre-digested the slide into the exact thing you want them to walk out remembering. Topics make them work; assertions hand them the takeaway.

🔄 Check Your Understanding. What is the single most reliable test of whether a slide follows the assertion–evidence approach — something you can check in two seconds by looking at it?

Answer

Read only the headline. If it's a complete sentence that makes a claim ("Our interphase doubles cycle life"), it's an assertion. If it's a noun phrase or topic ("Results," "Cycle Life," "Battery Performance"), it isn't — and almost certainly the body is bullets. The headline test catches the failure instantly: a topic headline is the tell of a topic-and-bullets slide. A secondary check: is the body a single visual (pass) or a list of text (fail)? Sentence headline + one visual = assertion–evidence.

18.3 Structuring and Timing the Talk

A research talk has a shape, and it's a familiar one, because it's the spoken cousin of the structures you already know. The audience is a listener doing a task — figure out what this person did and whether it matters — at your pace, with no ability to skip ahead or back. That single constraint dictates the structure: you signpost relentlessly, you front-load the point, and you repeat the through-line, because a listener who gets lost can't flip back to find their place.

The arc of a talk

A clear research talk moves through five beats. Think of them as the IMRaD of Chapter 13 reshaped for the ear and compressed for the clock:

1. The hook and the question (the first 60–90 seconds). Open with the problem and why it matters — concretely, fast — and state the specific question you set out to answer. This is the Specific Aims hook from Chapter 17 spoken aloud: here's the problem, here's the gap, here's what I did. Do not open with "Hi, I'm going to talk about…" and an outline of your outline; open with something that makes the room want to know the answer. The single biggest delivery upgrade most researchers can make is to put their result — or at least the promise of it — in the first ninety seconds, not the last ninety.
2. The roadmap (one slide, ten seconds). Tell them the path: "I'll show you the problem, our approach, the key result, and what it means." This is signposting (Chapter 4) — a listener needs to know the shape of the journey because they can't see how many pages are left. One line, then move.
3. The approach (the middle). What you did, at the level of detail a listener can follow by ear while looking at your visuals. This is where presenters over-share: you do not need every method, control, and parameter — you need the ones required to trust the result. Show the design, not the protocol.
4. The key result (the peak). The one or two findings that matter, each on an assertion–evidence slide. This is the part you most want remembered, so give it air: state the claim, show the evidence, pause, and say what it means. Don't bury the headline result in a thicket of secondary findings — pick the peak and let the audience stand on it.
5. The takeaway and implications (the last 60 seconds). What it means, what's next, the one sentence you want them to repeat to a colleague at coffee. End on the point, not on "so, yeah, that's it" or a wall-of-text "Conclusions" slide. Your last slide should not be "Thank you / Questions?" with nothing on it — make your final slide restate your single takeaway, so it's on screen during the entire Q&A.

Notice what's missing from a talk that a paper has: the exhaustive literature review, the full methods, the complete results, the caveats and limitations spelled out at length. Those live in the paper. A talk is not a compressed paper; it's a trailer for the paper — it makes people want to read the real thing and remember your one result. The Results-versus-Discussion discipline from Chapter 13 still applies (show the finding, then interpret it), but the volume is a fraction.

Timing: the discipline that respects the room

Running over time is the rudest thing a speaker can do, and one of the most common. It steals from the next speaker, annoys the chair, and signals that you didn't respect the audience enough to practice. The fix is mechanical and non-negotiable:

• Budget roughly one slide per minute as a starting estimate — a rough heuristic, not a law (a single data slide you dwell on might take three minutes; a section divider takes ten seconds). For a 15-minute talk, that's a ceiling of around 15 slides, usually fewer. If you've built thirty slides for fifteen minutes, you've built a talk you cannot give.
• Practice out loud, with the clock, at least three times. Not in your head — out loud, because you speak slower than you think and your in-head rehearsal lies to you. Time it. Most first-timers run long.
• Aim to finish at about 90% of your allotted time. Land a 15-minute talk in 13–14 minutes. This leaves margin for the inevitable: a slide that takes longer, a moment you lose your place, a clarifying aside. A talk timed to exactly fill the slot will run over, because live talks expand. Finishing early is a gift; finishing late is a theft.
• Know your cut points in advance. Mark two or three slides you can skip if you're running behind, so you can drop them gracefully mid-talk instead of speeding up into an unintelligible blur. The session chair's "2 minutes" card should trigger a plan, not panic.

⚠️ Warning: the "I'll just talk faster" trap. When you realize mid-talk that you're running long, the instinct is to accelerate. Don't. Speaking faster makes you less intelligible exactly when you most need the audience with you — at the result. The right move is to cut: skip a pre-planned cut slide, drop a secondary finding, jump to your takeaway. Better to deliver 80% of the content clearly than 100% of it as an auctioneer. Plan the cuts before you stand up.

🔄 Check Your Understanding. You've built a 15-minute talk and your practice run comes in at 19 minutes. A colleague suggests "just talk a bit faster and tighten the transitions." Why is that the wrong fix, and what's the right one?

Answer

Talking faster degrades comprehension precisely where it matters and still rarely recovers four whole minutes — 19→15 is a 21% cut, far more than pace can buy. The problem isn't your speed; it's that you have too much content for the slot. The right fix is to cut: remove slides (you likely have more than the ~one-per-minute budget), drop secondary results, shorten the background, and trim the roadmap. Cut the talk down to a 13–14-minute version (90% of 15) at a normal speaking pace, then practice that. The audience remembers a clear subset, not a rushed superset.

18.4 The Research Poster

A poster is a different animal from a talk, and people mis-design it by treating it like either a paper (too dense) or a slide deck (too sparse). A poster is a static visual document that you stand next to — it has to work two ways at once: it must draw a passerby from across a noisy hall in three seconds, and it must reward a genuinely interested person who stops to read for two minutes while you talk them through it. Those two jobs pull in opposite directions, and good poster design resolves the tension with hierarchy.

The one-takeaway rule

The first question to answer before you lay out anything: what is the single thing you want someone to remember after they walk away? A poster has one takeaway, the way a slide has one assertion. Everything on the poster either delivers that takeaway or supports it. A poster that tries to convey your entire paper conveys nothing; the hall is too crowded, the visit too short, and the competing posters too many. Pick the one finding. Lena's poster takeaway is the same as her talk's: a sulfide interphase nearly doubles solid-state battery cycle life. That sentence, in some form, should be readable from across the room.

Hierarchy and the reading path

The most common poster failure is the wall of text: a poster that is the paper, shrunk and pasted into columns, with paragraphs of ten-point prose nobody standing up in a crowd will ever read. The fix is the same hierarchy you learned for documents (Chapter 4) and design (Chapter 10), turned vertical and made bold.

Figure 18.2 (described): a well-designed research poster, top to bottom. At the very top, a title in large type readable from ~3 meters — ideally phrased as the finding ("A Sulfide Interphase Doubles Solid-State Battery Cycle Life"), not a topic ("Investigations into Solid-State Electrolyte Interphases"). Below it, smaller, the authors and affiliation. The body is arranged in columns read left-to-right, top-to-bottom, with a numbered or visually obvious reading path so the eye knows where to start and where to go next. The dominant visual element — the single most important figure, large and clearly captioned — sits near the center-top where the eye lands first. Text is in short chunks: a one-line takeaway box (often boxed or color-blocked, top-right or center, stating the finding in one sentence), brief bulleted method notes, a compact results figure with an interpretive caption, and a short "what it means" conclusion. Generous white space separates the blocks. The font for body text is large enough to read at arm's length (roughly 24 point minimum). Nothing is a paragraph longer than a few lines. A color-blind-safe palette and high contrast make the figures legible to everyone (Chapter 10).

The principles, distilled:

• Make the title the finding, big. It's the only thing most passersby read. "A Sulfide Interphase Doubles Cycle Life" stops the right people; "A Study of Electrolyte Interphases" stops no one.
• Give it a reading path. Number the sections, or use arrows, or rely on strong left-to-right column flow — but make it obvious where to start and where to go. A reader who can't find the entry point leaves.
• One dominant figure. The single graph that proves your takeaway should be the biggest thing on the poster after the title, with an interpretive caption (Chapter 9) that states what it means, not just what it is.
• Ruthless word economy. A poster has perhaps a tenth of the words of the paper. Use short chunks, boxes, and bullets — not paragraphs. If a block is more than four or five lines, cut it.
• White space is not wasted space. Crowding is the enemy of a hall reader. Empty space guides the eye and signals confidence; a packed poster signals panic.
• Design for the standing, distracted reader. Big fonts, high contrast, color-blind-safe palette, legible from arm's length. A poster that requires leaning in to read ten-point text has already lost most of the hall.

A useful sanity check: the five-second test and the one-minute test. From five feet away, in five seconds, can someone get your title and your main figure's point? Up close, in one minute, can they read the takeaway box, the key figure's caption, and the conclusion and understand your contribution? If a poster passes both, it works. Most failed posters pass neither — too dense for five seconds, too disorganized for one minute.

💡 Tip: the poster is a backdrop for a conversation, not a substitute for one. The real action at a poster session is you talking to people. The poster's job is to attract the right people and give them an anchor while you explain. So design it to support the conversation you'll have — big figure to point at, takeaway to gesture toward — not to replace you. The best poster in the world standing alone loses to a mediocre poster with a clear, enthusiastic presenter beside it.

18.5 The Elevator Pitch and the Lightning Talk

Two compressed formats deserve their own treatment because researchers face them constantly and prepare for them rarely: the sixty-second poster pitch and the five-minute lightning talk. Both punish anyone who hasn't decided, in advance, what to cut.

The elevator pitch: 60 seconds at your poster

Someone stops at your poster. They have maybe a minute of attention before they drift to the next one. They might be a potential collaborator, a future employer, the senior figure in your field, or a curious stranger from an adjacent discipline. You cannot give them the whole paper; you have time for one well-built minute. This is the elevator pitch, and it has a reliable structure — which is, not coincidentally, the Specific Aims structure from Chapter 17 spoken in sixty seconds:

1. The problem (one sentence). "Solid-state batteries could be safer and pack more energy than today's lithium-ion — but they die fast, because the interface between the electrolyte and the electrodes breaks down within a few hundred cycles."
2. What you did (one sentence). "We developed a thin sulfide interphase, deposited from solution, that stabilizes that interface."
3. The key result (one sentence, with the number). "It holds 94% of its capacity after 800 cycles — where the previous best held 80% after 500."
4. Why it matters / the hook to keep talking (one sentence). "That's nearly double the cycle life with no loss of energy density, which is the kind of margin that moves solid-state batteries toward real products. Want me to show you how it works?"

That's four sentences, roughly forty-five seconds, and it ends with an invitation that hands the conversation to the listener. If they're interested, you go deeper into the figure; if they're not, you've given them a complete, memorable contribution in under a minute. The discipline is the same as the assertion–evidence slide: lead with the point, support it with the one number that matters, drop everything else.

✏️ Try This. Draft your own 60-second pitch for a project you've worked on, using the four-sentence skeleton: problem, what you did, key result (with one number), why it matters (+ a hook to continue). Write it out, then say it aloud and time it. If it runs past 60 seconds, you have too much — cut until each beat is one clean sentence. The most common failure is starting with background ("So, there's this whole field…") and never reaching the result. Start with the problem, get to your number fast.

The lightning talk: 5 minutes, ~5 slides

A lightning talk is a five-minute presentation, often with a hard cap (some venues auto-advance your slides and cut your mic at the buzzer). Five minutes is not a short normal talk; it's a different talk that requires its own ruthless design. You have time for one idea, not three.

The structure that survives the format:

• One result, not a survey. A lightning talk makes a single point. Pick your best finding and build the whole talk around it. Do not try to compress a 15-minute talk into 5 — select, don't compress.
• Roughly five assertion–evidence slides. A title/hook, the problem, what you did (one slide), the key result (the heart — give it the most time), and the takeaway. No roadmap slide; in five minutes the roadmap is the talk.
• A hook in the first 15 seconds. There's no warm-up runway. Open on the problem or the result itself: "We nearly doubled solid-state battery cycle life — here's how."
• End with a pointer. "The details are in our paper / at poster 47 / talk to me after." A lightning talk is a teaser whose job is to send interested people to the longer version.

The lightning talk is excellent training even when it's not required, because the constraint forces the discipline this whole book preaches: if you can say it well in five minutes, you understand which one thing actually matters. Most researchers, asked to cut their talk to five minutes, discover their "essential" content was 80% optional.

🔄 Check Your Understanding. What is the key mental shift between preparing a 15-minute talk and preparing a 5-minute lightning talk — and why does "just compress the long version" fail?

Answer

The shift is from compress to select. A 15-minute talk can carry a result plus context, methods, and secondary findings; a 5-minute talk can carry one idea, period. "Just compressing" the long version fails because compression keeps all the points and shrinks each — producing a rushed, shallow version of everything, intelligible as nothing. Selection instead drops most points and gives the one survivor enough room to land. You don't speak the 15-minute talk faster; you build a different talk around a single finding, with a hook, the result, and a pointer to where the rest lives.

📐 Project Checkpoint

This chapter adds the sixth piece to your Communication Portfolio: a presentation — a slide deck with speaker notes. In Chapter 17 you wrote the Specific Aims page (or its equivalent) for your portfolio proposal; now you'll turn that same work into a talk, which is where the page becomes a podium.

This chapter's increment: build a short assertion–evidence deck (about 6–10 slides) presenting your portfolio project, plus speaker notes. Whatever your project is — a research result, a tool you built, an analysis, a proposal — design the talk a live audience could follow. Concretely:

1. Write the hook first (your opening 60 seconds), as speaker notes. Use the Specific Aims spine from Chapter 17: the problem, the gap, what you did, why it matters. Do not open with an outline; open with something that makes the room want the answer.
2. Build every slide as assertion–evidence. A full-sentence headline that states the slide's one claim; a single visual as the body; almost no bullet text. Run the headline test on each: is it a sentence (keep) or a topic (rewrite)?
3. Pick your one peak result and give it its own slide with the most air. State the claim, show the evidence, plan a pause.
4. Make the final slide your takeaway, not "Questions?" — one sentence you want remembered, on screen through Q&A.
5. Write speaker notes for each slide: what you'll say (the argument, the details the slide doesn't show). The slide is the visual aid; the notes are the talk.

Then practice it aloud against a clock and trim to ~90% of a target time you set (say, a 6-minute version). Time it honestly — out loud, not in your head.

Self-check before you file it: Read only your slide headlines, in order, ignoring the visuals and notes. Do they tell a coherent story — problem, approach, result, meaning — all by themselves? If the headlines alone make your argument, your deck follows assertion–evidence. If they read as a list of topics ("Background, Methods, Results, Conclusion"), you've built a document to perform, not a talk to give — rewrite the headlines as claims. Preview of the next increment: the portfolio's seventh and final piece, a blog post for a general audience, arrives in Chapter 28 — where the discipline shifts from compressing for experts to translating for everyone.

18.6 Handling Q&A

The talk ends and the hard part begins: the question period, the one stretch of a presentation you can't fully script. Q&A makes presenters more anxious than the talk itself, because it's unpredictable — and yet it's also where you can most visibly demonstrate command of your work. A few principles turn it from an ordeal into an asset.

Listen to the whole question, then pause. The instinct under stress is to start answering before the questioner finishes, or the instant they stop. Resist it. Let them finish, take a beat to actually think, and then answer. A two-second pause reads as thoughtful, not slow; it also stops you from answering a question they didn't ask. If a question is long or muddled, restate it in one clean sentence before answering ("So you're asking whether the interphase survives at higher temperatures — is that right?"). Restating buys you thinking time, confirms you understood, and lets the rest of the room hear the question if the asker mumbled.

Answer the question that was asked, briefly, then stop. Q&A answers should be short — a focused minute, not a second talk. Make your point and stop; if they want more, they'll follow up. The presenter who responds to a one-line question with a five-minute re-lecture loses the room and usually the chair's patience.

When you don't know, say so — and bridge. The single most credibility-saving move in Q&A is the honest "I don't know." It is far stronger than bluffing, because experts can smell a bluff and it poisons trust in everything else you said. But don't stop at "I don't know" — bridge: "I don't know whether it holds above 60 °C — we haven't tested that range yet, but it's exactly the next experiment, and I'd predict [X] because [reason]." That answer admits the limit, shows you've thought about it, and demonstrates scientific judgment. "I don't know, that's a great question" with genuine engagement beats a confident wrong answer every time.

Handle the hostile question by separating the substance from the tone. Occasionally a questioner is aggressive — a skeptic, a rival, someone showing off. Do not match their hostility; it's a trap that makes you look bad regardless of who's right. Extract the legitimate technical question underneath the attack and answer that, calmly, as if it had been asked politely. "That's a fair concern about the control condition — here's why we chose it…" If they're simply wrong, you can disagree without heat: "I see it differently, and here's the evidence." Composure under a hostile question is one of the most persuasive things an audience ever sees; you win the room by staying level while the other person doesn't.

For the rambling non-question, find the question and answer it — or, if there truly isn't one, acknowledge gracefully and move on ("That's an interesting perspective; happy to discuss it after the session") so you don't burn everyone's time. The chair will usually thank you for keeping things moving.

Here's the contrast that captures the whole section:

❌ Before (defensive, bluffing):

Questioner: "Have you checked whether this holds at elevated temperature? Batteries run hot." Presenter: "Well, our conditions are standard and widely used, so I don't think temperature would be an issue, and the result is very robust, so…" (defensive, vague, quietly bluffing past a real gap)

✅ After (honest, bridging):

Presenter: "Good question — we tested at room temperature, so I genuinely don't know yet how it behaves at 60 °C and above, which is the regime that matters for real packs. That's the next experiment on our list. My expectation is that the sulfide interphase is more temperature-stable than the baseline, because [mechanism], but I want the data before I claim it."

Why it's better: the second answer is more credible because it admits the limit. It names the gap precisely, frames it as future work, offers a reasoned prediction, and refuses to overclaim — which makes the audience trust every other claim Lena made. The first answer, by bluffing past the gap, invites the questioner (and the room) to wonder what else she's hand-waving. In research, calibrated honesty is not weakness; it's the whole currency.

🔄 Check Your Understanding. Why is "I don't know — but here's what I'd expect and why" usually a stronger answer than a confident attempt to cover a gap you can't actually fill?

Answer

Because your audience is expert enough to detect a bluff, and a detected bluff doesn't just lose that one point — it retroactively poisons trust in everything else you presented, since now they wonder where else you hand-waved. The honest answer does the opposite: admitting a specific limit demonstrates that you know the boundaries of your own evidence, which is exactly the judgment that makes the rest of your claims credible. Adding the reasoned prediction ("here's what I'd expect, because…") shows you've thought about the open question without overclaiming. Calibrated honesty (Chapter 7) raises your credibility; bluffing spends it.

18.7 Common Mistakes and Practical Considerations

The failures that sink research presentations cluster into a recognizable set. Most are fixable in revision and rehearsal — which is the good news.

Reading your slides verbatim. The cardinal sin, and the direct cause of death by PowerPoint. If your slide is a full paragraph and you read it, the audience reads ahead, gets bored, and tunes out. The fix is structural, not behavioral: build assertion–evidence slides with little text, and there'll be nothing to read aloud — you'll be forced to speak from knowledge while the visual supports you.

Too many slides for the time. A talk built with thirty slides for fifteen minutes cannot be given at a human pace; it can only be raced. The one-slide-per-minute heuristic is a sanity check: count your slides against your minutes before you rehearse, and cut early.

Burying the result. Researchers love to recreate their own journey — background, then methods, then finally the finding — which puts the one thing the audience came for at the very end, when attention is lowest. Front-load it: promise the result in the first ninety seconds, deliver it at the peak, restate it at the close. (This is BLUF from Chapter 4, applied to a clock.)

Tiny fonts and unreadable figures. A figure pasted from the paper, with eight-point axis labels and a dense legend, is invisible from the back of the room. Rebuild figures for the screen: large labels, direct annotation, one message per figure. The exploratory chart you made to find the result is not the explanatory chart you show to prove it (Chapter 9) — and on a slide, the gap is even wider.

Ignoring the clock in rehearsal. "I'll be fine, I know this material" is how talks run four minutes over. Practice out loud, with a timer, more than once. Your in-head rehearsal is always faster than the real thing.

A poster that's a shrunken paper. Columns of ten-point prose nobody reads, no reading path, no dominant figure, no takeaway box. A poster is a visual; if yours is mostly text, redesign around the one figure and the one sentence you want remembered.

No backup for a live demo or animation. If your talk depends on a live demo, a video, or an internet connection, it will fail eventually — the projector won't sync, the wifi will drop, the demo will crash. Always have a static backup (screenshots, a recorded clip on your laptop, the key numbers on a slide). Chapter 31 treats this at length; for now, the rule is: anything that can fail live, have a non-live version of.

The unprepared Q&A. You can't script questions, but you can anticipate them. Before the talk, list the five questions you'd least like to be asked — the weak control, the small sample, the alternative explanation, the scaling concern — and prepare honest, bridging answers. The "surprise" question is usually one you could have predicted if you'd looked for your own soft spots.

It depends — on the venue and the field. Conference norms vary widely. Talk lengths run from 5-minute lightning slots to 45-minute invited talks; some fields expect dense equation-heavy slides, others expect spare visual ones; some poster sessions are quiet and scholarly, others are loud and competitive. Some venues forbid you from advancing your own lightning-talk slides. Always check the specific session's format, length, and conventions, and design for that room — the principles here (the audience can't read and listen at once; lead with the point; one idea per slide; finish on time) are portable, but the surface expectations are local. When in doubt, watch a session in the same track before yours and calibrate.

Frequently Asked Questions

What is the assertion–evidence approach to slides?

It's a slide design method (developed by Michael Alley and colleagues) where each slide's headline is a full-sentence claim — the point you want the audience to take away — and the body is a single piece of visual evidence for that claim (a graph, a photo, a diagram, a big number), with almost no bullet text. It replaces the default "topic-and-bullets" slide (a noun-phrase title like "Results" over a list of text fragments), which causes death by PowerPoint because the topic title states no point and the bulleted text forces the audience to read instead of listen. The headline test tells them apart: a sentence headline is assertion–evidence; a topic headline is not.

How many slides should I have for a conference talk?

As a rough starting heuristic, about one slide per minute — so roughly 15 slides for a 15-minute talk, usually fewer. It's a sanity check, not a law: a data slide you discuss in depth might take three minutes, while a section divider takes ten seconds. The real test is timing: practice out loud with a clock and aim to finish at about 90% of your allotted time (a 15-minute talk landing in 13–14 minutes). If you've built thirty slides for fifteen minutes, you've built a talk you can't give at a human pace — cut it.

How do I design a research poster that people actually read?

Decide on one takeaway first — the single thing you want someone to remember — and build everything around it. Make the title the finding ("A Sulfide Interphase Doubles Cycle Life"), in type readable from across the room. Give the poster an obvious reading path (numbered sections, arrows, or strong column flow). Feature one dominant figure with an interpretive caption. Keep text in short chunks (no paragraphs over a few lines), use generous white space, and size body fonts to read at arm's length (roughly 24 point and up). Run the five-second test (title + main figure readable from five feet) and the one-minute test (takeaway, key caption, conclusion understandable up close).

How should I handle a question I can't answer?

Say so honestly, then bridge: "I don't know whether it holds above 60 °C — we haven't tested that range — but it's the next experiment, and I'd expect [X] because [reason]." Admitting a specific limit is more credible than bluffing, because an expert audience can detect a bluff and it poisons trust in everything else you said. The honest answer shows you know the boundaries of your own evidence — which is exactly the judgment that makes your other claims believable — and the reasoned prediction shows you've thought about the open question without overclaiming.

What's the difference between a 5-minute lightning talk and a short regular talk?

A lightning talk requires selection, not compression. A regular talk can carry a result plus context, methods, and secondary findings; five minutes carries one idea. Don't try to speak your 15-minute talk faster — build a different talk: a hook in the first 15 seconds, the problem, what you did (one slide), the key result (the heart), and a pointer to where the full version lives (your paper or poster). Around five assertion–evidence slides, no roadmap slide. The constraint is good training: if you can deliver it well in five minutes, you've found the one thing that actually matters.

Chapter Summary

Key Takeaways

• A talk is not a paper read aloud, and a slide is not a document. The audience listens at your pace and cannot read and listen at once — so the slide must show what words can't (a visual), and your voice carries the argument.
• Use assertion–evidence slides: a full-sentence claim as the headline, a single visual as the body, almost no bullet text. The "topic-and-bullets" slide causes death by PowerPoint.
• Structure the talk as hook → roadmap → approach → key result → takeaway, and front-load the result — promise it in the first 90 seconds, don't bury it at the end.
• Time it: ~one slide per minute as a sanity check, practice out loud with a clock, finish at ~90% of the slot, and plan your cut points.
• A poster has one takeaway, a finding-as-title, an obvious reading path, one dominant figure, short text chunks, and white space. It passes the five-second and one-minute tests.
• The elevator pitch (problem → what you did → key result with the number → why it matters) and the lightning talk (select one idea, ~5 slides) both reward deciding in advance what to cut.
• In Q&A, listen fully, restate the question, answer briefly, and when you don't know, say so and bridge to a reasoned prediction. Stay calm under hostile questions.

Action Items

1. Convert every slide to assertion–evidence: rewrite the headline as a sentence, replace bullets with one visual.
2. Read your headlines alone, in order — make sure they tell the whole story.
3. Front-load your result; make your last slide the takeaway, not "Questions?".
4. Count slides against minutes; practice out loud with a clock; trim to ~90% of the slot.
5. For your poster: pick one takeaway, make the title the finding, add a reading path and one dominant figure.
6. Draft and time a 60-second elevator pitch and (if relevant) a 5-slide lightning talk.
7. List the five questions you'd least like to be asked and prepare honest, bridging answers.

Common Mistakes

Decision Framework: Is this slide / poster / answer ready?

• [ ] Slide: Is the headline a sentence that makes a claim (not a topic)? Is the body one visual (not bullets)?
• [ ] Talk: Does the result arrive in the first 90 seconds? Do the headlines alone tell the story? Does it finish at ~90% of the time?
• [ ] Poster: Is there one takeaway? Is the title the finding? Is there an obvious reading path and one dominant figure? Does it pass the 5-second and 1-minute tests?
• [ ] Pitch: Problem → what you did → result-with-number → why it matters, in ~60 seconds?
• [ ] Q&A: For your five scariest questions, do you have an honest, bridging answer ready?

Spaced Review

A few questions reaching back, to strengthen retention.

1. (From Chapter 9) You learned that "a figure does not speak for itself" and that captions should interpret (climb from label → observation → interpretation). How does the assertion–evidence slide apply that exact idea — and what plays the role of the interpreting caption on a slide?
2. (From Chapter 4) What is the inverted-pyramid / BLUF principle, and where does it appear in (a) a conference talk and (b) an assertion–evidence slide?
3. (From Chapter 13, bridging) IMRaD keeps Results (what happened) separate from Discussion (what it means). A conference talk compresses IMRaD hard — what does it keep from that discipline, and what does it mostly leave in the paper?

What's Next

Part III ends here, at the podium and the poster — the moment your research meets a live audience. Chapter 19: Emails That Get Read and Get Results opens Part IV and shifts the whole context from the academy to the workplace: the most frequent, lowest-glamour, highest-leverage writing most people ever do. The medium changes from a fifteen-minute talk to a five-line message, and the audience from a study-section reviewer to a busy colleague with three hundred unread emails — but the discipline you've been building since Chapter 2 is exactly the same. Know your reader. Lead with what matters. Make every line earn its place. You'll meet the assertion–evidence idea again, in fuller form, in Chapter 30 (slide design) and Chapter 31 (delivery), when the Presentations track returns to the stage you just stepped onto.

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