Where we left the deck, and tonight's target
Session 1 measured the problem: 27 percent of a room can state your ask thirty seconds after a normal corporate slide. Session 2 fixed the argument underneath, and cut the deck roughly in half. Every surviving slide is still built the old way. That changes in the next forty-five minutes.
★ Tonight's target, stated as a number. Three levers: the sentence headline worth 21 points, the visual as evidence worth 16, and coherence worth 9 for deleting things. That is 27 to 73 percent understood. The remaining two levers, cueing the hero number and clearing WCAG contrast, are session 4 and session 6, which is why 82 is the ceiling until then.
LiveWhy this is the session people skip, and should not3 min▶
Rebuilding slides is not fun. It is the middle of the work, there is no new tool to play with, and it involves admitting that a slide you were quite proud of is a filing cabinet with a chart in it. So people skip to session 5, open a generator, and produce a beautiful version of the same failure.
The reason to sit through it: this is the only part of the course with a controlled study behind it. Not a framework, not a style, a measured comparison with a sample size and a p value. If you take one defensible thing out of six sessions, take this one.
The sentence headline, properly this time
7 min liveSession 1 gave you the rule in one line. Here is the rule with its edges. A sentence headline states the slide's main message as a complete sentence: present tense, saying what changed or what you want, under twelve words. Every part of that does work, and the part people drop is the "message" part, because a verb is easy and a claim is not.
LiveThe four rules, and the one that gets faked4 min▶
- A complete sentence. Subject, verb, and something at stake. "Revenue Trends" is a label. "Revenue fell" is a sentence.
- Present tense. "EMEA is the entire gap" lands harder than "EMEA was the largest contributor to the shortfall". Present tense is a room-shaped tense.
- What changed, or what you want. Nothing else qualifies. Background, context and process are not messages.
- Under twelve words. Not a law, a discipline. Past twelve you are writing a paragraph and the room reads instead of listening.
The rule that gets faked is the second half of rule three. People add a verb, feel finished, and ship a topic label in a sentence costume. "This slide reviews our Q3 regional performance" is grammatically a sentence and informationally a title. If the room could not disagree with your headline, it is not a claim yet.
LiveFive real labels, rewritten5 min▶
These are the six slide types in the golden set the simulator scores. Left column is what the slide is actually called in most decks. Right column is the same slide with a claim on it.
| Topic label | Sentence assertion | Words |
|---|---|---|
| Q3 Revenue Performance Review | Q3 missed by 8 percent, and EMEA is the entire gap | 10 |
| Headcount Planning FY26 | We need three hires in Q1 or support slips in March | 10 |
| Vendor Renewal Overview | The renewal auto-triggers in 11 days, so we decide today | 9 |
| Pricing Strategy Update | Moving the entry tier up 10 dollars keeps churn flat | 9 |
| Incident Post-Mortem Q2 | One config change caused all four outages, and it is fixed | 10 |
Notice what happened to the vendor renewal one. The label said "overview" and the assertion said "we decide today". Same slide, same data, and the second version makes it impossible for the meeting to end without a decision. That is not a design trick, that is the headline doing the argument's job.
An incident review presented as "Q2 Post-Mortem" ran forty minutes of blame-adjacent discussion. The same content re-presented as "One config change caused all four outages, and it is fixed" took eleven minutes and ended with two follow-up actions. The content did not change. The first slide invited an investigation and the second one reported a conclusion.
Self-studyThe study, cited so you can defend it5 min▶
The first time you present slides built this way, somebody senior will tell you they look unusual. This is the paragraph that ends that conversation politely.
Be precise about what it says, because overclaiming is how you lose the argument. It does not say bullets are always wrong. It does not say slides must be sparse. It says a sentence headline plus visual evidence beat a topic headline plus a bullet list, on comprehension, in a controlled comparison with 110 students.
And be precise about what is not evidence. Assertion-evidence is peer-reviewed. The 10/20/30 rule is convention with no study behind it, as session 1 covered in full. Keeping those two apart in your own head is what lets you defend the first one hard.
The visual is evidence, not decoration
6 min liveThe other half of assertion-evidence is the half people get wrong quietly. The rule is not "add a picture". The rule is one claim, and one visual that proves that claim. If your headline says EMEA is the gap and your chart shows all six regions equally, the chart is not evidence, it is company.
LiveWhen a table is genuinely the right answer4 min▶
I am not anti-table, and the anti-table version of this advice is wrong often enough to be dangerous. A table is the right visual when it is doing one specific job: a reference lookup, where a person scans for one value and does not need to see a pattern.
| Use a table | Use a chart |
|---|---|
| Somebody will scan it to find their own row: pricing tiers, SLA thresholds, the regional owner list | You want the room to see a shape: growth, decline, a gap, a concentration |
| The values are categorical and exact, and precision is the point | The comparison is the point, and the exact figures live in the notes or the appendix |
| It goes in the appendix, where people read at their own pace | It goes on the slide, where you have thirty seconds |
The failure mode has a name worth using out loud: a bar chart wearing a table's clothes. Four quarters and two regions in a grid is not a lookup, nobody is scanning it for one value, they are trying to work out whether the line goes down. You have made the room do arithmetic you could have done for them with one shape.
The one-question test for any visual. Cover your headline. Look at the visual alone. Can you reconstruct the claim? If the visual supports four possible claims equally well, it is not evidence for yours, and cueing it in session 4 will not save it.
Self-studyOne claim per slide, and what to do with the leftovers4 min▶
One claim per slide is the constraint that makes the rest of it work, and it is the one people negotiate with. The negotiation usually sounds like "these two points really belong together". Sometimes they do. Usually what belongs together is the claim and its evidence, and the second point is a different slide that has not been given its own headline yet.
Three honest routes for the leftovers, in the order I try them:
- Its own slide. If it carries a claim the thesis needs, it earned one. You cut half the deck in session 2 precisely so there was room.
- Speaker notes. Detail you will say but not show. Session 6 builds these out properly, and this is where most of your old bullets belong.
- The appendix. With a named person who will ask for it, per session 2's rule. No name, no appendix.
What you must not do is the thing every tool makes easy: paste the leftovers back onto the slide as bullets because there is white space. In the simulator, adding bullets to a finished slide costs 15 points from the 91 percent ceiling. It is a trap that feels like thoroughness.
Coherence, and the hero number
5 min liveTwo of Richard Mayer's principles decide whether a well-built slide actually lands. Coherence says people learn better when extraneous words, pictures and sounds are excluded. Signaling says people learn better when cues highlight the organisation of the essential material. One removes, one promotes, and they only work as a pair.
LiveNine points for deleting things3 min▶
Coherence is the cheapest lever in the course, because it is subtraction. In the simulator, stripping the gradient blob and the CONFIDENTIAL DRAFT v7 watermark is worth 9 points. You are not designing anything. You are deleting things somebody added to feel professional.
The usual suspects, and every one of them is on a template somewhere:
- Decorative gradients and shape blobs. They exist because the slide looked empty, and empty was correct.
- Watermarks and version stamps. Nobody in the room needs to know it is v7. They need to know what you are asking for.
- The logo on every slide. The room knows whose meeting it is. Title slide, and stop.
- The gridline forest. Chart gridlines, axis ticks and data labels all at once means three encodings of one value.
The honest caveat, same as session 1: Mayer's principles come from instructional-learning experiments, not board meetings. They transfer well, and the assertion-evidence study is the bridge that tested them in a presentation setting. Do not sell them as laws of business communication.
Self-studyThe hero number, and why its points wait for session 44 min▶
Signaling in practice means one thing on a business slide: the one number that decides the meeting is set differently from every other number on the slide. Bigger, bolder, or in the accent colour, and everything else quietly greyed back.
You meet the principle tonight because you cannot build a slide without deciding what the hero is. Its 9 points land in session 4, when the number lives inside a chart you built and labelled yourself and there is something real to cue. Tonight's job is to know which number it is.
Two ways people break signaling, both from good intentions:
- Cueing four things. Four bold numbers is zero bold numbers. Pick the one the decision turns on.
- Cueing with pale colour. The accent that looks tasteful on your laptop is often pale accent type on white, which fails the WCAG contrast minimum by a wide margin. Session 6 makes that a gate, with real arithmetic rather than an opinion.
On slide count and font size: the 10/20/30 rule sits next to this conversation and it is convention, not evidence. Session 1 handled it in full. Use it as a default if it helps, and never as an argument.
Rebuild three real slides, right now
★ 22 min · everyone buildsThree slides from your own deck, the ones you brought from session 2. Seven minutes each. We are not making them pretty, we are making them carry a claim. Prettiness is session 5 and it is much easier once the slide knows what it is for.
Slide one, headline only. Write the claim as a sentence, present tense, under twelve words. Then read the old headline and the new one out loud back to back. The difference is audible, and that is the 21 points.
Cover the headline and look at the visual. Can you reconstruct the claim from the visual alone? If not, the visual is company rather than evidence, and you need a different cut of the same data.
Decide table or chart, honestly. Is anyone scanning this for one value, or do you want them to see a shape? Lookup means table, shape means chart. A grid of quarters is a bar chart wearing a table's clothes.
Delete the dead zone. Logo on every slide, page numbers, the DRAFT stamp, the gradient blob, half the gridlines. Nine points of pure subtraction, and it takes under a minute per slide.
Name the hero number. One number per slide, written down. Do not style it tonight. Just decide, because session 4 needs the decision and most people have never made it explicitly.
Repeat on slides two and three. The second one takes half as long. The third one takes five minutes, and by then you will have opinions about your own template.
Run the real 30-second test on the best of the three. One colleague, one slide, one question: "in thirty seconds, what am I asking for?" Compare their answer to session 1's answer on the same deck.
All three switched on is 73 percent, and that is exactly where the ladder sits after rung three. The last two rungs, 82 and 91, are session 4 and session 6. Nothing you do to these three slides tonight gets past 73, which is worth knowing before you spend an evening trying.
If you only have twenty minutes this week. Headline rewrite on every slide, then delete the dead zone. That is 21 plus 9 points, no new charts, no redesign, and it is the highest return per minute in the whole course.
Before session 4
30 min- Finish the headlines. Every remaining slide, a full sentence, present tense, under twelve words. Count how many were topic labels wearing a verb.
- One claim per slide, enforced. Any slide with two claims splits into two, or the second claim goes to speaker notes or the appendix with a name on it.
- Audit every visual against its headline. Cover the headline, look at the visual, write down the claim you would guess. Where it does not match, note what cut of the data would.
- Delete the dead zone across the deck. Blob, watermark, per-slide logo, surplus gridlines. Do it in one pass with the slide sorter open, it goes faster than slide by slide.
- Write down each slide's hero number. A plain list, slide by slide. Bring it to session 4, because that list is what session 4 turns into charts that argue.
- Flag every figure you cannot source. Mark it OPEN. Session 4 is about charts and data integrity, and the fabricated-figure conversation is easier when you already know where your own gaps are.
Three questions
Clear all three and session 3 gets stamped in your passport.
1 · Your headline reads "This slide reviews our Q3 regional performance". Is that a sentence assertion?
Grammar is not the test. The test is whether it states what changed or what you want, in a way the room could argue with. "Q3 missed by 8 percent, and EMEA is the entire gap" is a claim. "This slide reviews" is an announcement.
2 · Your headline says EMEA is the whole gap. Your chart shows all six regions at equal weight. What is wrong?
Assertion-evidence means one claim and one visual that proves that claim. Cover the headline: if the visual could sit under four different headlines, it is company rather than evidence. Cueing it later does not fix a visual that is arguing something else.
3 · Which of tonight's three levers is worth the most in the simulator, and how much?
Headline 21, evidence 16, coherence 9. The headline is the single biggest lever on all six golden-set slides, dropping each by 20 to 25 points when removed. Coherence is the cheapest per minute, which is a different thing from the biggest.
What this session teaches, and where it came from
Every fact on this page, and its verification tier, is recorded in the course's source map. Certificates, videos and assessments stay with the official providers.
Session 3 cheat sheet · pin this
Your slides now carry a claim each, and every one of them points at a number you have not checked. Session 4 is where charts start arguing instead of reporting, where the cue lever finally banks its 9 points and the average clears 82, and where you find out exactly how a confident invented figure gets into a deck without anybody lying.