How this course works
Six sessions, 45 minutes each, and one real deck that you build across all of them. Not a practice deck. Pick something you genuinely have to present in the next month: a client pitch, a budget ask, a quarterly readout. That deck is the artifact, and every session moves it forward.
★ What you walk out with tonight. A test you can run on any slide in thirty seconds, the name and the citation for the only slide-design method with peer-reviewed evidence behind it, and an honest read on which presentation advice you have been given is research and which is just repeated a lot.
LiveThe artifact: your deck, in six moves3 min▶
Every session ends with your deck further along than it started. The order is deliberate, because the mistakes compound backwards: people design slides before they have an argument, and then no amount of design saves it.
| Session | What your deck gains |
|---|---|
| 1 · tonight | A measured score on your worst slide, and the reason it scores that way |
| 2 | An outline that is an argument, approved before a single slide is designed |
| 3 | Every slide rebuilt as one claim plus its evidence |
| 4 | Charts that argue, with every figure traceable to a source you can open |
| 5 | The real build, in whichever tool you actually have a licence for |
| 6 | Speaker notes, the Q and A you are dreading, and an accessibility pass |
If you only do three sessions, do 2, 3 and 6. That is where the score moves most.
Self-studyWhy this course exists at all4 min▶
There is no shortage of official training on AI presentation tools. Microsoft Learn has a seven-unit module on building presentations with Copilot. LinkedIn Learning has a 51-minute course from a genuine PowerPoint MVP. They are good, and this course maps against them honestly in the coverage table at the bottom of every session.
What none of them teach is the part that decides whether your deck works: audience-level framing, evidence integrity, and accessibility QA. They teach the tool surface, because that is what a vendor course is for. The only slide-design method with peer-reviewed comprehension data behind it lives in an engineering-education journal that no vendor course cites.
So the split in this course is: the tools are treated as tools, and the thinking is treated as the subject.
The test, and the number it produces
7 min livePut a slide up. Start a timer. At thirty seconds, ask one person in the room to say, out loud, what you are asking them to do. Not what the slide is about. What you want from them. That is the whole test, and the first time you run it on your own deck it is genuinely unpleasant.
LiveWhy "what is this about" is the wrong question4 min▶
Ask a room what a slide is about and they will answer, correctly, from the title. Ask what you want and the room goes quiet. That gap is where decisions go to die, and it is invisible from the front of the room because you already know the answer.
The three failures that produce the silence, in the order they cost you most:
- The slide states a topic, not a message. "Q3 Revenue Performance Review" is a filing label. It tells the room where they are, not what changed.
- The evidence is a bullet list. Six bullets is six things to read, in a room where you have thirty seconds and everyone is also reading their phone.
- Nothing is cued. The one number that should decide the meeting is set in the same weight as the four numbers that should not.
A renewal slide with no visual comparison gets deferred to next week. Next week is past the auto-renew date. Nobody in that room decided to renew the contract, and the contract renewed.
A title that is a filing label scores fine on word count and still fails the test, which is exactly why word count is not the measure. Hold that thought until Part 2.
The one method with real evidence behind it
8 min liveAlmost everything you have been told about slides is convention. One thing is not. In 2013, Garner and Alley put 110 engineering students in front of two versions of the same content and measured what came out the other side. The result is the spine of this whole course.
LiveAssertion-evidence, in one line4 min▶
A sentence headline states the slide's main message, and that message is supported by visual evidence rather than a bullet list. That is the entire method. It comes from Michael Alley at Penn State, and it has been tested rather than asserted.
The study worth citing by name, because you will need it the first time someone tells you your slides look unusual:
Note what the study does not say. It does not say bullets are always wrong, and it does not say slides should be empty. It says a sentence headline plus visual evidence beat a topic headline plus a bullet list, on comprehension, in a controlled comparison.
Q3 Revenue Performance Review
The room now knows which slide they are on. They do not know that a decision is being asked for, or which region caused the problem, or what you want them to do about it.
Q3 missed by 8 percent, and EMEA is the entire gap
The claim is on the slide. The room can now disagree with something specific, which is the fastest route to a decision that holds.
Self-studyMayer's three principles that touch every slide6 min▶
Richard Mayer's multimedia learning work is the other research pillar, and three of his principles land directly on slide design:
| Principle | What it says | What it costs you to ignore |
|---|---|---|
| Coherence | People learn better when extraneous words, pictures and sounds are excluded | Every decorative gradient and draft watermark is competing with your point |
| Signaling | People learn better when cues highlight the organisation of the essential material | The room has to find the important number themselves, and some will not |
| Redundancy | People learn better from graphics and narration than from graphics, narration and the same words on screen | Reading your slide out loud makes comprehension worse, not safer |
That last one is the counter-intuitive one, and the simulator lets you watch it happen. Read the slide aloud verbatim and the number drops twenty points.
One honest caveat to carry: 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. But do not oversell them as laws of business communication.
What is research, and what is just repeated a lot
5 min liveYou have been given a lot of presentation rules. Some are tested. Most are one confident person's experience, repeated until it sounded like a finding. Knowing which is which is what stops you defending an arbitrary rule in front of a client.
LiveThe 10/20/30 rule, handled honestly3 min▶
Guy Kawasaki's rule: ten slides, twenty minutes, no font under thirty points. His reasoning, in his own framing, is that a normal person cannot hold more than ten concepts in a meeting, that twenty minutes leaves forty for discussion in a one-hour slot, and that a large font forces you to explain your pitch rather than read it.
All three are sensible. None of them has a study behind it. No empirical validation of 10/20/30 exists. It is one very experienced investor's doctrine, and it has been repeated so widely that it now gets cited like a finding.
Use it as a useful default, not as an argument. "We keep pitch decks near ten slides" is fine. "Research shows ten slides is optimal" is not a sentence you can back up, and someone will eventually ask.
The thirty-point font point is the interesting one, because it is convention that lands near something real. The reason big type works is not concept-counting, it is that small text at the back of a room is unreadable, and that has a standard behind it. Which brings us to the lever that is locked until session 6.
That is real arithmetic, not a style opinion. Pale accent type on white is a favourite of AI deck generators and it fails the standard by a factor of two and a half. Session 6 makes this a gate your deck has to pass.
Watch the number climb, then run it on your own slide
★ 22 min · everyone buildsThe simulator below renders a real slide in your browser and rebuilds it as you toggle each design lever. The comprehension percentage is a teaching model. The contrast ratio is real WCAG maths on the colours the slide is actually painted with, and the chart is real SVG. Start with everything off, which is where most decks live.
27 percent. A topic label, six bullets, a decorative blob and a draft watermark. This is the median corporate slide and it is not a strawman.
48 percent. Swap the filing label for a sentence that states the message. One change, twenty-one points. This is the single biggest lever on all six slides in the golden set.
64 percent. Replace the bullet list with one visual that proves the claim. Now it is assertion-evidence, and you are at the level the 2013 study measured.
73 percent. Strip the gradient blob and the CONFIDENTIAL-DRAFT-v7 stamp. Mayer's coherence principle, worth nine points for deleting things.
82 percent. Cue the one number that decides the meeting. This is the ceiling until session 6, and no amount of further design gets past it.
91 percent. The text finally clears 4.5:1. The last nine points were an accessibility fix, which is why session 6 is not optional garnish.
Open your own worst slide. The real one, from the deck you actually have to present. Not a clean example. The one you already know is bad.
Score it by hand. Walk the five levers in the simulator against your slide and mark each one yes or no. Five noes is 27 percent. Be honest about the headline: a topic label with a verb in it is still a topic label.
Rewrite the headline only. One sentence, present tense, stating what changed or what you want. Nothing else about the slide changes yet. This is the twenty-one point move and it takes four minutes.
Run the real 30-second test. Send the slide to one colleague with one question: "in thirty seconds, what am I asking for?" Their answer is your actual score. The simulator is a model, that person is the measurement.
Check your deck's body text against the standard. Paste your text colour and your background colour into the contrast box above. If it is under 4.5:1, note it. Session 6 fixes it properly.
Toggle both traps and watch it backfire. Switch on "read the script aloud" and "add more bullets" from the 91 percent ceiling. Twenty points and fifteen points, gone. Both are things people do to feel safer.
If you are presenting this week. Do the headline rewrite on every slide and nothing else. It is the cheapest twenty-one points in this course, and it does not require you to redesign anything.
Before session 2
20 min- Choose the deck. One real deck you have to present in the next month. Write its audience and the single decision you want out of the room, in one sentence each. Session 2 starts from those two sentences.
- Rewrite every headline as a sentence. Do not touch the layout, the charts or the colours. Headlines only. Count how many were filing labels before you started.
- Get one real 30-second score. One colleague, one slide, one question. Write down what they actually said, not what you hoped.
- Find the pale text. Check your template's body colour against its background in the contrast box. Note the ratio. You will need it in session 6.
- Optional, and worth it: read the one-line summary of assertion-evidence at writing.engr.psu.edu/assertion_evidence_EA.html so the method has a home in your head beyond this page.
Three questions
Clear all three and session 1 gets stamped in your passport.
1 · What does the 30-second test actually ask the room?
"What is it about" is answerable from any title, which is why it hides the problem. "What am I asking for" is the question that exposes a slide with no message.
2 · Which of these has peer-reviewed comprehension data behind it?
Garner and Alley 2013, 110 students, significant at p under .01. The 10/20/30 rule is Kawasaki's doctrine asserted from experience, with no study behind it, and the one-font rule is taste.
3 · You read your slide aloud, word for word, so the room does not miss anything. What happens to comprehension?
Mayer's redundancy principle: graphics plus narration beats graphics plus narration plus the same words on screen. In the simulator it costs twenty points from the ceiling. The safe-feeling move is the harmful one.
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 1 cheat sheet · pin this
Tonight you rewrote headlines on a deck whose structure you never questioned. Session 2 goes underneath: your deck is an argument, and if the argument is wrong the slides cannot save it. You will generate an outline and then kill half of it, and the golden-set average moves off 48 for the first time.