Where we are
Session 3 got each slide down to one idea. That idea is now competing with the rest of the slide for attention, and the rest of the slide often wins.
★ What you walk out with tonight. One slide where the eye lands exactly where you intended, achieved entirely by turning other things down, plus the three-second test you can run on anything from now on.
Emphasis is relative, so the only way to add it is to remove it
7 min liveThis is the whole session in one line and it is genuinely counter-intuitive. There is no absolute quantity of "loud". A red bar is only loud among grey ones, and a deck where every slide has a red bar has no red bars at all.
LiveWhy greying out feels wrong, and why it is right4 min▶
People resist this more than any other technique in the course, and the resistance is worth naming rather than overriding.
- It looks like you are hiding data. You are not - the greyed bars are fully present, labelled and readable. Nothing was removed. Compare that to the usual alternative, which is filtering them out of the chart entirely.
- It looks unfinished. A chart in one colour plus grey reads as a draft to people used to categorical palettes. It reads as deliberate to anybody who has thought about it for ten seconds, and the second group includes everyone whose opinion changes the decision.
- It feels like devaluing your work. The other five cohorts took the same effort as Guji. Greying them acknowledges they are not the point of this slide, which is a statement about this slide, not about the analysis.
The grey is not a neutral background colour. It should be a genuinely legible mid-grey that a reader can still read values from. Greying out means turning down, not switching off.
Five channels, ranked by how hard they pull
7 min liveYou have five ways to make one thing stand out. They are not equivalent, they stack badly, and using three at once on the same element is a reliable sign of somebody who did not trust the first one.
LiveThe three-second test3 min▶
Show the slide to somebody for three seconds, take it away, and ask what they saw. Not what it said - what they saw first.
- They name your point. Done. Stop editing.
- They name something else. That thing is louder than your point. Turn it down rather than turning your point up, because turning up is how you end up with a slide where four things are shouting.
- They name nothing in particular. Nothing is emphasised. This is the most common result on a slide made from a default template, and it means the reader is doing all the work.
- They name the title. Also fine, if the title is a takeaway from session 3. A reader who reads the title and stops has still got your point.
Self-studyEmphasis in text, and the highlighting problem4 min▶
- Bold, not italic, not underline. Italic is for terms and titles and is weak at pulling the eye. Underline means link on any screen, and always will.
- Bold a phrase, not a sentence. A bolded sentence is a sentence in a different font. Three to six words is the range where bold does anything.
- Under a tenth of the text. Past that the bold becomes the baseline and the unbolded text starts to read as the emphasis, which is a genuinely disorienting effect to produce by accident.
- Never all-caps for emphasis. It removes word shape, which is what fast readers use, so it is simultaneously louder and slower to read.
The same rule as everywhere else in this session: the amount of emphasis in a document is fixed at roughly one thing per view, and every additional bold phrase divides it.
The grey-out pass
★ 22 min · everyone editsOne rule for the next twenty minutes, and it is strict: you may not add any colour, size or weight. You may only take them away. Every slide must end up with less ink than it started with.
Three-second test on the slide as it stands. Get a neighbour to tell you what they saw first. Write it down before you change anything.
Name the one thing. Write, in the margin, the single element the eye should land on. One element, not two. If you cannot pick, session 3 is not finished for this slide.
Grey everything else. Every other bar, line, box and label goes to a legible mid-grey. Turned down, not switched off - values must still be readable.
Add one annotation. An arrow and up to eight words on the one thing. This is the only addition permitted tonight, and it is the channel that explains rather than merely pointing.
Retest. Three seconds, a different neighbour. If they still name the wrong thing, something is still loud that you did not notice - usually a logo, a heavy border, or a total.
Run the test before you edit. Without the before result you cannot tell whether the edit helped, and people are remarkably bad at judging their own slides.
Grey the gridlines and borders too, not just the data. A heavy border is emphasis on the slide's edge, which is never the point.
Check the logo. Corporate logos are usually the most saturated object on the slide and they are on every slide. On a greyed slide they can win the three-second test outright.
Write the annotation as a claim, not a label. "12.4 percent, three times everyone else" beats "Guji cohort".
Count the colours at the end. Two is normal for a good slide: one accent and one grey. Three is defensible. Six means the template won.
If you take one thing from tonight. When somebody looks at the wrong thing, turn the wrong thing down. The reflex is to turn your point up, and following that reflex three times is how a slide ends up with nothing emphasised at all.
Before session 5
25 min- Run the three-second test on five slides from a deck you already sent. Record what people saw first, before you tell them what it was supposed to be.
- Do a grey-out pass on the three worst. Subtraction only.
- Count colours across one whole deck. If the accent colour appears on every slide, it has stopped being an accent.
- Bring your full deck, in order, to session 5. Next week is about the order itself, and you will need something to reorder.
Three questions
Clear all three and session 4 gets stamped in your passport.
1 · A colleague looks at your slide and their eye lands on the wrong chart. What do you change?
C and D are real channels and either might help, but both are additions, and the reflex to add is what produces slides where four things compete. Turning the distraction down is cheaper, more reliable, and does not accumulate: do it three times and the slide gets quieter, whereas adding three times makes it louder and no clearer.
2 · Why is annotation ranked as the weakest of the five emphasis channels, given how useful it is?
The ranking measures one thing only: how hard the channel pulls attention. Annotation is last on that measure and first on explanation, which is a different axis. That is exactly why the build-along uses colour to pull and annotation to explain, rather than choosing between them.
3 · You have greyed out your slide and the three-second test still fails. What is the most likely culprit?
Greying out is usually applied to the data and stops there. Corporate logos are typically the most saturated object in the entire template and they appear on every slide, so on a properly greyed slide they can win outright. B is a real failure mode but produces a different symptom, and C is the reflex this session exists to break.
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.
Session 4 cheat sheet · pin this
Every slide is now right on its own. Session 5 is the one where we measure what the order does to all of them together: a real deck on a bench that reads the point's position, the words before it and the number of claims straight out of the rendered slides - including what happens when you add the context slide everybody asks for.