Don't Read the Table During a Presentation: Three Sentences per Figure
At group meetings, there's a way of presenting that I heard for three years: put the table on the screen and read it from top to bottom—"The first row is Group A, mean 12.3; the second row is Group B, mean 11.8..."
After three minutes of reading, the audience remembered nothing.
I used to present this way too. My advisor interrupted me: "We can see all the things you're reading. Tell us what you see in them."
Three Problems with Reading Tables
**Problem 1: Numbers don't carry a point.** Every number in the table is correct, but the audience doesn't know which one matters.
**Problem 2: The audience is searching while you are reading.** Eyes are faster than ears; the audience is already on the fourth row while you are still on the first.
**Problem 3: No conclusion, no discussion.** After you finish reading the table, the meeting room goes silent for three seconds, and then someone says, "Let's take another look."
The Three-Sentence Method: What This Is, What We See, So What
Use only three sentences for a figure (or a table):
1. What this is: axes, groups, sample size (10 seconds)
2. What we see: one conclusion with one number (20 seconds)
3. So what: what it means for the research question (10 seconds)
Here's an example of my own:
"This figure shows the strength of three groups of samples at four temperature points, with three replicates per group." "After 120 °C, Group A decreases the fastest, 18% lower than Group B." "So we'll use 120 °C as the cutoff going forward and focus on the failure path of Group A."
Three sentences, forty seconds, and the audience knows where to look.
When to Keep Tables
Tables aren't unusable; just don't use them to talk about trends.
My current rule:
| Content | What to use |
| Trends, comparisons, distributions | Figure |
| Exact values, parameters, conditions | Table (place on backup slide) |
| Processes, structures | Diagram |
Put tables on backup slides and turn to them only when asked, "What's the exact value?"
One Page, One Figure, One Conclusion
My presentation slides now basically consist of: one figure + one conclusion (written above the figure) + one line of conditions (written in the figure caption).
The conclusion goes above the figure because the audience looks at the title first, then the figure. A title like "Group A decreases fastest after 120 °C" is far more useful than "Comparison of strength across three groups."
Two Tips for Presenting Data
Tip 1: State the difference first, then the numbers. "Group A is clearly lower" registers before "11.8 versus 14.4."
Tip 2: Give a reference point. "18% lower" is easier to understand than "2.6 lower," because percentages have a reference.
How I Changed It
The first step wasn't changing the figures; it was changing the script: write what you'll say for each slide in three sentences. If you can't write three sentences for a slide, that slide can be deleted.
The second step was moving tables to backup slides.
The third step was redrawing the figures: remove gridlines, add units, and put the conclusion in the title.
For the layout step, I use MuchSci's office assistant agent to produce a first draft; it can lay things out according to the "one page, one figure, one conclusion" format. MuchSci only handles the layout—I still decide the figures and conclusions myself.
A Real Comparison
With the same data, the table-reading version took four and a half minutes, and no one asked questions afterward; the three-sentence version took one and a half minutes, and there were three questions afterward.
More questions mean it was explained clearly—the audience can only ask questions if they understood.
A Few Takeaways
A presentation is not "showing the data"; it's "reading the data on the audience's behalf."
The essence of the three-sentence method is: you look through it for them first, then tell them where to look. For me, MuchSci only appears at the layout and sentence-editing step; which sentence is the conclusion is something I have to read out of the data myself. Once the script is finalized, I save it in MuchSci's personal knowledge base, and next time I present, I just revise three places.