You will be able to simplify a chart so the audience sees its point within seconds.
Open the last slide deck you presented and find a chart. Count everything on it that isn't the data itself: gridlines, borders, a legend, axis titles, a label on every point, shadows, any 3D effect.
Picture a chart with twelve months along the bottom, five coloured lines, gridlines every ten units and data labels on every point. It comes up on screen and the speaker says "As you can see." The audience can't see it yet, because they are still matching colours to the legend. Charts in talks most often fail because they show everything in the data, when the audience only needs to see one thing. A chart in a report can carry detail because the reader has time. A chart on a slide gets a few seconds of attention while the speaker talks, so it has to make its point in that time.
Before you pick a chart type, ask what comparison you are making. Most charts in work talks make one of two: a comparison between categories, or change over time.
To compare categories at one point in time, such as regions, products or teams, use bars. Bars are easy to compare by length, and the eye reads them quickly. If the category names are long, use horizontal bars so the labels fit without being tilted.
To show change over time, such as monthly sales or weekly sign-ups, use a line. A line shows direction and turning points at a glance. Twelve months of data as twelve separate bars can work, but a line usually makes the trend clearer.
Pie charts are popular, but they are usually the weakest choice for a slide. People find it hard to compare slice sizes, especially with more than two or three slices. To show that one part is a large share of a whole, a single bar or a plain sentence with the number often works better.
Charting software adds clutter by default: gridlines, borders, a legend, axis titles, data labels on every point, shadows and 3D effects. Remove all of the defaults first, then add back only what the audience needs to read the point.
Gridlines can usually go. If an exact value matters, label that one value directly. A legend makes the eye travel back and forth, so label each line or bar directly at its end instead. You can often drop axis titles if the headline and labels already make the units clear. Never use 3D effects, because they distort the sizes of bars and slices and add nothing.
Once the chart is clean, show the audience where to look. Put the one bar or line that carries the point in a strong colour and make everything else light grey. The grey elements still show the comparison, but they don't compete for attention.
This works for line charts too. Say one region fell while four others grew. Give the falling line a strong colour and turn the other four grey, and the audience sees the story first and the detail second.
Don't rely on red and green together to carry meaning. People with the most common forms of colour blindness find that pair hard to tell apart. A strong colour against grey avoids the problem.
The headline above a chart should state what the chart shows, as a sentence. This applies the rule from Lesson 4.1, "One idea per slide, stated in the headline". A label names what the chart contains without making a point, for example "Sales by region, FY2024". A point states what the chart shows, for example "East region sales fell while every other region grew".
With a sentence headline and the key line highlighted, the audience reads the sentence, sees the line and understands the slide in a couple of seconds. You can then spend your speaking time on why it happened and what to do about it.
Put the data source on the chart in small type beneath the headline. Lesson 3.3, "Say numbers people can understand", explains why this matters.
Arjun is presenting a chart of missed appointments. His first version has all the default clutter. He strips it back to seven bars, one per day of the week, with each bar's percentage above it and nothing else. He colours the Monday bar dark blue and leaves the other six grey, so the audience knows the chart is about Monday before anyone reads a label. His headline reads "Mondays lose almost twice as many appointments as any other day." Under it, in small type, he adds the source: booking system, January to June.
Test every chart before you use it. Look at it for five seconds, look away and say what it showed in one sentence. If you can't, the audience won't be able to either. Then ask someone else to do the same check. If what they say differs from your headline, the chart and the headline are not yet telling the same story.
Now take one real chart from your own work, such as the one you counted at the start, and put it through every step: match the chart type to the comparison, strip the defaults and add back only what the point needs, highlight the one bar or line that matters, add the source in small type, and run the five-second check yourself and then with someone else. Finish with a headline that states what your chart shows.
Pick one chart from your work. Strip it back, highlight the key data, and write a headline that says what it shows.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).