Effective data visualization can make or break a presentation. I saw this firsthand during a performance review for the airline’s loyalty program, where poor charts completely derailed the meeting.
The speaker was just beginning her presentation when she made two fundamental mistakes.
Not even two minutes into the meeting, she said:
“Do not believe all the data that you’ve seen, because we’ve found a lot of discrepancies.”
The second mistake came right after: she showed a crowded chart with conflicting axes that accounted for different things. In other words, the chart itself was confusing and misleading.
She lost the entire audience in just three minutes.
You can imagine how the rest of the presentation went. You can also imagine why the whole thing was wrapped up in 10 minutes.
Discrediting yourself is a recipe for disaster.
But this post is not about that; it’s about data visualization.
What Makes Effective Data Visualization: A Conversation with James Eagle
I recently met with data visualization legend James Eagle (KILLER CHARTS) and spent some time discussing what actually works in data visualization, and what we should be teaching.
In our Podcast we discussed a lot about Data and Visualizations: Here are a few of his key nuggets:
1. No story, no charts
If you don’t have a story, don’t even start putting charts together.
A chart without a clear narrative is just decoration. Before you open your BI tool, ask yourself:
- What is the question I’m trying to answer?
- What decision should this chart help someone make?
- What is the one message I want them to walk away with?
2. If words do it better, skip the chart
If something is better said with words, you might not even need a chart.
Sometimes a single, clear sentence like:
“Customer churn increased by 15% in Q3, mainly driven by price-sensitive segments.”
…is far more effective than a complex visualization.
3. You have three seconds
When it comes to visualization and decision-making, you literally have about three seconds to make an impact.
If your chart is not clearly understood almost immediately, you’ve failed.
Pledging Allegiance to Simplicity in Effective Data Visualization
With all this in mind, James makes a couple of pledges for simplicity, even though his own charts are some of the coolest, most dynamic data visualizations you’ll ever see.
Here are some of his practical recommendations.
Static Bar Charts: The Backbone of Effective Data Visualization
If you’re unsure which chart to use, go for a static bar chart.
Why?
- It communicates a lot with very little noise.
- It shows differences very clearly.
- It can represent evolution over time (with grouped or stacked bars, if needed).
- It’s easily understood by almost everyone.
Simplicity beats sophistication when decisions are on the line.
Pie Charts and Effective Data Visualization: Use with Care
The most overused chart: the pie chart.
Imagine slicing an actual cake like this
Guidelines:
- Anything with more than five slices will likely confuse people.
- Too many segments make the chart visually cluttered and hard to read.
- Humans are not great at comparing angles; bars are usually better for comparison.
If you need to show many categories, consider:
- A bar chart sorted from highest to lowest
- Grouping smaller categories into “Other.”
Reserve complex charts for expert audiences
Charts like scatter plots and other multivariate visualizations are powerful, but they are better suited for very knowledgeable audiences.
Yes, cool, but what do I do with this?
If you have a mixed crowd: executives, non-technical stakeholders, or people seeing the data for the first time, keep it super simple:
- Lines
- Bars
- Clear labels
Clarity beats cleverness.
The Goal: Support Decisions, Not Impress People
Watch the full interview with James for a deeper dive into these ideas. He shares a huge amount of practical wisdom about how to make data visualizations that are:
- Effective
- Easy to understand
- Truly supportive of decision-making
At the end of the day, your job is not to impress people with fancy charts.
