Data

Data Visualisation Best Practices

Effective charts make decisions easy. In presentations, a chart's job is to prove a single point quickly. Follow these rules to make your data work for you, not against you.

In business settings, weak visualisation creates expensive misunderstandings. Teams delay decisions, executives ask for follow-up analyses and presenters lose credibility because the audience cannot see the point fast enough. Strong visualisation prevents this by making one conclusion obvious at a glance.

Choose the right chart

Bar charts compare categories. Line charts show trends. Scatter plots reveal correlations. Tables are for exact values; use them sparingly in slides.

Use a chart selection decision tree

When deciding on a chart type, start with the question you need to answer:

Are you comparing categories? Use a bar chart.
Are you showing movement over time? Use a line chart.
Are you showing composition? Use a stacked bar chart, not a complex pie chart.
Are you testing a relationship between two variables? Use a scatter plot.
Do you need exact values for audit purposes? Use a table in the appendix.

If your slide needs two chart types to tell one story, split the slide into two. Clarity usually improves when each chart answers one question only.

Choosing by business context

The chart type decision tree above tells you what is technically correct. Business context tells you what is persuasive. The same data can be presented in multiple valid ways and the best choice depends on what your audience needs to do after seeing it.

Approval context: Your audience needs to say yes or no. Show a single metric against a threshold. One bar, one line, one number with a target line. Remove everything else.
Diagnostic context: Your audience needs to understand why something happened. Use a breakout view: trend line with annotated events, or a waterfall chart to show where the change occurred across contributing factors.
Monitoring context: Your audience needs to detect drift. Use consistent period-over-period comparisons. A rolling four-week trend will reveal noise; a year-on-year comparison will reveal signal.
Exploration context: Your audience is forming hypotheses. A scatter plot with labelled outliers is often more useful than a summary chart because individuals want to probe specific points.

Knowing the context before choosing the chart prevents the most common error in business presentations: giving an exploration chart to an approval audience. They will produce questions instead of decisions.

Highlight the single metric

Annotate the key number and make it visually dominant, a coloured callout or large bold label, so your audience immediately understands the takeaway.

Simplify axes and labels

Remove unnecessary gridlines and tick marks. Shorten axis labels and use consistent units. If the audience needs detail, provide a downloadable appendix, not a crowded slide.

Build for accessibility and room conditions

Charts often fail in live rooms because they were designed on a high-quality laptop display and never tested in a typical meeting setup. Treat accessibility as a baseline requirement.

Use high-contrast colour pairs and avoid pale lines on dark backgrounds.
Keep chart labels at readable sizes for projected screens.
Do not rely on colour alone to signal category differences. Use labels or patterns where needed.
Avoid red-green only distinctions when status or variance is critical.

Common anti-patterns to avoid

Dual-axis confusion: Two scales can create false relationships.
Decorative 3D charts: Depth effects hide values and reduce trust.
Overloaded legends: If the legend is doing too much work, the chart is too complex.
Axis manipulation: Truncated axes can exaggerate minor differences and undermine credibility.

When in doubt, choose the simpler visual and write a clear takeaway sentence above it.

Real-world scenario walkthroughs

The rules above are easier to apply when you can see them in concrete situations. The three scenarios below cover the most common failures in business presentations and the specific fix for each.

Scenario 1: The quarterly revenue review

What most presenters do: Place a twelve-month bar chart showing all revenue lines by product, region and channel on one slide. The chart has eleven data series, a cluttered legend and no annotation. The title reads "Q3 Revenue by Segment."

Why it fails: The audience is looking for the reason revenue missed forecast by 8%. A full-coverage chart forces them to do the analysis themselves, which is the presenter's job.

The fix: Split into two slides. Slide one uses a waterfall chart showing the four categories that drove the shortfall, with the largest bar annotated: "Enterprise segment down 14% due to delayed contract renewals." Slide two shows the full breakdown for those who need the detail. The title on slide one becomes "Enterprise contracts drove the entire Q3 shortfall." That title restates what the audience now needs to act on.

Scenario 2: The sales pipeline update

What most presenters do: Use a pipeline funnel that shows raw opportunity counts at each stage. The funnel looks healthy, but the sales director is asking why close rates have dropped.

Why it fails: The raw counts hide the conversion ratio between stages. A funnel of 200 opportunities at proposal stage means nothing without the conversion rate from the previous month to compare it against.

The fix: Replace the raw funnel with a stage-by-stage conversion rate comparison, current month versus prior month, shown as two simple bar pairs side by side. Annotate the stage where conversion dropped. The title becomes "Proposal-to-negotiation conversion fell from 42% to 28% in October." This is one insight, one chart and one decision: investigate proposal quality or response time at that stage.

Scenario 3: The operational performance dashboard

What most presenters do: Copy five KPI widgets from a business intelligence dashboard into a slide. The result is a slide with five numbers, five trend lines and five status indicators in five different colours.

Why it fails: A dashboard is designed for self-service exploration. A presentation slide is designed for a shared conclusion. The audience scans five metrics and forms five different opinions about what matters.

The fix: Select one metric that is off-track, or one that is performing well above expectation and give it the full slide. Show the metric as a large number, with a twenty-six-week trend line behind it and a red or green annotation marking the threshold. If multiple metrics need discussion, use one slide per metric. The additional slides can stay in the appendix for sessions that need them.

Industry-specific metric selection

Different functions and audiences have different default tolerances for chart complexity. Calibrating your chart choices to the industry context reduces back-and-forth requests for clarification.

Sales and revenue

Sales audiences respond fastest to charts that show momentum: are we ahead or behind and by how much? Avoid full-year cumulative views unless the meeting is a year-end review. Use rolling four-week or quarter-to-date comparisons. Annotate the gap to target visibly. The key calculation your chart should answer: "What is the gap and how many weeks do we have to close it?"

Finance and reporting

Finance audiences expect precision. Truncated axes undermine credibility in this context more than in any other. Use waterfall charts for variance explanations, not bar-over-bar comparisons. Always show the prior period as a reference. Label exact numbers when the difference between 4.1% and 4.3% matters to the decision. If your chart cannot show the number to the required precision, use a table for that row and a chart for the trend.

Operations and logistics

Operational audiences need to detect and locate problems quickly. Control charts, which show a metric over time alongside its expected range, are more useful here than simple trend lines because they make anomalies visible at a glance. If a control chart is too technical for the room, use a simple line chart with a shaded band indicating the acceptable range. Any point outside the band should be annotated with the date and the cause, if known.

Product and growth

Product audiences work with ratios and cohorts. Absolute numbers often mislead in a growing user base. Use rate-based metrics (activation rate, day-seven retention, weekly active users as a percentage of registered users) as the primary chart and show absolute counts in a supporting note below the chart. Cohort comparison charts, which show behaviour of different user groups across the same time window since signup, are difficult to replace with simpler alternatives when retention is the subject of discussion.

Executive and board level

Executive audiences want fewer charts with stronger conclusions. Three charts with clear recommendations outperform fifteen charts with ambiguous data every time. Prioritise charts that compare actuals against a plan, forecast, or prior period. Always provide the implication on the same slide; do not make the executive derive it from the visual. Treat any chart that requires more than five seconds of reading as a candidate for simplification or removal.

Design Better, Faster

SlideCut automates alignment and formatting so you can focus on your story.

Install SlideCut Free

Tell a story with data

Arrange charts in a narrative sequence: context → trend → implication → recommendation. Each visual should answer one question that moves the story forward.

Data slide quality checklist

The title states the conclusion, not the chart type.
The key number is highlighted and labelled clearly.
Axes, units and dates are consistent and visible.
Only one primary message appears on each chart.
The recommendation is explicit on the same slide or the next slide.

Frequently asked questions

How many metrics should one chart include?

In most presentation contexts, one primary metric and one supporting comparator are enough. Additional series should be moved to backup slides unless they are essential for the decision.

Should I show raw data tables in the main deck?

Only when precision is the objective of that slide. For executive communication, show interpreted visuals first and provide raw tables in an appendix.

What is the fastest way to improve an existing weak chart?

Rewrite the title as a conclusion, remove non-essential series, increase contrast and add one annotation that explains the implication.

My executive said the chart was confusing. What should I check first?

Start with the title. If it describes the chart type rather than the conclusion, rewrite it. Then check the number of data series: if there are more than three, remove the least important ones. Finally, check whether the most important number is the most visually prominent element on the chart. If you have to look for it, your audience will too.

I have twelve data series and they are all important. What do I do?

They are not all equally important to the decision being made in that meeting. Identify the two or three that drive the current decision and give them the main chart. Move the remaining series to a small supplementary table below the chart or to an appendix slide. If every series genuinely affects the decision being made, you likely need a different meeting structure, not a bigger chart.

When should I use a pie chart?

Pie charts work when you have two to four segments and your goal is to show part-to-whole composition, not to compare values precisely. If the segments are similar in size, the audience cannot reliably distinguish them by angle. Use a bar chart instead. If you are tempted to use a pie chart with more than five segments, use a stacked bar chart and sort the segments by size to make comparison immediate.

How do I handle charts that look fine on my laptop but become unreadable when projected?

Project the deck before the meeting if you have any access to the room or a similar screen. If not, apply the following rules as a minimum: font sizes should be at least 18pt for labels and 24pt for titles; line weights on charts should be at least 2pt; colour differences between series must be distinguishable without colour by using different line styles or textures. Test by standing three metres from your own laptop screen and checking whether you can still read the annotation without leaning in.

My chart shows a positive trend but the audience is still sceptical. Why?

The most common reason is a truncated axis. If your y-axis does not start at zero, a modest gain can appear dramatic. Your audience may have noticed this, even without articulating it. Use a full-axis view to establish trust and if the change genuinely looks small, annotate the percentage or absolute change explicitly so the relevance is clear regardless of the visual scale.

Conclusion

Good visualisation is ruthless editing. Remove distractions, emphasise the insight and make the action obvious. When in doubt, simplify.