Presenting data to non-technical stakeholders
Most data presentations follow the order the analysis was done in. Your audience wants the opposite: the answer, what to do, and how sure you are. Here is a structure, chart rules and phrasing that work for owners, boards and department heads.
You have done the analysis and the result is clear. Then you present it and the room is lost by the third slide. The analysis was fine; the translation was not. People who run a business think in decisions, costs and risks, not in distributions and correlations. This guide covers how to present for them.
Start with the answer
Most data presentations follow the order of the work: the data we collected, how we cleaned it, the method, the findings, and only at the end what it means. That is backwards for the audience.
A non-technical audience wants three things, in this order:
- So what? What does this mean for the business?
- What should we do?
- How sure are you? What is the risk if we act on it?
Lead with the conclusion, support it with a few pieces of evidence, and keep the method for questions or an appendix.
A structure that works
1. The question (one sentence). "We were asked whether to add two more people to the sales team next quarter."
2. The answer (one or two sentences). "Yes, if we keep the current win rate. On today's numbers each new rep pays back in about eight months."
3. The evidence (three to five charts). Each chart supports the answer. If a chart does not, cut it, however interesting it is. Typical evidence: the trend so far, how the pipeline looks, what happened last time you hired, and the simple financial model.
4. The risks. "This assumes a win rate of 30%. At 20%, payback stretches to about a year. The main risk is hiring time: if the roles take three months to fill, we miss the busy season."
5. The recommendation. "Hire one now, and the second if the pipeline reaches a set level by the middle of the quarter."
This fits in ten minutes and leaves the rest of the meeting for the decision.
Choosing charts for this audience
The charts that are useful while exploring data are often the wrong ones for presenting it. Keep to a small set.
Use:
- Bar charts to compare categories. Sort them by value. Use horizontal bars when labels are long.
- Line charts for trends over time, with three or four lines at most.
- Single numbers with a comparison: this month, last month, the change.
- Small tables when exact values matter: five to seven rows, three or four columns.
Avoid:
- Pie charts with many slices. Use a sorted bar chart.
- Dual-axis charts. Two scales on one chart confuse most readers. Show two charts.
- Heatmaps, scatter plots and box plots, unless you explain how to read them. They are great for analysis and slow for an audience seeing them once.
- Stacked areas with many layers. Only the bottom layer is readable.
More on this in data visualization mistakes to avoid.
Every chart title is a finding
Not "Revenue by quarter" but "Revenue grew about 25% in Q3, mostly from larger customers." The title tells people what to look for before they look. Someone who reads only your chart titles should still get the whole story.
Making numbers stick
Compare instead of stating. A monthly churn rate of 3% sounds small. Over a year it compounds to about 31% (1 minus 0.97 to the power of 12), so "we lose close to one customer in three each year" lands harder, and is still accurate.
Round. "Revenue grew about 25%" is easier to remember than 24.7%, and the decimal rarely changes the decision. Keep exact figures in the appendix.
Use familiar units. "Our software spend equals roughly two salaries" lets people judge the number against something they understand.
Show the size of the gap, not only the direction. "Down 4%" means little without knowing that the usual monthly swing is 5%.
Handling pushback
"That number is wrong." Stay curious: "What figure are you seeing, and where is it from?" Most disputes come from different definitions, periods or sources, not from errors. Having agreed definitions in advance avoids most of these; the single source of truth playbook shows how to set them up.
"What about X?" If you do not have the data, say so: "We don't track that today. Here is what it would take to get it, and what we can say without it." Honest gaps build more trust than confident guesses.
The discussion drifts. Bring it back to the decision: "Worth looking at separately. For today's decision, the number that matters is this one."
"How sure are you?" Give a range, not a single figure. "Between six and ten months, most likely eight" is more honest and more useful than "eight months".
Dashboard or slides
For a one-off decision, a short deck or written report with five charts is usually best. For numbers the same people review every week, a dashboard is better than rebuilding slides. Dashboard vs. report explains how to choose.
Where clariBI can help
- Reports with the right structure. clariBI reports start with an executive summary, then key metrics with charts, findings, risks and next steps, which is close to the structure above. Every number in the text is checked against the chart data before the report is saved. See report sections. Download it as a PDF and add your own context and recommendation; the report cannot know your plans or constraints.
- Live dashboards instead of screenshots. On Professional and Enterprise, a dashboard or report can be shared through a public read-only link, or added to a team workspace where people can comment. The sharing docs explain each option. A link always shows the current numbers, so nobody argues about which version of the deck is right.
- Answers during the meeting. When someone asks "what about last year?", a plain-English question in Ask clariBI can often answer it on the spot, with a drill-down that shows where the number came from.
For your next presentation, try one change: rewrite every chart title as a finding and put the recommendation on the second slide. Then watch how much shorter the discussion about the numbers gets.