Data storytelling
How to present data that actually gets decisions made
Knowing SQL and Tableau is not enough — you need to tell the story behind the numbers. Here is how data analysts turn insights into decisions.
Why data storytelling matters
Most dashboards are ignored. Stakeholders open them, see numbers, and move on without changing anything. The problem is rarely the data — it is the presentation.
The analysts who advance are the ones who connect data to decisions. Technical skills get you the job. Communication skills determine how far you go. Your job is not to show data — it is to drive action.
A chart that surprises no one and recommends nothing is just decoration. A chart that changes what someone does next is analysis.
The data story structure (3 acts)
Every strong data presentation follows the same underlying structure — the same one screenwriters, journalists, and lawyers use. It is not a coincidence. The structure works because it matches how humans process information.
Setup
What is the business context? What question are we answering?
Ground the audience before you show them anything. What decision is on the table? Who is affected? What time period are we looking at? Without this, data floats in a vacuum.
Confrontation
What does the data show? What is surprising or concerning?
This is the finding. Present it directly — the number, the trend, the comparison. Name what is unexpected. If nothing is surprising, question whether the analysis is worth presenting.
Resolution
What should we do about it?
Give a recommended action with a confidence level. 'We are 80% confident this is the cause and recommend pausing the feature until we investigate.' Uncertainty is not a reason to omit a recommendation.
The 5 chart types and when to use each
Most analyses need only one or two of these. The mistake is picking a chart because it looks interesting. Pick a chart because it matches the question you are answering.
The most common chart mistakes
These are the mistakes reviewers notice immediately — and that undermine trust in the entire analysis, regardless of how good the underlying work is.
- ✕
Truncated y-axis (makes small differences look huge)
- ✕
Too many metrics on one chart (pick the most important one)
- ✕
No chart title that states the insight (title should be the takeaway, not a label)
- ✕
Wrong chart type (pie for time series, bar for proportions)
- ✕
Unlabeled axes
The insight sentence formula
Before you open a slide deck, write this sentence. If you cannot write it, you do not yet have an insight — you have observations. The sentence forces you to connect a metric to a business consequence.
Formula
[Metric] went up / went down / stayed flat by [X%] vs [comparison], which means / suggests / indicates [business impact].
Example
"Day-7 retention dropped by 18% vs last month, which suggests the onboarding changes in Sprint 12 reduced early user engagement."
Notice: the sentence names a metric, a magnitude, a comparison period, and a hypothesis about cause. All four elements are required. Missing any one of them makes the sentence weaker and the action less clear.
How to structure a data presentation
The biggest mistake analysts make is building their slide deck in the order they did the work — cleaning data, exploring, hypothesizing, concluding. Your audience does not care how you got there. They care about the conclusion and what to do about it. Lead with that.
Slide 1
The one-sentence takeaway
Your conclusion, not your method. Lead with the finding.
Slides 2–4
Evidence
2–3 charts that support the conclusion. Each chart earns its place.
Slide 5
Recommendation
What to do and why. Be specific — 'investigate further' is not a recommendation.
Slide 6
Caveats and limitations
Data quality issues, sample size concerns, and what you could not measure. This builds credibility.
Keep building
Build data skills in the Data Analyst track
SQL, visualization, statistical thinking, and stakeholder communication — the full skill set that separates analysts from spreadsheet users.
Explore the Data Analyst track