When you're using data to tell a story, one number is a fact, two numbers are a comparison, and three or more become a story only if you sequence them so each point answers the question the one before it raised. I've sat through quarterly reviews where every single slide was accurate and the whole thing still went nowhere. The numbers were fine. The order was an accident.
Why this matters (and what most executives get wrong)
A narrative arc in a data presentation works the way it works in a film. There's a starting point that establishes the world as it is, a development that shows what changed, and a destination that says where things are heading. Your data points are the evidence at each stage. The narrative is the thread.
That container has a name I use constantly: setup, conflict, resolution, the three-part cinematic structure for a data story. Most people can describe the shape. Far fewer can tell you which of their numbers belongs in which beat, or why point four should come before point two.
Communication researchers call the underlying idea the narrative paradigm: people judge what they hear partly by whether it hangs together as a story, not only by whether the evidence is sound. Give a room an accidental order and it builds an accidental story.
The 4-step method for using data to tell a story across several points
Here's the practical method I use and teach, all of it before you open the presentation software.
1. Write the three sentences first
Put three sentences on paper. Where things stand right now. What changed to create that situation. What it means and where we go next. That's your arc, compressed. Everything else in the deck is evidence for one of those three sentences. Can't write them? You don't have a story yet. You have a folder of charts.
2. Choose the order that builds, not the order the data arrived
Chronological is a default, not a decision. Most decks run in the order the analysis happened, which made sense to the analyst and nobody else. Sort your points by what each one sets up. Sometimes the newest number goes first, because it's what makes the older ones matter.
3. Make every point answer the question the last one raised
This is the test that fixes most of it. After each data point, ask what the person in seat three now wants to know, then make your next point that answer. Retention dropped. Why? The drop sits in accounts under a year old. What changed there? Onboarding time doubled after the migration. Now the room follows you instead of waiting.
4. Send everything that serves none of the three sentences to the appendix
Run every candidate slide through the filter. Serves sentence one, it's setup. Sentence two, development. Sentence three, recommendation. Serves none of them and it stays out of the flow, no matter how many hours went into it. The appendix is right there.
The mistake most executives make
They present in the order the work happened: data gathered, then cleaned, then findings, then a summary slide. That's a tour of everything you know rather than a path somewhere that matters. I've watched a senior analyst spend six minutes on methodology before a single result reached the screen. The room had already decided how much to care.
There's a second version, and it's the one I'd cut first: announcing the arc out loud. "I'll cover the background, then the analysis, then my recommendation." Don't. Those sentences are a roadmap for you, not for them. Slides are the evidence. You're the narrator, and the clicker doesn't get to lead.
Case study: the clinic review that changed only the order
What follows is a composite, generic on purpose. A clinic operations lead at a hospital system runs a monthly review that always lands in one order: visit volume, then cost per visit, then wait time, then staffing, because that's the order the report generates. Every slide true. Nothing connected.
Nothing gets cut in the fix. The points get reordered. Wait time moves to the front, because it's the number the room actually feels. Visit volume follows, because it explains the wait. Cost per visit comes next, since it shows what absorbing that volume costs. Staffing lands last, as the recommendation rather than a footnote. Same numbers, new sequence. The room changes too: people ask about the decision instead of what the definitions on slide 14 mean.
| Your sentence | Where it lands | The question it answers |
|---|---|---|
| Where things stand right now | Setup | What is true today? |
| What changed to create that | Conflict | How did we get here? |
| What it means and where we go | Resolution | So what do we do? |
| Everything else | Appendix | Nothing the room asked |
Frequently asked questions
How many data points does it take to tell a story?
One data point is a fact. Two are a comparison. Three or more, connected deliberately, become a story. The count is not the thing. What matters is whether the points are sequenced so each one leads into the next on purpose. Three unrelated numbers are still three facts.
Should a data story always run in chronological order?
No. Chronological order is a default, not a decision. Time order is right when change over time is the point, and wrong whenever a later number is what makes the earlier ones matter. Choose the order that builds toward the decision instead.
How is a narrative arc different from setup, conflict, resolution?
Setup, conflict, resolution is the three-part cinematic structure for a data story, the container the whole presentation sits inside. The narrative arc is what you do with individual data points inside it: which number belongs to which beat, in what order, and how each one sets up the next.
Should I tell the audience which structure I am using?
No. Your three sentences are a roadmap for you, not a structural announcement for the room. Saying out loud that you have a beginning, a middle and an end makes a presentation feel like a syllabus. Let the sequence do that work silently instead.
What to do next
Open the deck you're presenting next and write its three sentences before you touch a slide. Then lay your data points in a row and check one thing: does each point answer the question the one before it raised? For the wider method, start with what data storytelling really is, the pillar behind Story-Driven Data™, and how storytelling works inside a presentation. If the sequence has to survive a hostile room, get a quick quote and we'll rebuild the order on camera.