A spreadsheet is not a story, and neither is a slide carrying six charts and no throughline connecting them. How to tell a story with data starts from a simple observation: a story has three parts a spreadsheet doesn’t have on its own: a setup that says what was normal, a tension that says what changed or is at stake, and a resolution that says what should happen next. Remove any one of the three and the numbers stay accurate while the audience forgets them by the next morning.
The setup: what was normal
The setup is the piece most reports skip, usually because it feels too obvious to the person who has spent a month living inside the data to bother stating. It is rarely obvious to the audience encountering the finding for the first time. “Monthly churn has been reviewed for the first half of the year” is a setup sentence: plain, almost boring, and doing real work, the same way a film spends its first few minutes establishing a character before anything happens to them. Skip the setup and a reader has to reconstruct the baseline themselves, mid-sentence, while trying to absorb the actual finding at the same time.
A good setup names the measure, the period, and nothing else. It resists the urge to preview the conclusion, because previewing the conclusion in the setup removes the very tension the next part of the story depends on.
The tension: what changed, and why it matters
Tension is the turn: something moved, something is now at risk, or two things that should agree with each other don’t. “Churn fell from 7.1% in January to 4.0% in June, after a change to onboarding in March” carries a cause, not just a number, and the cause is what makes the sentence memorable rather than merely accurate. Strip the “after a change to onboarding in March” clause out and the sentence becomes a fact nobody has a reason to act on. Leave it in, and the same number now has a story attached that a reader can repeat to someone else without needing the original chart in front of them.
This is also where most charts earn their place. A chart is not decoration bolted onto a finding; a well-chosen chart is frequently the clearest way to show a tension exist at all. A line dropping sharply right where an onboarding change happened does, visually, what the sentence above does in words. The which chart should I use guide covers picking the chart that actually carries a tension like this one, rather than one that merely decorates the slide it sits on.
The resolution: so what happens now
A resolution answers the question every good story eventually forces its audience to ask: so what. Keep the onboarding change in place. Watch the next two reporting periods to confirm the improvement holds rather than reverts. Apply the same change to a different part of the funnel where a similar problem exists. A story that ends on a number and stops, without a stated next step, hands the audience a half-finished thought and asks them to supply their own ending. Some will, most won’t, and the ones who do will not all land on the same one.
The resolution does not need to be dramatic or definitive. “Monitor for one more quarter before drawing a firm conclusion” is a legitimate resolution when the data genuinely doesn’t yet support a stronger one, and stating that limitation plainly is more honest, and often more persuasive, than overreaching on a small sample.
Why the stakes of skipping this are real
One 2015 study on this exact problem is worth citing directly rather than taking on faith. Adobe’s “State of Content” research, covering more than 12,000 consumers across six countries, reported that a majority of respondents abandon content that runs too long, and a similarly large share said poor presentation alone was enough to make them stop reading altogether. That is a decade-old, single-year study of that era’s consumers, not a live measurement of behaviour today, but the underlying mechanism it describes — attention is finite, and structure decides how much of it a piece of content actually earns — has not gone anywhere in the years since. A chart-heavy report with no narrative thread is asking for exactly the kind of patience that research found in short supply even a decade ago.
Examples for beginners: turning a chart into a sentence
The fastest way to practise this, for beginners especially, is to take any chart already sitting in a deck and write the one sentence it is supposed to prove before looking at it again. A sorted bar chart of regional sales becomes “Region three grew fastest this quarter, ahead of the other four combined.” A line chart of weekly signups becomes “Signups doubled the week the referral programme launched, and have held at that level since.” Neither sentence needed new data or a new chart; both simply stated the setup and the tension the chart was already showing, in words a reader could repeat in a hallway without pulling the chart back up.
A second exercise worth trying on the same chart: delete the chart and see if the sentence alone still carries the finding. If it does, the chart’s job is to make the sentence easier to believe and remember, which is exactly what a chart should do. If the sentence collapses without the chart, the chart was probably carrying information the sentence should have stated directly instead of leaving to be inferred from a shape on a screen.
How to tell a story with data: putting the framework to work
Most reports already contain all three pieces somewhere inside them; they are just scattered across an executive summary, a methodology section and a set of appendix charts rather than organised into one throughline. Pulling a report’s setup, tension and resolution into three plain sentences, before touching a single chart, is usually enough to reveal which finding actually deserves the slide’s attention and which chart, of the six that could be built from the same dataset, is the one that shows the tension rather than merely a number.
That is really the whole answer to how to tell a story with data: name the setup, name the tension, name the resolution, then pick or build the one chart that makes the tension visible rather than merely decorative. The assertion-evidence slides lesson picks up directly from here, turning that same three-sentence structure into a specific slide format built around one message and one piece of visual evidence per slide, rather than a topic phrase and a bulleted list underneath it.
Beginners often assume this framework only applies to a full report or a long presentation, but it scales down just as easily to a single slide or even a single sentence in an email. A one-line update carries exactly the same three parts as a thirty-slide deck, just compressed rather than abandoned: setup, renewals have run at 82% for two years; tension, this month they dropped to 71%; resolution, pausing the planned price increase until the cause is confirmed. The framework is a way of organising a finding, not a minimum length requirement, and the smallest data story still needs all three pieces to land.