Data Storytelling Guide: Use a storyboard to present data like a Pixar filmmaker

Stop overwhelming people with numbers. Start with the takeaway.

This article aims to help designers, product managers, and data scientists pitch data-driven ideas and present UX research. Write stories that don’t overwhelm people with numbers. Prevent data paralysis.

Let’s assume that dozens of data points have been pulled as part of an exploratory analysis. Perhaps to assess impact of a product change, to estimate the value of a new market opportunity, or to guide a Go, No-Go decision. Insights are littered all over the place.

So what’s next? How do we synthesize all these insights? How do we go from dozens of plots to “Go, No-Go” decisions?

Toy Story 3 cost ~$200M to make. Not cheap. So before the team invests in building 3D models, composing music, and hiring actors, they storyboard. The entire film is planned by sketching potential shots, re-working them, and getting the storyline just right. So only the necessary animations are produced. Nothing more.

Designers, product managers, and data scientists can also leverage storyboards. They can help visualize user experiences and draft analysis stories. They allow us to see the big picture, organize plot points, and add/remove insights to better convey ideas. It prevents us from wasting time investigating hypotheses that don’t matter or creating beautiful charts that won’t be used. It serves as a plan for our analysis write-up.

Instead of sketches, analysis storyboards are made from takeaway statements. These takeaways are written by reviewing each data point from our exploratory analysis and asking: So what? What is the data telling us? What action does it empower? What does it explain?

This leads to takeaway statements that look like:

• Opportunity to ___
• ___ caused ___
• ___ is growing due to ___

Next let’s group takeaways into three key themes of our storyline:

1. What is the problem?
2. Why is the problem important?
3. What is the solution and its expected/actual impact?

#1 and #2 are often explored together as part of problem definition, while #3 tends to be performed later during solution design. This follows the double diamond process and works for projects of all sizes: From why visits were down today to why our market share declined x% last year.

Now that our takeaways are grouped, let’s tidy things up. Three ways to sort takeaways include:

• Top-down: From highest level insight to more granular segments. This works best when the audience has the time to learn the ins and outs of the analysis. For example, we’d start by revealing total annual sales, then segment to monthly, by regions, by product lines, etc.
• Decreasing ROI: From highest impact information to lowest impact data points. This tactic works for busy executives that may only have time to learn and provide input on key issues and skip the rest. So don’t waste time and start with a bang.
• Surprises at the end: The idea is to keep debates and discussions for the end and not let hot topics distract from important and non-debatable insights at the beginning. Start with non-debatable insights and move toward more contentious points.

Pro Tip: Simplify. Move any takeaway that distracts from the main storyline to the appendix.

Avoid keeping all insights and data for completeness sake. They can make the story difficult to follow. Imagine Jason Bourne is being chased and ends up on the stage of a Pitch Perfect singing competition. It makes no sense. So remove irrelevant takeaways.

At this point, takeaways are grouped by themes and logically sorted. A storyline is born. It should read like a novel with a start, a middle, and an end. Otherwise, try re-wording the takeaways to make them flow.

It’s very difficult to edit one’s own story. Tunnel vision and blinding biases tend to set in. So ask for help.

It’s a good idea to include rough numbers and visuals to complement takeaways for a peer review. A complete analysis yields more holistic feedback. But don’t spend time on making them look pretty just yet. We’ll likely have to iterate many times before aligning on what charts and numbers to show. Just copy/paste numbers from the exploratory analysis.

Beyond validating the numbers and insights, ask the reviewer:

• What data is missing? These are gaps to fill
• What insight is too much information and distracting? Remove.
• What made no sense? Change the wording or remove.
• How was the pace or flow? Re-organize.

Make sure the takeaways still flow one after another after changes.

On data visuals, I won’t reinvent the wheel that Edward Tufte built in The Visual Display of Quantitative Information. I will however share the 3 key elements to driving a point across:

1. A section title that communicates the takeaway and answers “so what?”
2. Evidence in the form of data points, charts, or tables
3. Supporting context that explains why the data behaves the way it does

This works for essays, articles, and also decks as long as the text is short. Decks benefit from longer talking points that are unwritten, so people focus on what you say instead of reading ahead.

Other helpful references to make things pretty include HBS’s guide on basic chart types, Colombia University’s guide on the basic structure of a consulting deck, and fellow UX writers’ tips here and here.

Notice we saved beautifying for last. Time would otherwise be wasted on takeaways that are moved to the appendix.

Pro Tip: Feel the data

Data alone is often insufficient to drive a point across. It feels impersonal. So complement with a backstory. Showcase a character or person facing the problem. This helps the audience relate on a human level.

For example, telling the story of a driver who killed two pedestrians while running Tesla Autopilot generates much stronger reactions to the challenges facing self-driving technology than the fact Autopilot was linked to 273 crashes over 2021–2022.

So share a backstory using film, pictures, or even personal quotes.

Pro Tip: What good looks like

Here are a few of my favorite analyses and presentations:

• Pew Research center provides critical research on changing behavior of Americans, including this article on American’s sentiment on driverless cars (where I pulled the data and charts for our article).
• Our World in Data’s article on plastic pollution masterfully shows key insights accompanying the data.
• As example of an effective backstory to complement the data, here’s National Geographic’s article on plastic pollution.
• Finally, here’s a simple and data-driven fundraising deck made by my former colleagues to launch a new startup called Common Paper, resulting in millions in seed money as they standardize legal contracts.

Pro Tip: What bad looks like

It’s just as important to see what bad looks like, so allow me to showcase three common mistakes:

Bad Example 1: Section titles describe the content type rather than the content itself. It’s a waste of space.

Bad Example 2: Text that describes the chart rather than complement with context and answers why the data behaves this way. It’s redundant.

Bad Example 3: Exposing too many data points in the same section can overwhelm the audience and lead to data paralysis. Split into separate sections or move distractions to the appendix.

For more Design x Data guides, see creating your analytics strategy and using data to drive product design.