Data Storytelling: Separating Fiction From Facts

Data is playing a larger role in day-to-day business conversations than ever before. The ability to communicate with data is now a necessity for business leaders, frontline employees, and everybody in between. People who may have easily avoided discussing data in the past are finding numbers being thrust upon them. When data is a foreign language to you, it can be frustrating to not understand what’s being said or be able to use it effectively in communications with others. Not being conversant or fluent in data is quickly becoming a liability in today’s fast-moving data economy.

Fortunately, we can turn to a timeless, highly effective form of communication—storytelling—but with a modern data twist. Now instead of illustrations, pictures or film, we can communicate with data by bringing stories to life with compelling data visualizations. For example, the combination of a compelling narrative and insightful data charts can help explain why your sales are down and create urgency to fix problems in your sales channel. Data storytelling isn’t just a skill reserved for analysts or data scientists—it’s one I believe we must all master. Self-service analytics solutions are increasingly placing a wealth of information at people’s fingertips. Now more people are put in a position to formulate and share data stories that can drive action and change.

Even though the field of data storytelling is still relatively new and uncharted, it has become a hot topic in both analytics and business circles. As various people step forward to provide opinions on how to tell data stories, I’ve seen misinformation creep in which—if left unaddressed—could lead aspiring data storytellers astray. I’d like to separate the fiction from the facts by highlighting some of the recent misconceptions about data storytelling I’ve encountered:

Fiction #1: Every data visualization tells a story. While a story may potentially exist within any data visualization, that doesn’t mean the chart or graph actually communicates it clearly or effectively. Every data story should convey or build up to a main point or idea. A data visualization fails to tell a story when your audience doesn’t clearly understand the central insight or idea that’s being conveyed. Multiple data points in the same data visualization (or across multiple data visualizations) can often compete for attention and create noise. Unless you purposefully draw your audience’s attention to a specific insight or series of insights (the signal), your data visualization won’t support a data story. I’ve seen interactive data visualizations where the author had no intended destination for his or her audience. While these charts may empower audiences to discover their own insights and data stories, they don’t tell a story.

Fiction #2: Data storytelling is only about data visualization. While data visualizations are often a central focus of data stories, the narrative aspects are equally important to the success of a data story. A story is not just a bunch of pretty pictures—it is an account of a causally related or connected series of events. The narrative elements in a data story provide structure and help to create meaning from an assortment of data points that might otherwise feel quite random. In addition, narrative elements also help your data story to engage your audience and connect with them on an emotional level, which data could never do on its own. The skillful combination of data, visuals, and narrative is essential to any meaningful data story.

Fiction #3: Data stories have a beginning, middle and end. In this case, the fictional aspect is not that this statement is incorrect but that this piece of advice is actually useful to would-be data storytellers. The idea that stories are structured this way came from the Greek philosopher Aristotle. However, a textbook has a beginning, middle and end—so too does a financial report or a phone book—yet none of them would be considered stories. In terms of the dramatic structure of a data story, I find Freytag’s pyramid more useful with its introduction (exposition), inciting incident, rising action, climax, resolution and conclusion (denouement). Establishing the setting and characters at the beginning of a data story is important as well as highlighting how something has changed in the environment that introduces a new problem or opportunity (inciting incident). Leading your audience to the main insight (climax) and then to recommendations / next steps (resolution) is a more helpful structure for data stories than Aristotle’s model.

Fiction #4: Data are the characters of your story. I often see some confusion about how the different narrative elements are represented in a data story. A common mistake centers around data being perceived as the characters or heroes of a data story. While you might be enamored with a particular insight, people—not data—should be the heroes of your data story. Most business data is either directly or indirectly related to people—your customers, prospects, employees, partners, investors and so on. An audience is going to care about what happens to different groups of people. For example, sales is going to be worried about prospects, and management is going to be focused on customers. So while insights can shape and influence the plot turns of a data story, humans—not numbers—should be at the center of the story (especially if your audience are actually the characters of your story). Adding a human face to your data story will help it to resonate more strongly with your audience.

Fiction #5: Data stories start with a hypothesis or question. How you start a data story is critical to its success. Just because your analysis started with a hypothesis or question doesn’t mean your data story should begin that way too. Starting with a hypothesis or question places too much emphasis on the analysis journey (your approach to analyzing the data) rather than the insights you uncovered. Just like an author needs to establish the story’s setting (e.g., a galaxy far, far away), a data storyteller needs to provide ample background information to help orientate the audience. The beginning of your data story should focus on providing sufficient context (time period, data source, previous trends, etc.) so your audience can properly understand the insights you’re going to share. It’s also an opportunity to introduce the main character(s) that will bring the story to life for your audience.

Fiction #6: Data storytelling can be automated. I’ve seen various vendors tout the power of natural language generation (NLG) as a means of providing automated data storytelling. Rather than relying on “confusing” data visualizations, they feel descriptive text about the data is somehow better and easier for people to follow. In a world where humans have a lower attention span than goldfish, text is never going to beat images. Furthermore, the main advantage of having text is not to describe the data but to explain it. I should be able to see revenue increased 72% from Q3 to Q4 in a well-designed chart without descriptive text telling me the same thing. If that’s not the case, fix the chart instead of adding redundant text. However, knowing the 72% increase was attributed to the success of a new marketing campaign would be valuable context. Text can add value if it’s explanatory, not descriptive. Without adequate context and domain expertise, it’s going to be difficult for machines to “automate” data storytelling.

With so many important data stories to tell, it’s our duty and privilege to ensure the insights we discover are properly shared, understood, and acted on. Misconceptions about data storytelling can inadvertently weaken and interfere with the message you’re trying to convey through a data story. It’s often our fault—not the data’s—when an insight fails to resonate with an audience and ends up being ignored.

William Playfair, the 18th-century Scottish inventor of many modern-day charts noted, “No study is less alluring or more dry and tedious than statistics, unless the mind and imagination are set to work.” Data storytelling has the power to open people’s minds to new possibilities and ideas they would have never imagined.  As data is democratized throughout more and more organizations, it is no longer left to the back alleys and side streets of your IT department or the gated community of the executive boardroom. Data is increasingly parading down the main street of your business for all eyes to see, and how you choose to craft stories from this valuable resource will be essential to your success. Don’t let the misconceptions I’ve highlighted water down or ruin the power and effectiveness of your next data story.

Follow Brent Dykes on Twitter:

Follow @analyticshero