Best Practices for Data Storytelling
The secrets of data storytelling and how it can be leveraged to improve how you communicate key business information.
- By Michal Baumgartner
- August 25, 2022
Introduction
Telling stories is an essential element of the human experience. From the earliest days of the oral tradition to modern films, stories have been central to the way information is conveyed. Yet, in today’s business world, data storytelling has often taken a back seat to raw data. In PowerPoint presentations, whiteboard meetings, and elsewhere, the need for speed and efficiency often results in the mere delivery of facts, often incomplete and without context. This, in turn, leads to misunderstanding of data, inevitably resulting in poor decisions and faulty strategies.
Storytelling doesn’t have to come at the expense of productivity. It’s the key to a more effective business model. With modern data management and conditioning, creating stories that draw real meaning out of vast stores of data is proving to be a crucial component of digital transformation.
What is Data Storytelling?
Data storytelling is the art of visualizing insights from data using a narrative form. Traditionally, data visualization comes in the form of a dashboard or a chart, which provides the viewer with an easier way to digest information than digging through raw numbers on a page. The storytelling component is added when we provide context to data, giving a more complete picture and understanding. Context can be added by presenting data in an ordered fashion (chronological or otherwise), adding digestible explanations and viewpoints around what data shows to aid comprehension or suggesting steps to be taken to drive action.
Data storytelling does not have to be overly complex. It can be as simple as two charts side-by-side, or it can be a lengthy series of interactive graphs, connected text, varying versions of the same chart, comparative images, and a host of other elements. The goal remains the same, however: to guide people through a narrative to reach a more informed conclusion.
The key difference between data storytelling and data visualization is this added context. This context should provide the connecting tissue between each presentation -- whether pictorial, graphic, or alphanumeric -- which helps both the viewer and the presenter focus on the broader message being conveyed.
A good way to think of it is as a history lesson. If you ask students to simply memorize endless names of people and places, dates, and events, they will come away with only limited knowledge about the past. If you combine all of these things into an engaging and explained story, students gain a much deeper understanding of what happened, why it happened, and what lessons they can take into their lives.
In a business setting, we can see how visualization can convey a limited message based on relevant data: “All our offices met their goals this month; therefore, we can open a new office.” Using a storytelling approach, however, we can add necessary context, such as: “Every office reached its goal this month after we launched the new incentive program. These were our last expectations and these are the results. We can now open a new office and this is how much growth we anticipate once it is up and running.”
Sure, data storytelling is not suitable for every data presentation. For people who track particular data frequently or daily, an operative dashboard would still work the best.
Maximizing Storytelling’s Benefits
Storytelling has a critical advantage over visualization: without context, information becomes hard to digest, particularly among numerous and diverse audiences. Even in a visual form, data is still data, and can be understood differently by different people. The best use of data storytelling is for audiences that may not be familiar with the presented data or business context. Storytelling is also useful in the current environment of home offices and asynchronous meetings.
Most business intelligence and data visualization platforms are very good at data visualization, but they lack the capacity to place all of the data they present into a cohesive model of understanding. An important aspect is that usually reporting can’t be disseminated across an entire organization. Each business unit tends to pull what it needs in order to satisfy its own mandate without giving much thought to how their use of data affects other units or whether it is helping guide the organization as a whole to a successful outcome.
With a narrative-and-explanation approach, data is placed into a broader and more accurate context in which all viewers can see the beginning, middle, and end of the function that the data supports. Also, additional information about reasoning or next steps creates a greater level of understanding. Knowledge workers are then able to make not just better-informed decisions but decisions that help propel the story -- the whole story -- forward.
Tips for Telling Your Story
The most important aspect is that you actually understand your data and its business context -- you need to have command of the data you wish to present. If sales are reported at, say, 100,000, is this an increase or a decrease? How significant is this? Is this monthly, quarterly, or yearly? What currency does this represent? This can greatly affect the outcome of the story, which, if misinterpreted, can vastly alter strategic decision-making.
The second most important aspect of storytelling is to understand your audience and, in particular, identify gaps in their knowledge that need to be filled. From there, you can begin compiling the data needed to fulfill their needs -- not simply reporting that there was an increase in sales last quarter but why there was an increase in sales using diverse data sets such as sales reports, product launches, and performance metrics -- and what next steps should be taken. Again, though, the aim is not to simply report these numbers but to use them to tell the tale of how sales got from there to here.
Keep in mind that understanding the audience is a multifaceted project in itself. Not only must you ask about the composition of this audience but also about their specific relationships to the task at hand. What information do they already possess and how did they react to previous presentations? Data stories can (and should) contain explanations -- or even better, next steps and recommendations. In today’s business environment, there is a vast amount of data making it harder for business people to make decisions alone. Data storytelling can help here as business analysts (or AI) can prepare data stories with suggested action points and next steps enabling easier and faster decision-making.
Finally, don't forget that telling a story requires equal participation from both the presenter and the receiver, so the more you know about your audience, the more likely they are to be able to digest and understand the story. You will need to have command of the data you wish to present. It must be in the right structure and format, and it must be presented in a recognizable pattern that provides a complete view of the situation. All this data must be made relevant to business goals and objectives.
The Problem with Pie Charts
Pie charts have largely been banished from the data analytics and visual analysis communities for a number of reasons. First, they require too much space to convey a limited amount of information. Even in a digital slide presentation, real estate is limited, and pie charts must be large to provide meaningful communication to the user. Studies show that pie charts are inherently less readable than other charts. The human eye is much better at discerning minute differences in distance than area, which is why it is much easier to compare data sets on a bar chart or a graph versus a pie chart. These deficiencies become magnified when the data becomes more complex and the pie segments become smaller.
Final Thoughts
Data conditioning is a vital asset in your data storytelling quest because it helps determine the proper accuracy and validity in order to deliver a convincing narrative. This is crucial in order to avoid two key mistakes in data storytelling: overloading the audience with too much (or superfluous) data, and presenting data in the wrong order.
Also, note that simple profiling is not enough in most cases because it fails to account for how data is created, utilized, and stored. Proper preparation can overcome many of these issues by ensuring that data is placed in the correct standardized format and then further parsed and enriched to provide no gaps or duplicate sets that can lead to false conclusions.
By building a foundation of trust in the data, you can begin building your story using combinations of text, numbers, visuals, and other elements, all exquisitely layered and presented in the right order to ensure your audience receives the right message in the right context so they can make the right choices going forward.