Data storytelling: how to blend narrative and data visualisation to improve business communication

‘The Power of Data Storytelling’ by Sejal Vora shows how the art of storytelling can be effectively used with the science of data so that organisations can make better decisions.

In a world where decisions are becoming increasingly data-driven, it is stories that bring data to life, grab user attention, increase audience buy-in, improve retention, build emotional connections, and trigger desired actions.

Frameworks and case studies of data storytelling are explained in the well-illustrated book, The Power of Data Storytelling, by Sejal Vora. Storytelling is hardwired in the human brain, and comes to us unconsciously and consciously, the author explains.

Sejal Vora has been an investment research analyst for over a decade, and is now a corporate trainer. She graduated from NMIMS, and worked at Morgan Stanley and RBS. She is a CFA and MS Finance (ICFAI).

The book spans 220 pages, and is well-illustrated with charts and figures. More diverse case studies would have helped, as well as a review of related publications and online resources. Here are my three sets of takeaways from the book. See also my reviews of the books Stories for Work and Analytics and AI, and YourStory’s Changemaker Story Canvas and Pitch Tips for Startups.

In data storytelling, the characters are not raw data points but measures (eg. absolute and proportional growth rates, composition of market share), and events are the trends or relationships between these characters (eg. growing gaps, upward rise, dominance of one factor).

I. Story frameworks and structure

Sejal advises data storytellers to first use frameworks like story wheels, structures, and arcs while developing their stories. She defines a storywheel as consisting of data, events, characters, assessment, outcomes, purpose, and the ‘aha’ moment of revelation.

These are the six elements of the data story, and bring out the where, when, why, how, what, who, what next, and so what aspects of decisions. This will help determine what is important in the data insights, why it is important, what the implications are, and who will take the desired actions.

The author classified business stories into four types: reporting (eg. business performance, market updates, attrition trends, monthly costs), decision (eg. recommendation, approval, transaction), probing (eg. exploring competition, macro-trends, regional comparisons), and pitching stories (eg. startup pitch to an investor, attracting a new client).

Stories have a context, sequence of actions, climax, and conclusion. Sejal draws on Kurt Vonnegut’s work, The Shapes of Stories, to focus on five types of story arcs: upward sloping arc, downward sloping arc, sharp arc, growth arc, and hills-and-valleys.

The downward sloping arc can be used, for example, to show how discounting strategies over a period of time may improve market share but not profitability. An upward sloping arc can be used to make an investment recommendation despite a small risk.

The context or summary (eg. in the opening slide) introduces the topic, provides background information, and gives a hint of the conclusion. This helps set the tone of the pitch and expectations of the audience, according to Sejal. The flow should link and connect the events as the story progresses, in order to keep the audience engaged.

The climax should bring out an element of surprise, curiosity, and even excitement. Risks and challenges should also be flagged off, but without distracting the audience. The conclusion should bring closure to the story, and justify the decisions to be made.

In addition to bottom-up frameworks, Sejal recommends a top-down blueprint framework called the story map. It breaks down the story into three or four key themes. Each theme consists of a few events which generate insights, and each event has underlying data descriptions.

For example, an investment recommendation story consists of company background (operational statistics), investment decision drivers (market leadership, future potential), and valuation (equity, peers, exit).

It is also important to understand the awareness, preferences, and needs of the audience, and the intended impact of the story. This will impact the depth, style, and presentation of the story, eg. detailed written report with colourful charts.

II. Stories and data

Data visualisation and storytelling can make business communication more effective, productive and actionable, Sejal emphasises. Data storytellers have an important role in ensuring that insights are revealed from data and brought to the attention of decision-makers.

Data by itself can be boring or complex to some professionals. Data storytelling highlights key data points, relationships, trends, and effects. It helps data analysis deliver breadth, meaning, and significance to audiences, Sejal explains. Data storytelling is more than data visualisation or savviness used in the latest tools. It blends narrative, oral, and written presentation skills as well.

Unfortunately, many charts and presentations tend to have too much data ('data dumps'), bad visual aesthetics, and lack of insights, Sejal cautions. Many professionals also have misconceptions that stories are fictional, flowery, unnecessarily emotional, and not serious enough for decision-makers.

To be effective, the data story should have flow and rhythm, along with emotional elements such as surprise, empathy, caution, disappointment, satisfaction, and pride. These should be brought out be framing the actions, trends, risks, and recommendations.

Good stories have episodes, scenes, and dialogues. In the context of data stories, episodes can be clusters of entities (eg. developed and emerging economies), scenes are activities (eg. absolute and relative business performances), and dialogues are the data description.

III. Textual and visual elements

Sejal provides a wide range of presentation tips for data storytelling. For example, the choice of text in the chart titles play an important role in this regard, and shows the narrator has gone beyond bland data description to deeper insights in the form of patterns.

More text should focus on key themes, and important points should begin on new paragraphs. Supporting information should be moved to an appendix or provided in footnotes. Common mistakes like overly long sentences, wrong units, missing sources, and inappropriate font styles should be avoided.

Practice and revision makes it perfect, Sejal urges. At least three rounds of edits will be needed to identify structure, analysis, and mistakes. “In data analytics, the why becomes more important than the what,” she emphasises.

One whole chapter is devoted to visual elements of data stories and when to choose the appropriate ones, such as stacked bar charts, bubble charts, line charts with markers; axis labels, smoothening of sharp lines, and the use of colour. Charts can be augmented with images and even photographs.

But too much colour and 3D charts should be avoided, the author cautions. The choice of adjacent colours should be smooth in transition and not jarring. There should be consistency between the colours of a line and its label.

Sorting of bars; horizontal versus vertical displays of bars; adjusting the width of bars and the spaces between them; and darkening or drawing a dotted box around a significant bar, are other presentation decisions to make.

Sejal also shows other examples of sorting plots into quadrants, demarcating bars into clusters, using arrows and labels to highlight trends, adding a few grid lines, thickening key lines, including ‘call-out’ text, and adding new baselines (eg. five-year profit level, ten-year profit level).

All these principles are effectively illustrated in the last chapter, with two hypothetical case studies. For example, the story map of a company’s performance is broken down into volume, revenue, and profit, drawing on historical and current growth in domestic and export markets. Data is presented in well-designed combo charts, and the story arc reveals good growth (rising action) despite some mitigants (falling action).

The story map of a hotel chain’s performance is divided into demand-supply dynamics, revenues, and future demand impact. The headline of the story captures the effect of macro-tailwinds in the industry, and the final slides show line charts with smoothed lines. The title of the concluding slide answers the ‘what next’ aspect of the story (good growth prospects).

The author ends by urging companies to create a culture that blends a data focus with storytelling practices. Support, constructive feedback, and coaching are needed to ensure that such a culture sustains.

(Edited by Megha Reddy)