Storytelling with Data

Data provides us with powerful opportunities to tell a story. With the advances in data visualization tools over the recent years, the possibilities are truly endless. However, with the vast quantities of data at our immediate disposal, storytelling with data has become more difficult as authors try to do too much or lose focus on their original intent.

Over the years, I have created countless reports for dozens of clients. In some cases the data was limited and straight-forward, while other times it was large and extremely complex. In either case, it was my job to tell a story with the data at hand through education and engagement. What’s helped me tell these stories is following the guiding principles below:

  • Find the Right Data – Compelling data makes storytelling immensely easier and will engage the end user. However, this is usually easier said than done as many times the data set contains stale metrics like bounce rates, downtime, product costs, and transactions. Regardless of the domain or data set, it’s important to make the data fun and interesting. When presented with less than stimulating data, my goal is to bring in additional sources of data to augment it in a way that is both relevant and contextual. Blending data takes an old report and gives it new life.
  • Make a Hypothesis – Before even starting to build a visualization or report, start with a hypothesis. Starting with a hypothesis allows you ask a question and begin with a data dialog. Is productivity improving over time? Do younger consumers engage with the brand in the same way an older audience would? Does the day of the week impact click-thru rates? It also helps to analyze the data from a different perspective. Sometimes a hunch will pay off and other times the results will be counter intuitive. Regardless, it’s a good starting point. However, if the findings are not insightful it might be best to go back to the drawing board. If you don’t find your story compelling, neither will your audience.
  • Start With Simple – The worst thing is overwhelming the audience at first blush. The job of a data storyteller is to get them comfortable and ensure they understand what is being presented to them. If their first reaction is “Someone will have to walk me through this,” the storyteller is probably on the wrong track. In most cases, it’s best to start the story telling with the answers and let the audience work backwards. This is beneficial when there are different types of people viewing the reports and dashboards. Executives can instantly get the macro level information that is important to them, while those closer to the day-to-day operations can dive into the details if needed.
  • Put It In Context – Too often data is presented without a frame of reference. Without context, the audience may not understand the impact of the data, or worse, draw the wrong conclusions. Find points of comparison such as: variance to budget, increase from prior period, or indexing to standards. On April 3, 2015, The New York Times wrote an interactive article estimating how long baseball records would stand before each would be broken. The article was powerful because the viewer could see how today’s top players performed in relation to those who have held various records over the years.

Capture Given the current trajectory, the complete game record held by Jack Chesbro in 1904 may never be broken.

  • Use the Right Visualizations – There are several great books and blogs about using the proper visualizations for the right circumstances. We won’t go into the details here, but a critical takeaway that’s common in all these publications is to adhere to best practices. Not utilizing the most effective way to present your data will significantly limit the ability to tell a complete story.
  • Remember Your Audience – Whenever I begin to tell a story with data I first think about my audience, regardless if I’m building a dashboard, executive summary, operational report, or blog post. How technically savvy are they? Are they comfortable with statistics and complex metrics? Will they be satisfied simply viewing Key Performance Indicators (KPIs) or do they want to dive into the details? Data is often presented in the way the author wishes to view it, rather than the way the audience will understand it. Not being cognizant of your audience is quick way to lose the effectiveness of a report. By catering a product to your audience, it will allow them to focus on the story you wish to tell, rather than figuring the technical aspects of the report.
  • Let the User Create Their Own Story – Storytelling with data is all about engagement and there is no better way of accomplishing this than making it interactive. The story can be done through drill-downs, context filters, time-series analysis, animation, and segmentation. Present data that needs to be seen, but ensure the audience can test their own hypothesis and discover insights into the data that may have missed. Individuals approach data from their own point of view and letting them explore it on their own will be far more beneficial to the organization.
  • Continually Evolve – Very rarely will the story be right the first time. Ernest Hemingway is reported to have rewritten the ending of Farewell to Arms 39 times before he was satisfied. Building a great story with data is often a journey and requires several iterations and edits before it is complete. Start with prototypes and continually solicit feedback. With patience and the right audience focus, a great story will be told to an audience ready to understand and interpret it.

I hope these guiding principles help you as they have helped me through the years for a variety of clients. With the ever evolving data analytic capabilities, data storytelling will become more and more exciting, engaging, and critical for business decision making. What’s your story?

David Newman

About David Newman