"Every picture tells a story, don't it?" If you think it's scary that I started this article with a quote from a Rod Stewart song, I did have a reason for doing so. Data storytelling is the big message being touted by some key BI and data visualization vendors these days. But much like that song quote, which dates back to 1971, telling stories with data is nothing new.
I'm a marketing guy, and I like strong marketing messages. And yes, data storytelling does matter. But what BI and visualization software vendors are enabling now is just an evolutionary step in how we tell data stories in organizations -- and it isn't just a matter of creating data visualizations.
In reality, businesses have used data to tell stories for much longer than there have been computers. A profit-and-loss statement tells a story. An income sheet tells a story. A record of the last month's manufacturing output tells a story. The advent of computing helped us analyze more data, faster, and tell more accurate stories. But the fact is that we were telling them when I was a COBOL financial software programmer on mainframes -- in the mid-1980s.
In the 1990s, Microsoft PowerPoint took over business storytelling, often to the detriment of the storytelling process. It's not that PowerPoint was bad in and of itself -- it's that far too many people spent far too much time designing slides with fancy fade-ins and other transitions and animations instead of focusing on the information they were trying to convey.
At the same time, the growth of computing power and data volumes highlighted a big limitation of the storytelling techniques then available to analysts: The static nature of reports essentially froze the data that they presented to business executives.
Storytelling meets data analytics
Fortunately, BI technology began to move past that. The 1990s also saw the growth of vendors such as Business Objects, Cognos and MicroStrategy, which worked to provide lower-latency analytics capabilities to business analysts and BI developers. Initially, the primary benefit was a way to get more current information into still-static reports and presentations.
As processing performance continued to improve, more became possible, including ad hoc querying of data. That was a big step forward for business operations -- and data-driven storytelling. Take managing a sales force, as one example. Before the emergence of BI software, printed reports or spreadsheets were used to provide sales executives with information on how regional teams and individual sales reps were performing. The ability to use an analytics tool to explore and query sales data provided faster, and better, insight into performance.
Since the mid-2000s, we've seen newer BI vendors such as Tableau and Qlik, as well as the established players in the market, move to a much more dynamic method of data analysis and storytelling based on self-service BI and data visualization software. The result wasn't the invention of data storytelling -- it was a change in the types of data stories that can be told.
Think of the older styles of telling stories with data as similar to writing novels. They required a lot of advanced thinking and the setting of a plot; then they needed to be written and produced. Everyone who read one of the resulting reports, whether a day or a year after its creation, read the same story.
The new storytelling capabilities are more similar to extemporaneous speech or even improv. They've narrowed the time frame for data analysis and reporting to near real time or even actual real time, allowing the presenter and the audience to interact with one another and adjust a data story as it unfolds.
Risks, not only rewards, in data stories
While the increased flexibility has great advantages, it also includes some risks. To get an idea of them, let's go back a lot further in time. In the preliterate world, storytelling was critical for passing on knowledge to people. However, there were two issues: Individual storytellers would learn slightly different versions of stories, and even the same person could tell a story in different ways each time.
Such variations could cause some big problems in the business world. When a sales director discusses what's happening in a certain territory, you want the region's manager to understand exactly what is meant. Key performance indicators used to measure sales performance must be the same or translatable among each geographic area so data analysis makes sense from a corporate perspective.
The immediacy of self-service analytics and data visualization adds another dimension of difficulty to that process. An oral story might change over time, yet the parable remains the same. What about comparing sales data before and after a change in the delineation of sales territories? Financial metrics aren't parables and need consistency.
As a result, the freedom of modern data visualization and storytelling can't be absolute. That's why metadata and master data management are absolutely required. They provide the consistent data syntax and semantics necessary to tell data stories that can be understood across geographies and business units, without creating misunderstandings and errors in decision making.
Just as the early growth of the PC both empowered end users and created some confusion in organizations, the new data analysis, visualization and storytelling methods have benefits and downsides. To be fully successful, they require a common language and shared context. We've been telling stories with data in organizations for a long time. Let's make sure that the new ways of doing so really do add to the business picture.
About the author:
David A. Teich is principal consultant at Teich Communications, a technology consulting and marketing services company. Email him at [email protected].
The role of data stories in motivating decision makers
Why the data storytelling trend matters to the business
Filling the role of data storyteller
To tell a compelling data story, use effective data visualization