SXSW 2014: Big Data, Personalized Stories

We hear often about the power of “big data” to do magic things like identify supply-chain problems, drive conversion, optimize direct mail and other impressive tasks. But what effect could big data have on storytelling? After all, isn’t that the other side of marketing besides the actual technologies used to deliver the stories? Today at the Sheraton, amidst an audience made up almost entirely of marketers and product managers (as determined by a show of hands), Francois Ajenstat from Tableau Software and Eric Shoup from held an interesting panel discussion about “Big Data” as it relates to storytelling.

Mr. Ajenstat stressed the need to try and break out of displaying data simply in rows and columns and show information more visually, as a way to help tell a story. Marketers today have access to mountains of data: purchase activity, social media engagement, device GPS data and even government data. Beyond simply using this data for business decision-making, could it be used to tell a story back t consumers? He described data as a new form of multimedia that brands could turn to as a way to enrich their own content. As an example of data telling a powerful story visually, he cited a TED talk by Hans Rosling that used data visualizations to challenge the audience’s perception of global health and trends.

Mr. Shoup showcased a new product recently launched by called Story View. The new tool draws upon Ancestry’s 10 petabytes of data (across 55 million family trees) and assembles what it knows about a person into a sort of timeline. It effectively pieces together a biography, with dates and places and images assembled to tell a person’s life story (as best they can) in a linear fashion. They used A/B testing to try different layouts and templates, and used the number of social shares as a metric to measure whether a given approach was working.

The speakers agreed on this tip for turning data into stories: Start with the question; what are you trying to get at? Then see how many data elements you can you remove and still answer the question. In short, Simplicity helps the storytelling.

Also discussed was technologies such as those offered by Narrative Science, which automatically generates prose summaries (e.g. of sporting events) based purely on raw data. A potential application of this technology could be to auto-generate direct mail or mass e-mail based on big data customer profile, changing the narrative text on the fly for each customer.