Creating a Story with Big Data
EJ HowardNovember 13, 2014
Big data. It can deliver actionable insights that accelerate efficiency and results. The catch? Big data doesn’t come with a pre-built narrative. It arrives as a jumble of triggers and attributes. The data itself needs to tell you how those different, seemingly unrelated points are connected. To do that, you need to work with skilled storytellers.
Now, this doesn’t mean creating stories from whole cloth. Quite the opposite – these stories are grounded in observations and hypotheses, told by proficient analysts and strategists. But they are stories nonetheless with the same essentials of any good tale:
- Setting – What is the campaign? What does the competitive landscape look like? When does the story take place and in what environment?
- Plot – Like a good whodunit, big data tells a story in which one clue leads to the next. This iterative process builds until it reaches a crescendo: the insights.
- Action – In a bad story, actions only happen to the characters; in a good story, the actors actively make choices that drive the plot forward. Selecting audiences, establishing test and control groups, and piloting new media partners are the actions that drive big data storytelling.
- Characters – The dramatis personae of the story? The agency and client. Some characters are peripheral, others are central. But a good story recognizes the different motivations each character has and how those motivations inform their actions.
- Conflict – The heart of every good story is what must be overcome. It could be a lack of understanding or different viewpoints between characters. The conflict defines the characters, shapes them, and changes them.
- Dialogue – In storytelling, dialogue is the window to a character’s soul – it drives the story. With big data storytelling, dialogue establishes the hopes and fears of clients, illuminates the approaches and expertise of the agency, and drives the team towards collective understanding.
Here’s an example of how this storytelling plays out for a healthcare client. The setting was the first significant quarter’s worth of advertising for a portfolio of products. The primary action? Establishing a new product in that portfolio and understanding the interrelation of the audiences of the portfolio. The clients had a mix of perspectives: how did my brand fare versus the others? Are my dollars being spent well? How to best drive prospects to their doctor for more information? A steady dialogue prioritized the needs and interests of the characters while ensuring their different motivations were known and established.
As for the plot, the team exposed a trail of clues that led to actionable insights. The homepage activity on one brand was unfocused, with more on-target activity for deeper-site activity like “find a doctor.” For a more established product in that same portfolio, the general site traffic was on-target while the deeper site activity skewed disproportionately towards users whose demographics indicated they’d be unable to take advantage of the product. This led to a wealth of media tactics that built upon the findings and prepped for a continuing story of efficacy.
The example above highlights a few parting thoughts about storytelling with big data:
First, that storytelling is the art of exploration, requiring a strict methodology to identify which threads are worth following, and which trails will prove less fruitful.
Second, storytelling is about “pleasing one writer” – the author. Had the analysts ignored data on how audience compositions differed wildly on one product in the attempt to please one of the clients, they would have missed the bigger narrative. Third, storytelling is a continuum, one story’s end is the next one’s beginning.
Last, and perhaps most importantly, storytelling is a skill that must be regularly honed. Only through the rigor of constantly plotting and presenting can our storytellers identify and tell the stories that our clients want – and need – to hear.