Storytelling has become central to discussions on marketing in 2016. We’ve even contributed our own thoughts on the power of verbal and written communication to share ideas, persuade and incite action in ways that visualizations or numbers alone cannot.
This concept now has a firm set of followers/believers. Some, such as Narrative Science, have taken huge leaps in using Natural Language Generation combined with analysis to describe performance and its key drivers in written language that reads as if it were produced by a human. Adam Kanouse, CTO of Narrative Science, shared an extremely fascinating example of how their software can ingest data, perform analysis of it, isolate key facts and run these through a Natural Language Generation system recently at a Data Driven NYC event.
But… is merely describing the information shared in visualizations or drawing conclusions from metric performance or data sets across time sufficiently valuable as to constitute storytelling?
Simply put, no.
On one hand, maybe Storytelling is an accurate word. One of the key principles of a story is that it has a beginning, middle and end. Describing past performance is similarly an action that stops at a finite location in time (any moment prior to now). On the other hand, however, the value of stories doesn’t end when the story does. Stories are written to; challenge us, make us consider our position on issues, make us rethink our morals, to teach us lasting lessons about the world, etc. Have you ever finished a story and made a change in your life? Or seen a shift in your perspective? Or simply thought about a new concept?
That’s what storytelling in reporting is attempting to achieve.
So then, would the following invoke an emotional response and cause you to make a change?
“Revenue has decreased across two consecutive quarters. In terms of a percentage, the decrease was largest during the last month of each quarter.”
How about this?
“20% of our conversions were converted at a cost per conversion of $70 and have an expected average lifetime value of $5,000, while 20% of conversions were converted at a cost per conversion of $390 and have an expected average lifetime value of $250. Let’s focus on the first segment by targeting women aged 18-34 who live in a specific area and also purchase from brand X. Specific campaigns built around ‘this theme’ and run through ‘this channel’ have historically led to an increase in the conversion rate of this group by ‘X%’.”
Much more likely, right? So why?
Not only are we describing what has happened here, but we are providing a recommendation for an action: to focus on a specific segment, with a specific campaign type, and a specific channel. Perhaps even more powerful than this is the forecast for the expected impact that such an action will have: conversion rate increase of X%.
While we aren’t tugging on any emotional heartstrings with this type of storytelling, we are inciting action. By including an outline of the expected impact of carrying out the action, we are making the potential consequence of not doing so abundantly clear.
And in a world where we are so often held accountable for the influence we have at work, I wouldn’t want to have to tell my boss that I had missed the chance to increase conversions by X% because I didn’t follow a recommendation shared with me by a colleague who had ample access to historical data and a strong ability to interpret it to make recommendations. I’m not sure you would either!
So while we are extremely excited by the potential of such language processing solutions combined with analytical solutions, in the short term we will continue to promote the value that an experienced analyst with the ability to interpret data, and also apply common sense and contextual knowledge has.
We will also continue to encourage storytelling that not only describes historical performance, but that also communicates recommendations and their potential impact.
Not Another Dashboard.