Harnessing the Power of Data Storytelling
It’s easy to collate some data, maybe even make a graph from it. But can you convince anyone with it?
This is a question I pose to my (mostly actuarial) students right at the start of my Data Visualisation and Communication course at UNSW. It doesn’t help that many actuarial students (and perhaps many practitioners) lean towards the introverted stereotype. Observationally, the phrase data communication is often (mis)interpreted as “here’s a list of numbers I will now read out to you”.
There is a wide chasm between being able to analyse data for your own understanding and using these data insights to persuade decision-making.
The key tool I teach for bridging this great divide is data storytelling. In this article, I’ll share some of my experiences and insights in this arena and provide some tips for harnessing the power of data storytelling in your own practice.
Why do we need storytelling?
Decision-making is an emotional process — see the vast work of neuroscientist Antonio Damasio. People tend to be more receptive to a decision, especially a bad one, if they understand why it was made. Unfortunately, some of my students seemed upset that data communication could not be solved through R (or a comparable tool).
One student’s feedback even questioned how one of my activities which involved role-playing negotiating a lower rent from a landlord (using some hypothetical summary statistics) was relevant to data communication.
Data communication in business applications is rarely an exercise of sheer curiosity; to play with the data for the sake of playing. Instead, it is almost always used to drive some decision or action.
To effectively influence decisions with data, we need to trigger an emotional response in our audience. Data storytelling is a great tool for achieving this and simultaneously revealing your why.
OK, so what is (data) “storytelling”?
Forget about data for a moment and just think about stories. Think of a great movie that you’ve seen or a brilliant novel you’ve read. More than a mere list of sequential facts, these stories featured an overarching purpose and a clear direction that connected the facts for the intended audience.
This is the interpretation of storytelling I use for data as well. The most insightful analysis in the world, without this sense of overarching purpose and story, would simply be a list of facts. Data storytelling is about taking the analysis and connecting this seemingly disjointed list of facts into a path that leads your audience to a particular conclusion.
I suspect that, in the case of many students, they are hoping that showing more analysis being done will translate to more marks. Unfortunately, this approach is unlikely to influence any actual decisions (at least, not in their favour).
The power of purpose
I frequently warn my students that, by the end of the course, they’re going to be sick of hearing me say the word “purpose”.
Having a clear purpose not only helps you stay on message, but gives your audience a sense of transparency, leading to trust. To highlight the converse, imagine the cliché of your romantic partner giving you a vague “we need to talk”. The lack of transparency here will likely trigger a fight, flight, or freeze response; not conducive to getting someone on your side.
To borrow another cliché, having a purpose is more about the journey than the destination. There needs to be clear linkage between the facts (data insights) you present, the interpretations you draw from these insights, and the action or decision you’re asking of your audience. In other words, you need to frame the results of your analysis as the why of the action/decision you wish to drive.
A framework I give to my students for building this linkage is a backward-designed plot for a data story:
- What is the decision or action you want from your audience?
- Why should they do or choose that?
- What’s in it for them? What are the consequences of not choosing or doing this?
- What are the underlying problems or issues?
- What does the data say?
When you read the list backwards, it forms a sequence of a logical flow. It also forms a hanging vine (inverted tree?) of ideas, since there could be multiple pieces of data that can flow into the final decision or action.
I also tell students that this framework will help them filter out irrelevant insights. If an insight either doesn’t lead to the final decision/action at all, or requires too many steps to get there, then cut it.
Hooking your audience
How long does it take you to form your first impression of a person? A cursory look at the available psychology research suggests it takes somewhere between a fraction of a second to around 7 seconds. Whilst there are other variables to consider in the context of communicating data (e.g., what is your relationship with your audience, how far you are into the project), the idea of making a good first impression also applies to your data story.
Re-read the first sentence of each section (including this one). They were all chosen to trigger a sort of curiosity reflex in the audience. Your audience needs to be hooked into your story right from that very first sentence. A great hook gives a clear sense of where you are going, and simultaneously makes your audience want to know more.
Think about the opening scene of Star Wars – A New Hope. Right at the start of the film, you know that Darth Vader is the bad guy, that the good guys are the underdogs, and that they need to rescue Princess Leia and stop Vader. This clear sense of purpose and a great hook then motivates the rest of the film.
By the same token, a data story’s hook needs to also provide this sort of context. I teach my students a few basic formats they can use for writing their hook:
- give a surprising statistic;
- challenge their existing beliefs; or
- ask a confronting question.
To be clear, these are not mutually exclusive; a statistic can be surprising precisely because it challenges existing beliefs. I give my students a fourth option, build an emotional connection, which I deliberately omit from the list above as it may not apply as directly in an actuary’s typical work.
Less is more
Going back to A New Hope, do you remember the scene where Luke does a load of dirty laundry? Or how about the scene where Leia goes to the hairdresser to get her iconic hairstyle done? No? A good story contains only what is strictly necessary for the audience to see. It deliberately excludes the excruciating and unnecessary details in the background.
We see this idea in the “data-to-ink principle” for a single data vis, but it also applies to the broader data story. I implement a 3-minute time limit for the final video presentations of my students.
A typical response is along the lines of “but I did so much research, I need more time to present all my findings”, however a lot of preparatory work goes into whittling down to only the essential information.
One famous non-data example is Abraham Lincoln’s famous Gettysburg Address (“Four score and seven years ago…”), which lasted around 2 minutes. What I only learned in recent years was that preceding Lincoln was a man named Edward Everett, whose speech lasted 2 hours! It’s clear which speech had a greater impact.
Remember, your goal lies not in how much you say, but how much your audience remembers.
Data storytelling and actuaries
Imagine you went to your General Practitioner (GP) for a check-up, and your GP simply reads out every result of the medical tests that were performed. Very likely, you would not know how to interpret these results yourself. You would expect your GP to flag and explain any problems, and give recommendations to improve your health.
The same idea applies for actuaries. Although quantitative analysis forms part of an actuary’s skillset, it is not the main value they bring. Like with the GP, an actuary’s main value is in understanding the stakeholder’s needs, providing the professional judgement to flag and explain key insights, and giving meaningful recommendations and advice based on these insights. Taking this perspective, data storytelling is just a convenient tool for an actuary to provide their full value.
CPD: Actuaries Institute Members can claim two CPD points for every hour of reading articles on Actuaries Digital.
