Storytelling is the Way to Leverage Insights

I recently gave this talk at the Open Data Summit and at the Open Data Institute in London. I have added the slides and a transcript below the video. 

With the rise of analytics, open data, or simply big data, we’re literally drowning in information. The ability to extract actionable insights from it (and by actionable insights I mean some sort of understanding leading to an efficient business decision) requires particular skills. Analytic skills.

 

We haven’t yet reached a stage where everyone has developed this kind of knowledge. At the moment, the world seems divided between the ones who manage the data (people like data analysts, data scientists, data engineers…) And everyone else!

 

But what happens when the data specialist is not the decision maker? When the decision maker is not data literate, or even simply not as literate as a data specialist?

 

From my experience working closely with analysts I can tell you what generally happens:

Data people get excited and they provide loads of information - so much so that decision makers don’t even know where to start with it. And forcing an audience to make sense of your content is never a good idea. Especially now that studies have shown that the average attention span officially stretches to 6 seconds – as short as that of a goldfish!

 

In general 2 things happen when decision makers are not engaging with the analysis:

  1. Data people have just spent an awful lot of time working on a piece that will not be used, in a word, they’ve lost their time

  2. Therefore, business decisions are not as efficient as they could be as they don’t rely on data.  

So, what can we do to improve data specialists communication and decision makers' understanding?

We use storytelling to add meaning to data. Everyone likes a good story. It appeals to people. It bridges the gap between raw numbers and meaningful content. It encourages a reluctant audience to engage, and a wilful one to remember.

 

Now the question is: how do we create good storytelling? And my answer is that there is not one path, but many.

 

Which brings me to my position as an information designer at Lloyds Banking Group.  My daily task is to go through reports, find insights and then link them together in a way that will appeal to a defined audience. Unfortunately, due to their commercial sensitivity I cannot show you any of these specific creations otherwise my boss might have a heart attack. 

However I can still show you some stuff:

I created the above infographic (or slice of infographic as it was too long to fit here) for my colleagues.

 

As a bank, we’re really concerned by online criminality; for both our clients who see their card fraudulently used, but also for employees, who could turn the bank upside down by clicking on the wrong link.

 

The problem is that we’ve all seen this documentation a thousand times. Anti-phishing campaigns have been going on for a while, so long that we don’t even read the documentation when it comes our way. My goal here was to raise concern in a way that wasn't patronizing.

 

Let’s have a look in detail.

 

Informative title, we know what we’ll find on that document. As an audience I can decide straight away if it’s worth my time.

 

The hook: the incredibly high 27 billion figure. Quite surprising, I want to know how come.

 

Plus the spread of the problem – you could be (even if you have never been) one of the victims. Here I connect with my audience.

 

And another worrying element, that I could potentially repeat at the dinner table. Hey darling, do you know that today we’re 20 times more likely to be robbed online than on the street?

The piece above is an interactive dashboard showing the diverse ways in which customers interact with the bank.

 

Always focus on your audience – here it’s a dashboard aimed at executives – that’s why I’ve considered the time they could spend on it at the top e.g. 3 mins reading / 6 or 9.

 

The order in which the information aligns is crucial. Put the interesting content at the top – consider that your audience may drop before the end. 

 

See the little hand? It’s a hyperlink to the sources. Always make it easy to find the dataset so that people can dig into it on their own (if indeed they want to) – make it accessible. 

The chart above is one that I have produced for Arcadis which displays 100 cities ranked by their level of sustainability.

 

What is really important here is the level of detail. Do I need to compare each of the 100 cities to have an idea of global sustainability? No, I am selecting some of them in order to give the big picture without tiring my audience with too much detail.

I’ve designed this image in my spare time. It’s part of an article titled “Hilarious” situations all Londoners know about. I selected it for the  "Open Data Summit"  presentation because I wanted to show you that sometimes, it doesn’t take much to produce storytelling.

 

With a single image and 2 sets of digits, you intuitively got a picture of London, the tube, the overpopulation, the emotion and maybe even a personal connection as you may have lived that situation for yourself.

 

 

As I’ve said, there is not one but many ways to create an engaging narrative and the good news is that we’re all naturally geared up to produce stories. Not convinced? Let’s try and see.

Let’s consider the text.

 

Why did Sharon weep? Why did she go hungry?

 

You may have deduced that Sharon wept because Andy died. Or even that Fred went to a grocery store and that’s why Sharon went hungry.

 

Read again. Are those three sentences really linked together in this way? Couldn’t “He” be John or Arthur instead of Andy? Is Sharon really weeping because of Andy’s death or could she be sad for another reason?

 

The brain, by default, will create a story from a series of events even if they are unrelated.

 

Scientists are particularly familiar with the dangers of this cognitive bias and seek to negate its influence by applying the mantra “Correlation does not imply causation”.  

 

Follow these simple steps to help bring your data to life...

 

As I’ve said and hopefully shown, there are many ways to infuse storytelling in your data. There’s no perfect recipe, but personally I always keep in mind a series of points to help me expand the narrative:

In order to avoid making a mess of the information define what your key message is upfront. Think of it this way: if your reader had to remember only one thing, what would it be?

 

Align your narrative around this central fact. What is not backing up your key point may lose your audience. It becomes a distraction to your argument.

Defining who you are producing your content for is as important as defining your key message.

 

Ask yourself: is your content answering or helping to answer the question your audience has in mind?

If not clearly defined upfront you may deliver information they don’t need and therefore lose your time or theirs.

You may be familiar with the jargon or processes but your audience may not be.

 

Ask yourself: how knowledgeable is the audience?

When reaching for a wide audience I use the pub test: I imagine that I am telling the story to a friend in a pub. Could he understand clearly the narrative if I present the story that way? Hopefully it should reduce the odds of being misunderstood.

Too often people totally neglect this stage.

Craft your title carefully as it’s the first thing your audience will read, it will tell them in one second if your content has to be read or if it can be ignored.

For inspiration consider newspaper headlines. I'm loathe to admit it but the Mail Online is very good at this game.  Perhaps that’s why they’re the most visited online news outlet in the world.

 

Well, this concludes my presentation for today.  I hope you have found it useful and please do stay in touch if you have any comments.