Data analysis and exploration can answer many of the questions that decision makers have, and even generate questions that some businesses have not yet even figured out. Of course it also sparks curiosity around the value that data brings. Data has purpose and meaning – but not all organisations understand this – which is what makes data science so necessary. Data scientists have become the narrators of the stories the data is trying to tell to the business. In this piece, I want to focus on seven ways to improve data storytelling.
Telling stories has always been a wonderful human experience. It all started with cave drawings, the most primitive form of storytelling that already used visual imagery. Today, we are spoiled for choice – from reading to watching, with comics being the most humorous and then movies, television and online video clips being the more modern way of telling stories. However, in order to tell stories with data, we need to look at the essence of storytelling and adapt that to help our audiences make sense of their data. Below are seven ways to improve your data storytelling by adding context, improving engagement and adding emotions.
Focus on the audience
The first area to focus on is to make a strong connection with the intended audience. You need to understand the audience, where they are coming from, their background, their degree of interest and the level of detail and the type of “language” they understand. You need to make sure that you can present the story to be appropriate from the listener’s point of view. If the listener can identify where they fit in the story, it becomes a conversation and a real experience.
In the TDWI report “Ten Mistakes to Avoid In Data Storytelling”, Dave Wells emphasises that every audience is different. You need to understand the characteristics, interests and needs of the individuals that make up a small audience or the segments that make up a big audience. You need to make sure you identify with their stake in the story, especially when you want them to respond in a particular way. There are various aspects to consider, such as roles, level of authority, background knowledge, domain knowledge and experience. You may even have to consider demographics like age, gender, language and cultural perspective.
In some cases it may be best to partition the audience in appropriate groups and tailor the narrative and its accompanying visuals specifically for each group, in order to get a suitable message across and enable an appropriate conversation. For example, what and how you report on a project’s completion will differ whether you are sharing with an operational staff or managers, executive decision-makers or whether you are asking for more investment to continue – while the underlying work was all the same.
Structure the story
When presenting data – whether to an existing or a new audience – you need to have the whole story planned out, much like an author plans a book or a short story. You need to have the full outline worked out – which includes a beginning, middle and conclusion. This is what will allow your audience to actually follow the ‘story’.
Instead of just blasting them with millions of data points, you need to establish rationale and context first, then build trust around where the data comes from and what it represents and then you need to build up an understanding of how the results were achieved from the data. At the conclusion of your data story, you should have guided your audience to a decision, a solution or most preferable, an action that must be taken.
You need to organise the story to flow logically and be easy to follow. Instead of thinking of the data first, you need to focus on the narrative first and then bring in the right data to illustrate and support the narrative. It is the plot and the narrative that gives the story purpose and brings it to life. The data is simply the supporting evidence.
Make the story memorable
Jennifer Aaker, Professor of Marketing at Stanford Graduate School of Business, stated that ‘Stories are remembered up to 22 times more than facts alone’, with which I completely agree. Unless you are a total data nerd, columns and columns of numbers in an excel spreadsheet can hardly be interesting – in fact, in some cases they can just be plain confusing.
When we are able to turn the data into a story, this not only makes the numbers and facts digestible and understandable, but it gives them context, and makes what they stand for far more memorable. Remember, at the end of the day, you do not even necessarily need the audience to remember the actual numbers, but they need to remember the story. It is the story that needs to prompt them into making the right decision or to take the appropriate action. That’s why the story has to be memorable.
Use the correct language
Statistics form the basis of most visualisations and data analyses, but it is often not the most effective way to communicate the findings and conclusions of your analytics. Even when statistics are presented visually as charts and graphs, these represent isolated views and do not necessarily demonstrate cause and effect. So, the story needs to speak louder than the statistics. The language of the particular business makes the story accessible, credible and it improves the connection with the audience. Note, I’m not saying you should ignore or discount the statistics. You should analyse statistically but tell the story using a language that is familiar to and comfortable for the audience.
You also need to adapt terms particular to the industry at hand. For example, the concept may be similar, but in telecommunications it’s called churn, in insurance it’s called a policy lapse and in education it’s called attrition. I know this can be challenging for data scientists, and that’s where the requirement for business knowledge and acumen becomes a necessity. Given that your audience may not understand how their data might answer their business questions, you need to turn those figures, data points, trends and graphs into a story that will make sense to the business – in business terms – in a format that they will remember when they walk out of the presentation.
Make the story real
There are two ways to make a story “real”, and that is to put your emotions into the story and to use characters.
A boring story has no impact, unless you’re trying to put your children to sleep – but that’s cheating them out of a good story! Don’t only state the facts—give your interpretation, take a stand, show your beliefs and opinions and express the reasons why you think the story needs to be told. Don’t steer away from conflict — it is often a useful tool to engage the audience. If you appear to be disengaged, you will very soon have a disengaged audience too – this is often covered in presentation courses too. Speak of your own experiences in the first person (“I”) and involve the audience directly in the second person (“you”) – people take more interest and connect directly when the story directly addresses them.
A story about things is boring, especially data. Interesting stories involve characters—people who take actions and have experiences. They bring the story to life, they do things and show behaviours that are relevant. The audience often identify with a character and imagine themselves in the story. So, if you have to use third-person narrative, speak of characters by name instead of an impersonal “he” and “she”. Snow White, Aladdin, the Little Mermaid, name your favourite classic story – none of them would be so memorable if the characters didn’t have names and weren’t brought to life through their identifiable characteristics. Business stories also have characters. People cause and facilitate change. The change agent is the protagonist in a business story, he or she will naturally have opposition, and the other stakeholders are the supporting cast. The characters represent real people—customers, employees, suppliers and partners. They take action, have influence in a cause-and-effect chain, or have an interest in outcomes. If it’s too sensitive to use the real characters – you can’t really portray the unwilling-to-pay CFO as the “opposition” – then use abstract characters using personas. Well-researched personas will help your audience to identify with the characters that they interact with, which will make the story more real for them.
Bringing these two aspects together is how you draw the audience into the story. Show emotional reaction — sympathy, distrust, fear, anticipation, etc. — for the characters in the story in a way that guides the audience to find their own emotional responses. That way you guide them to make a decision in a way that they feel it was their own decision.
Make data persuasive through storytelling
We as data scientists are also the ones that need to give the data a clear and a persuasive voice. Remember, a big part of decision making is convincing others to adopt your point of view. Sometimes hard numbers and facts get you the buy-in or at least it draws the attention to the problem area or opportunity. But then it all comes down to making the connection – how to match the correct decision or action to the insight.
Storytelling allows you to talk about how the data relates to people and processes. You can use data-driven scenarios to illustrate what should be done, and what the potential impact or outcome would be. This means that you can use the data not only to give the insight, but also to inspire imagination.
So, in the conclusion of the story, we need to lead the business to make the right decision or take the appropriate action. And we need to make sure that decision or action and its implications have been properly researched and validated.
Listen
This may seem weird to require from a storyteller, but the real purpose of storytelling is to communicate, and communication is a two-way process. I have discussed listening in a previous blog – it is as important a skill as storytelling. It is the active, real-time interpretation of the audience’s response to the story. A business story will only achieve its purpose if the audience responds – and you need to be aware of that response, and the sentiments around that response, in order to steer or adapt the story to the desired conclusion. You need to seek feedback, invite questions, read facial expressions and body language, and respond to visual cues from the audience.
There are many methods to employ, such as asking direct questions, open-ended questions, leading questions, intentional pauses or gaps, and so on – many of these are useful to turn the story into a conversation too. For complex audiences or situations you may even need to take an additional person along who has a dedicated role to fulfil – a listener, observer and scribe to make sure you don’t miss or forget important bits and pieces. A good scribe will poll you to address outstanding issues.
More details about listening in the context of storytelling can be found in this post about communication in data science.
Concluding remarks
So, whether you are pitching to a new audience or a known one, remember the power of data storytelling and why it is necessary – it keeps humans infatuated by what the data is trying to tell them.
We all know that data has a great role to play – and it has become an integral part in many organisations’ strategic thinking. However, I would argue that it can only play a naturally strategic role within the business if the data is presented in a digestible and understandable format that leads to a proper conclusion – which is typically done through data storytelling. I believe following the tips outlined above will help to make your data storytelling alive, necessary and relevant today.
References
http://global.qlik.com/uk/explore/resources/e-books/3-reasons-why-you-need-a-story
https://tdwi.org/research/2016/05/ten-mistakes-to-avoid-in-data-storytelling
https://womensleadership.stanford.edu/stories
