Data storytelling: creating insightful narratives from raw data

We’ve all heard the saying ‘a picture is worth a 1000 words’; it’s a powerful statement that, when Googling the phrase, is often incorrectly attributed to Albert Einstein. Wasn’t E = mc2 enough for him?

The saying is regularly used in advertising, but when used in this medium, it’s not so much the 1000 words that matter - it’s the feelings that they induce, and the connection it builds with that particular brand.

And when it comes to data, it’s no different. Providing context to data can change the way we understand it, and the narrative we intertwine plays a huge role in the way we interpret and connect with it.

To put it another way, the picture you paint is the key influence in the reactions and results we draw from any set of data.

Of course, this is great news when this data story is told correctly. After all, data has become one of the most important assets at your business’s disposal, and is a critical component in your decision making, growth, and strategy implementation. As such, any way of making this data more accessible and easier to interpret is invaluable…right?

In theory, absolutely! But beware: data stories can also be misleading, and this misguidance can have significant implications.

So, how can data be misleading but tell a good story? Let’s look at an example.

Using percentages is an age-old way of presenting a data story in a way that will only ever support the narrative you want to tell, as opposed to painting the true picture. For instance, your marketing team might provide an update that downloads have increased by 500% - and of course, that’s great news at face value! That’s until you discover that you only had one download the month before, meaning you now have a grand total of…5. Not so great news.

In this case, month-on-month percentage growth, while telling a clear and accessible story, isn’t the right metric for creating a supportive narrative that will be meaningful to the decisions you make. In fact, by not painting the full picture, it’s the complete opposite.

Instead, a fixed numerical value that measures progress against your objectives would be more effective here. Depending on your KPIs and the North Star metric you’re working towards, you could compare this fixed numerical value against a growth target (perhaps around 10%-20%) and use data visualisations to track and measure this progress more effectively. This enables you to tell a story that manages expectations and is less likely to be misconstrued.

Using data visualisation in these instances is a super effective way of breaking down data and presenting it in a way that’s more visually engaging (yep, hence the name). Think graphs, charts, maps, and the like. This makes the data much easier to comprehend, and when it’s combined with an engaging data story, it humanises the data by giving it real-world meaning.

And it’s this real-world meaning that’s the real difference between a good data story and a bad one. Narrative is universal in that it helps us process and remember information in a way that keeps us engaged, and helps us communicate ideas in a way that’s both digestible and impactful. By providing a new layer of context through real-world meaning, you’re able to enhance the narratives in your data stories by reinforcing understanding.

To put it in another way: isolated numbers numb us, while stories stir us. Through data storytelling, we can become less concerned with proving, and more focused on moving. We know, we know, we’re expecting to be named Poet Laureate any day now.

To prove our point, take this statistic as an example: in America, Just Eat (or as they call it in the US, GrubHub - yuck) receives 8,683 orders. On its own, this data is pretty meaningless - the number is simply too large for us to interpret or contextualise with any real value. However, if we humanise this data through a narrative, it becomes easier to assign real world meaning to it. For instance, every ten minutes, GrubHub receives enough orders to feed a capacity crowd at Wembley Stadium. Suddenly, this data is a lot easier to comprehend - it’s a whole lot of chicken tikka masalas.

An excellent example of effective data storytelling using this method is Spotify’s annual Spotify Unwrapped. In this campaign, Spotify takes its users’ listening data and tells the story of their year in music. For example, let’s say we’ve listened to 10,000 hours of Rick Astley in 2023 (what’re you laughing at, Rick is cool again in 2024!). This data doesn’t really tell us anything when isolated (other than we’ve probably listened to too much Rick Astley), so Spotify uses narrative and data visualisation methods to enhance its meaning. For example, they might tell us that we were in the top 0.5% of Rick Astley listeners last year, provide a graph comparing how much we listened in one month compared to the next, and even map out how our listening habits compared with others around the globe.

Using similar methods in your own data storytelling adds significant value, no matter if you’re sharing internal insights or telling data stories to your customers, and ensures your audience remains informed, invested and engaged. In other words, tell a good data story, and your audience is never gonna give you up…we’ll see ourselves out.

The do’s and don’ts of data-driven storytelling
Here are the techniques to embrace and mistakes to avoid in order to tell a powerful data-driven story that provides meaningful insight.

Do

• Verify your sources: before you can begin telling enticing stories with your numbers, you need to establish what data you need to collect. Remember that it’s unlikely all your data will be coming from one place, so whether it’s user data, customer insights, or internal metrics, ensure you’re always collecting from reliable and accurate data sources
• Remain objective: as a good rule of thumb, remain objective by letting the data guide your narrative. It’s generally considered more ethical to tell the story around the data rather than shaping the data around a preconceived story, and although the latter does happen in some instances, it is not advisable in order to maintain trust with your audience and ensure your decision making remains well informed
• Use visuals wisely: employ charts, graphs and infographics to enhance understanding, not to mislead. When adopted correctly, data visualisation methods can transform numerical and non-numerical data into an engaging visual summary that’s far more effective than looking at rows and columns stuffed full of numbers, all while reinforcing that ever-important context in your narrative
• Keep it accessible: remember that the whole point of your narrative is to enhance the accessibility of your data. With this in mind, make sure your story is understandable to your audience by providing a relatable context, personalising where relevant, and avoiding jargon and overly complex analysis
• Make an impact: whether it’s to educate, inspire or evoke action, all good stories have a purpose. Ensure there’s meaning behind your narrative by establishing the reason for telling your story; what narrative are you trying to tell and how can you paint that picture for everyone to understand?

Don't

• Cherry pick data: avoid only selecting the data that supports your narrative, and never intentionally alter or misinterpret data to make it fit your story. Always be weary of the quality of your data too: never prioritise incomplete, inconsistent, or outdated data simply because it tells a better story. This is a sure-fire to compromise trust and misguide your decision making
• Overcomplicate: don’t overwhelm your audience with too much data or overly complex visualisations. For example, avoid cluttering data dashboards with too much information, as this can quickly become overstimulating and undermine the entire point of your approach. Similarly, focus on one main area in your narrative to avoid diluting your takeaways by trying to communicate too much in one message. We recommend focusing on the areas you know best - deep insight on one takeaway is a lot more valuable than top-level insights on a number of different takeaways
• Ignore Context: always provide the necessary context for your data to avoid misleading your audience. We all make assumptions and miscalculations based on perceived biases, especially when it comes to data and seeing the progress we want to see. Providing that all-important context to your narrative ensures understanding and impartiality, meaning insights remains consistent and well informed
• Breaching data ethics: when collecting and analysing the data to drive your story, it’s super important that you respect the rights and privacy of data subjects and sources in order to ensure integrity and honesty in your analysis and presentation (oh, and the small matter of also ensuring you’re abiding by the law!)

*How to tell your data story *
It’s important to establish that data storytelling isn’t just another passing fad like the latest TikTok trend (not that camping overnight for the viral Stanley tumbler isn’t completely normal behaviour or anything…).

In fact, as businesses continue to harness more and more data every day, nailing your data storytelling approach only becomes more integral to your ability to leverage these insights effectively, and gain a competitive advantage as a result.

So, what steps must you take to create a clear and cohesive data-driven narrative that achieves its purpose?

1. Start with a clear question or statement It’s important to set the tone early.

Your story should be driven by a clear, concise and compelling question or objective that is relevant to your audience. This focal point guides your data exploration and analysis, and ensures that your narrative has a defined purpose and direction.

By considering what your key message is and what emotions you want to drive in your audience, you can better understand the language you must lead with in order to support that goal and maximise the chances of your desired outcome.

For example, if you’re analysing sales data, a question like, ‘what factors contributed to the highest sales quarter in the past two years?’, sets a clear path for investigation. In turn, this should then help you map out the direction you take, exploring what has actually led to the highest sales rather than assumptions about what you think contributed.

1. Use visuals to enhance understanding Use visualisations such as charts, graphs, and maps to make complex data more accessible and engaging.

And don’t just take our word for it. To dig into the science behind this for a second (we can hear Sheldon Cooper cheering from here), studies have shown that, of all the information transmitted to the brain, 90% is visual. And this is why, as proven by Robert Horn at Stanford University using studies from other academic institutions, data visualisation can harness significant results like:

• A 21% increase in a group’s ability to reach consensus
• A productivity gain by shortening meetings by 24%
• A 43% increase in persuading audiences to take a desired course of action ‍ However, it’s crucial to choose the right type of visualisation for the data you’re presenting. For instance, use line charts for trends over time, bar charts for comparisons among categories, and maps for geographical data.

Ensure these visuals are clear, labelled correctly, and free from misleading scales or distortions to paint the clearest picture possible. Get it right and you can unlock a series of benefits, like breaking down complex data sets more effectively, identifying patterns and trends quicker, and making it easier to measure progress and outcomes - all key elements in extracting the most meaning and enhancing understanding to help inform your narrative.

1. Tell a story with a beginning, middle, and end Good news: you’re now in a position to begin crafting a narrative around your data. Without coherence and without a story, your data simply remains a collection of uncoordinated facts - or to steal Anthony Tasgal’s very fancy coinage, a ‘spewed litany of inert factoids’. Although we can’t be the only ones that think ‘inert factoids’ sound like the next Doctor Who villain?

But what makes a good story when it comes to presenting data?

Narrative serves to establish patterns with meaning. Therefore, a strong data story doesn’t just provide an overview, but rather frames your insights in a way that’s relatable and meaningful. In simpler terms, it provides meaning through structure.

Structure your data narrative like a traditional story in order to capture your audience’s attention and create a framework that they’re already familiar with. To hark it back to your old English classes, adopt a simple narrative arc of beginning, middle and end.

Start with an introduction that sets the scene and outlines the question, mission statement or problem you defined earlier - this is your beginning. Simples. At this point, you can also opt to establish a conflict - author John le Carré (you’ve no doubt at least heard of Tinker Tailor Soldier Spy) once famously said that, “the cat sat on the mat is not a story, but the cat sat on the dog's mat is”, and embracing this idea in your narrative will help to establish both relevance and importance in the story that the data is telling. What’s the problem you need to overcome?

The middle of your narrative is where you’ll highlight your key findings and how they relate to your question, statement or conflict. It’s important to note that, in order to keep things engaging, this part doesn’t just include the quantitative data you’ve collected, but also qualitative data that adds context, texture, and nuance. The combination of the two is what ensures your data story resonates with your audience.

Finally, conclude with a summary of your insights and their implications. What has the data told you, what does this mean for your initial question or statement, and how does resolve your conflict? This is your ending, and the point at which you really drive home the purpose of your story - what are the key insights your audience needs to come away with and how does this enhance their understanding of the matter at hand? Where relevant, you can also include a call to action or suggestions for further inquiry at this stage; again think back to the purpose of your story, and consider what actions you wanted to inspire.

Adopting this simple but effective narrative structure helps maintain attention, enhance understanding and improve the memorability of your data story, and makes complex data more digestible by framing it in a way that’s familiar and engaging. There’s a reason it’s the go-to structure for storytelling!

TL;DR
Not a fan of long posts? Too busy to read the whole thing? Need to go out and queue for a Stanley quencher? No worries, we’ve got you covered! Here are the main takeaways you need to know:

• Providing context to data can change the way we understand it, and the narrative we intertwine plays a huge role in the way we interpret and connect with it
• Data stories can be misleading when data sets are picked to support a narrative, and this misguidance can have significant implications
• When telling a data-driven story, you should always verify your sources, remain objective, use visualisation wisely, keep it accessible, and strive to make an impact
• When telling a data-driven story, you should avoid cherry picking data, overcomplicating, ignoring context, and breaching data ethics
• Your story should be driven by a clear, concise and compelling question or objective that is relevant to your audience
• Use visualisations such as charts, graphs, and maps to make complex data more accessible and engaging
• Structure your data narrative like a traditional story with a beginning, middle, and end in order to capture your audience’s attention and create a framework that they’re already familiar with

Remember, the goal is not just to present data, but to make it tell a story that is both informative and captivating. This involves not only showcasing the data in a way that’s engaging, but also connecting it to a larger context that resonates with your audience.

So the next time you’re faced with a data set, remember to ask yourself: what story does this tell and how do I communicate this most effectively? That’s the secret to a good data story.