Data storytelling is changing the way we communicate data insights – for the better. By shaping our facts using the right techniques we improve understanding and actions. In this Data Storytelling Guide, we give our 2019 updated point of view demystifying many open questions such as what is the difference between data storytelling and data visualization or why is data visualization important as an asset for your stakeholders and decision-makers.
Our Data Storytelling Guide summarizes a series of best practices that will make your message remarkable. We will guide you through some of our data visualization examples prepared by our data analysts at Ladder.io when adding value delivering data insights to our happy clients.
INDEX:
We tell a story when sending suggestions to support a decision-making process or when spoiling the last tv show on social media to a friend. We tell a story when discussing the scores of last night’s game or when presenting the last quarter’s marketing customer acquisition numbers.
At the end of the day, what differentiates stories that stick from stories that are lost is the narrative’s engagement.
Storytelling became an umbrella term for a more humanized approach to convey information, especially for the digital era. Data storytelling leverages the same core concepts to better communicate data analysis outputs. Data storytelling is considered a new trending discipline and a valuable “data-driven” skill, translating complex analysis outputs into impactful, clear narratives that lead to business decisions. This mix of art, science, and presentation skills is grabbing the eyes of C-level executives. A new wave of data visualization jobs filling the gap between operations and decision-making teams.
Data visualization deliveries analytical information in a visual format that is easier and quicker to understand, instead of sifting through spreadsheets. Good data visualization simplifies data sets (large and small) into simple visual analogies. This is important in this new era of data, where we have more data then we know what to do with. The overwhelming amount of data makes it harder to understand what stories it’s actually telling, or the questions it’s answering. Its prime value is that it can provide business leaders with confidence and rationale for better decision making and stronger investments.
Have you heard that “a picture is worth a thousand words”? This is why data visualization is so important. Children’s books make a great analogy, we don’t read them often but still, most would remember the climax and conflict between Little Red Riding Hood and the Wolf.
Connecting what your audience sees with what they hear is what makes your sales pitch to stick, and this is why data visualization is important.
Now, more than ever, we need to be able to decipher truths in data and be able to accurately communicate a central idea, a value. Without good data visualization… without good data storytelling, it becomes impossibly difficult to be “data-driven”. And if you’re not data-driven, you’re leaving money on the table.
At Ladder, there are three major elements responsible for transforming a visualization element into a complete story:
Data visualization is not restricted to highly technical professionals and developers working with lines of code. In fact, this guide aims to turn even the least technical person into a competent storyteller. And the amount of data visualization tools created to facilitate the process of extraction, transformation, and digestion is incredible.
Now that you understand why data visualization is important and you have the core elements of our framework in mind, let’s cover the most relevant techniques (with examples!) to make sure you’re able to create and share your own meaningful data stories.
*Here’s a few free sources to get started:
Context is everything when communicating intricacies that are meant to drive business decisions. At Ladder, we refrain from reporting on insights without context. It makes deciding on next steps much easier.
*Interested in a closer look at our input-context-action framework? See ‘The Marketing Analysis You Need To Grow’ for our ICA framework in funnel analysis.
There’s no point of getting extra technical with stakeholders that are not concerned with the method but the output. The added noise is distracting and will make them feel less confident in trusting your opinion.
^This is especially harmful if you have client relationships!
When considering the context of your data story, think about the Who, What and How:
Who is receiving the intended information and how they usually interpret this type of information? Do you have a trust established? Do they prefer brevity or detail? These are important questions to assess how comfortable you are with the target audience and which tone will be used to address the major takeaway;
What do you want to communicate, or what action do you want to drive? Effective storytelling should prompt action and answer the “so what?” question that is raised when presenting data insights. Guiding your audience to take action and plan next steps is vital to the storyteller skillset;
Pro tip: As a best practice when action can’t be suggested, guide the audience towards an action plan discussion.
How can you communicate the data story most efficiently? Should it be a bar chart or line chart? Would an animation help? Will your color choices confuse or clarify? Is your chart to complex and hard to understand, or too simple and hard to grasp?
Pro tip: Whenever possible, a personal explanation can help your story to stick… Don’t have time for an in-person meeting? Jump on a screenshare. Or record a screenshare/walkthrough (we use Loom, it’s free).
One of the biggest challenges while presenting a data story is bringing attention to the points of interest that support your insight. We live in a distracted society. Below are some of the best practices to make the most out of your visualizations.
At Ladder, we maximize impact and communication success by using two simple strategies:
Both measures will reduce unintended friction between the desired takeaway and the audience. Let’s go through the basic concepts of both, starting with the cognitive load which adds important insights to chart selection.
Cognitive load is the sensation felt when facing a non-conventional design or intricate visualization. Cognitive load is a mix of confusion and laziness that immediately creates a mental barrier for consuming information.
Typical causes and consequences from Cognitive Load, from Paul Boag.
When the required effort to understand the message is too high, your audience will disengage. Simple visuals are preferred for easier consumption and understanding.
Cognitive load dictates both the ability and the motivation of your audience to interact with your information. If your story is based on simple visual elements that are familiar, the probability of your audience taking the intended action is significantly increased.
Let’s repeat this to make it very clear, bar charts are “boring” but they’re simple. As data storytellers, we want to reduce friction to maximize understanding. Keep it simple.
Pro tip: In the next section you will learn how to use visual queues to transform a boring chart into a professional good looking one without much effort.Work as an analyst, think like a designer!
This idea of reducing friction is backed up by much user experience (UX) research. At Ladder, whenever we design behavior we come back to BJ Fogg’s Behavior Model, the behavior psychology engine behind famous solutions and services such as Instagram. When we consider that every data story has a behavior change as a goal, reducing friction by understanding cognitive load means maximizing the likelihood of the goal.
At Ladder, apart from choosing an effective visual, we pay attention to the saturation and data-ink ratio, alignment, contrast, and pre-attentive attributes. Optimizing these parameters are key for successful storytelling.
Increase the amount of ink (or pixels) used to display data. Don’t be afraid of white spaces, they should be used to reduce saturation and guide attention. Eliminate visual elements that do not add information to make your data stand out more.
Note for a moment the “Z pattern” your eyes make while reading. This behavior repeats for every visual-oriented page. Only after the brain realizes that the content is not text-based that the eyes focus attention on things like labels or data points. Planning visual hierarchy will make it easier for your audience to focus on what’s important.
From the past example, the planned visual hierarchy helps to organize information.
Use color, elements size and bold text, only to guide the audience’s eyes to where the story is. Still, from our last example, note how blue tones were used to direct attention to the darker color while connecting the sub-headline to the data without adding a legend element.
Taking the time to understand how to display the information efficiently for your audience is a critical step in data visualization/storytelling.
Let’s go through some core principles we use at Ladder when designing our visual elements:
When working with small data sets, focus on the message you want to inform. If your visual looks too simple try to summarize it using text and a single number or statistic.
Ladder tool suggestion: Google Slides offers a lot of styling flexibility when summarizing charts into single numbers or statistics. There is no reason to go complex with lines of codes for this matter.
^This small data set is better represented as a statistic.
Use line charts when analyzing before and after comparisons or continuous data over a time period, i.e. Monthly sales over a year period.
Ladder tool suggestion: In most cases, Google Sheets for plotting and Google Slides for annotating will do the job, this was the exact process for generating the inclination graph on the example below. Google Data Studio has great features for time series visualization and it’s available for free. On the coding side, Seaborn has a beautiful API for time series plots.
Now, let’s look at an inclination graph used for pre/post analysis. In the example below, this is a visualization we created for a client explaining keyword movements. Depending on the context, the slope of the lines alone explains everything.
Pro tip: Note the pre-attentive attributes used for displaying maximum and minimum values in Google Analytics time series and the color contrast used to connect line and legend on the inclination graph.
Bar charts are one of the best and preferable visualizations due to its simplicity and familiarity. Horizontal bars are the easiest to represent categorical data, especially when having long labels. Horizontal bars follow the same logic of reading, the eyes scroll from left to right, starting on labels and reaching the data, which is the most intuitive way to read information.
Stacked bar charts, for total numbers or at 100%, add an extra layer of depth but also uncertainty to the visualization. In the example below, for two subcategories the stacked at 100% worked perfectly, when added the third subcategory, however, it is difficult to visualize the difference between January and February. Adding data labels would assist, but it goes in the opposite direction of keeping visuals clean from unnecessary saturation.
Ladder tool suggestion: Every visualization tool will have a bar chart option available. We suggest using Google sheets or Google Data Studio for design flexibility and Google slides for annotations. On the coding side, pandas df.plot(stacked=True) is probably the easiest way to generate stacked bars. In general Seaborn does an incredible job with sns.barplot().
Pro tip: It is extremely important to keep the y-axis at zero for accurate reference in bar charts. The rule is particular for bar visuals, and when not followed it can introduce a strong visualization bias. Cole Knaflic from Storytelling with data has a dedicated post for addressing this bias with a Fox News flawed example.
Tables interact with the verbal part of the human brain, therefore we read tables instead of visualizing them. The approach of scanning a table while pointing a finger, or the mouse pointer, is a good example.
Use tables when it is necessary to display many categories at once. The best use case is having stakeholders from multiple departments in the same meeting. Instead of presenting fractions of data on different slides addressing different managers at a time, a table might fit to organize information. This allows the audience to focus on different segments that are relevant to their individual objectives.
Also, there are still many hardcore Excel users out there and sometimes a table will be the most familiar visual for your audience. Remember, context is everything!
Ladder tool suggestion: There is no way to go wrong with tables, Google Sheets is powerful with great sharing capabilities. On the coding side, pandas DataFrame is your best friend. Seaborn heatmap’s with annotations are great for illustrating simple correlations. For a table with many columns try using pandas style.background_gradient() method instead for reducing saturation.
Let’s get this straight, no pie chart, no pizza chart, no donut chart. The human eye is not good identifying and correlating areas on a chart to numbers. The optimal choice here is the waterfall chart as a replacement.
Ladder tool suggestion: The good news is that the waterfall chart is now available in Google Sheets, but, unfortunately, not featured in Google Data Studio yet. On the coding side, here you can find a good implementation using matplolib. By the way, put PBP in your bookmarks.
^Waterfall chart representing a categories breakdown on a sales report
Pro tip: From Storytelling with data podcast episode 4, titled ‘It depends’, the question “Should I use a pie chart?” is brought up again from one listener. Cole mentions new studies from 2016 by Robert Kosara and Drew Skau, where they found that pie charts are actually visually efficient to illustrate a contrast between parts, saying that one part is very small or one part is really big compared to the whole.
Simplicity is key. Aim for maximum understanding at a glance. Refer to cognitive load principles when structuring your reports to better align with human behavior.
And, the easiest way to bake storytelling into your data visualizations is to frame each visualization and an answer to a question… what question does your report answer at a glance?
Sign up to our newsletter to stay up to date with all the latest movements in the field.
More from Ladder Blog
The art world is being disrupted by generative AI, and artists aren’t happy. Generative AI models like DALL-E, Midjourney, and Stable Diffusion were trained by scraping millions of images from the internet, without permission. Artists claim this violates copyright, while the companies training these models argue this falls under fair use.
Read More →In the world of digital marketing and data-driven decision-making, creative testing is a pivotal tool in achieving business growth. Gone are the days of relying on gut feelings or guesswork; now, business decisions are powered by data-validated insights, meticulously collected, analyzed, and validated. This transformative process empowers businesses of all sizes, from established enterprises to budding startups, to thrive in an ever-evolving digital market. This article looks at the practical applications, challenges, and innovative transformations associated with creative testing, offering you valuable insights and actionable strategies to implement in your own digital marketing efforts for achieving growth and success.
Read More →
