Photo by author
# Introduction
Why do people misread your data? Because they are data illiterate. This is your answer. What is the end of the article? We can go home.
Image source: Tanner
Yes, it’s true; Data literacy is still at a low level in many organizations, even those that are “data-driven.” However, we don’t have to go home, but stick around and try to change the way we present our data. We can only improve our data storytelling skills.
If you’re looking to improve how you describe data in a narrative with structure, stories, and visual appeal, check out this guide. Building an Impressive Analyst Portfolio. It offers practical tips for creating data stories that actually resonate with your audience.
Knowing all this, we can make sure that our data is understood as we intended, which is, in fact, the only thing that matters in our work.
# Reason #1: You assume logic always wins
It doesn’t. People interpret data emotionally, through personal narratives, and pay selective attention. The numbers won’t speak for themselves. You have to get them to speak without ambiguity and room for interpretation.
For example: Your chart shows that sales are down, but the head of sales rejects it. Why? They feel that the sales team has worked harder than ever. This is a classic example of cognitive dissonance.
Fix it: Before displaying the chart, display this option: “Despite increased sales activity, sales fell 14% this quarter. This is likely due to reduced consumer demand.” It gives context and clearly provides a possible reason for the decline in sales. The sales team doesn’t feel under attack so they can accept the cold reality of falling sales.
# Reason #2: You rely on the wrong chart
A shiny chart may attract attention, but does it really present the data clearly and unambiguously? Visual representation is exactly that: visual. Angles, lengths and areas matter. If they are sketched, the annotation will be sketched.
For example: a 3D pie chart shows a budget category as more than it is, changing the perceived priority for funding. In this example, the Sales slice looks the largest because of the perspective, even though it’s exactly the same size as the HR slice.
Fix it: Stick to using chart types that are easy to interpret, such as a bar, line, 2D pie chart, or scatter plot.
In the 2D pie chart below, the size of the budget allocation is very easy to interpret.
Use fancy plots only if you have a good reason for it.
# Reason #3: The reason for correlation
You understand that correlation is not the same as causation. Of course, you do; You analyze the data. The same doesn’t often apply to your audience, as they often aren’t well versed in math and statistics. I know, I know, you think the difference between correlation and causation is common knowledge. Trust me, it’s not: the two measurements go hand in hand, and most people will assume that one causes the other.
For example: an increase in social media mentions of the brand (40%) is accompanied by an increase in sales (19%) in the same week. The marketing team doubles the ad spend. But the spike was caused by a popular influencer doing unpaid reviews. Extra costs had nothing to do with it.
Fix it: Clearly label relationships with “correlation,” “causal,” or “no proven link.”
If you want to prove causation, use experiments or additional data.
# Reason #4: You present everything at once
People who work with data tend to think that the more data they cram into a dashboard or a report, the more credible and professional it is. It is not. The human brain does not have an infinite capacity to absorb information. If you overload a dashboard with information, people will skim it, miss important data, and misunderstand the context.
For example: you can show six KPIs at once on a single slide, such as customer growth, churn, acquisition cost, net promoter score (NPS), revenue per user, and market share.
The CEO fixated on a small dip in NPS, derailing the meeting, completely missing the 13 percent drop in premium customer retention, a huge problem.
Fix it: Be a Slide Nazi: “One slide, one chart, one critical path.” An earlier example might go something like this: “Premium customer retention declined 13% this quarter, primarily due to service outages.” This keeps the discussion focused on the most important issue.
# Reason #5: You are fixated on health
You might think that showing granularity and raw numbers with six decimal places is more credible than rounding numbers. Basically, you’d think that more decimal places shows how complicated the calculation behind it is. Well, congrats on that complication. However, your audience gets stuck on round numbers, trends and comparisons. Sixth decimal place of accuracy? Vaguely disturbing
For example: Your report says: “Defect rate increased from 3.267481% to 3.841029%.” wtf!? People will get lost and lose sight of the fact that change is important.
Fix it: Circle the numbers and frame them. For example, your report might say: “Defect rate increased from 3.3% to 3.8% – a 15% increase.” Clean and easy to understand conversion.
# Reason #6: You use vague terms
If the terminology you use is vague, or the metric names, definitions, and labels are unclear, you leave the door open to multiple interpretations. Wrong in them too.
Example: Your slide shows “retention rate”.
Who or whose maintenance? Half the team will think it’s customer retention, the other half it’s revenue retention.
Fix it: Say “customer retention” instead of just “retention”. Be precise. Also, whenever possible, use concise and precise definitions of the metrics you use, such as: “Customer retention = customers active this month who were also active last month.”
You’ll avoid confusion and help people who may know what metric you’re talking about, but aren’t exactly sure what it means or how it’s calculated.
# Reason #7: You use the wrong context level
When presenting data, it is easy to lose context and present data that is too zoomed in or zoomed out. It can distort the perception. Minor changes may seem significant and vice versa.
Example: You show a 10-year revenue trend at a monthly planning meeting. Well, kudos for showing the big picture, but it hides a smaller, much more important picture: a 17% decline last quarter.
Fix it: Zoom into a relevant period, eg, the last 6 or 12 months. Then you can say: “Here’s the revenue for the last 12 months. Note the drop in Q4.”
# Reason #8: You also focus on averages
Yes, averages are great. Sometimes however, they don’t show a split. They hide the extremes and thus the story behind them.
For example: Your report says that the average customer spends $80 per month. Cool story, bro. In reality, most of your users spent $30-$40, which means that only a few high-spending users push the average. Oh, yes, the campaign that’s marketed based on your report, targeting \80 customers. Sorry, it’s not working.
Fix it: Always demonstrate distributions using histograms, boxplots, or percentiles of error. Use the median instead of the mean, such as “The median spend is \$38, with 10% of users spending \$190 or more.” With this information, marketing strategies can be significantly improved.
# Reason #9: You overcomplicate the visuals
Too many colors, too many shapes, too many labels, and legend categories can turn your chart into an unsolvable puzzle. Visuals should be visually appealing and informative. Striking the balance between the two is almost a work of art.
For example: Your line chart tracks 13 products (that’s 13 lines!) over 12 months. Each chart has its own color. By month three, no one can follow the same trend. On top of that, you added data labels to make the chart easier to read. Well, you failed! Data labels begin to resemble Jaime and Cersei Lannister – they are disturbingly intimate.
Fix it: Simplify the chart. Show the top three or five categories, group the rest as “Other”. Provide only important information; You don’t deserve to imagine all the data you have. Leave something for later, when users want to string.
# Reason #10: You don’t know what to do
Data is not the goal in itself. It must lead to something, and that is some action. You should always provide recommendations on next steps based on your data.
For example: you show that the rotation has increased by 14% and the presentation ends there. Well, everyone agrees that roaming is a problem, but what should be done about it?
Fix it: You should pair each big insight with an actionable recommendation. For example, say “Tuesday grew 14% this quarter, primarily among premium customers. Recommend starting a retention offer for this group within the next month.” With that, you’ve reached the ultimate goal of data storytelling – making business decisions based on data.
# The result
As any data suggests, you sometimes need to be an amateur psychologist. You should think about the people you serve: their backgrounds, biases, emotions, and how they process information.
The ten points I talked about show you how to do just that. Next time you present your findings, try to implement them. You will see how the possibility of misinterpretation is reduced and your work becomes much easier.
Nate Rosedy A data scientist and product strategist. He is also an adjunct professor teaching analytics, and the founder of StrataScratch, a platform that helps data scientists prepare for their interviews with real interview questions from top companies. Netcareer writes on the latest trends in the market, gives interview tips, shares data science projects, and covers everything SQL.