MLOps Zoomcamp: Free MLOps course. Register here!


Transforming Data into a Narrative

Characters, conflicts, and a conclusion

25 Dec 2020 by David Gates

The Challenges of Storytelling

Whether you’re working in data analytics or data science, the ability to turn your data into a story is an essential communication skill. Your insights are only valuable if you can communicate them effectively. Many professionals clearly see what the data is telling them. They present their findings by repeating these insights as they see them. They are then surprised when others fail to see the value in their data. The issue is not the quality of the insights, it is the lack of a story. We all process information a bit differently. Facts connect with some, and not others. Stories, however, are universal.

Image by Free-Photos from Pixabay

Data professionals do not need to be natural storytellers. Instead, they simply must understand the elements of a good story.

Great stories don’t simply communicate events and facts. They frame them in a way that engages the audience from start to finish. Great stories have 3 essential parts: characters, conflicts, and a conclusion.

I have a good friend who is terrible at telling stories. No matter how interesting the topic, his stories always seem to fall flat. He tends to focus on facts and is generally unable to create conflict or tension with relatable characters. Simply put, his stories are boring. When listening to him, I am rarely concerned about the conclusion. His stories end with a casual laugh and then someone quickly changes the topic.

My sister, however, is a fantastic storyteller. She is able to make the most mundane events engaging. Her stories seem to connect with ​everyone. ​ People laugh. Listeners pay attention. Others retell her stories. She combines characters, conflict, and a conclusion seamlessly. A good business story must do the same.

Characters + Conflict + Conclusion

First, every story needs characters. Data affects somebody. Identify these people. It could be stakeholders such as customers or employees. It could be the general public. It could even be your audience members.

Next, identify your conflict. It should relate directly to what the characters are experiencing. What problems are they facing? Why are these problems so important? How are these issues affecting their well-being? Maybe the characters are your audience members. In that case, ​ what problem are they causing?

Finally, a good story must have a conclusion. A conflict needs resolution. This is where your insights into the situation become essential. As a data expert, you understand the problems and know how to solve them. You can restore balance by eliminating the conflict. This motivates your audience to act and provides a satisfying ending to the story.

Creating a narrative

Let’s take a look at a simple example using demand forecasting at a shoe store. We could directly state the problem:

We do not have enough sizes or styles at the right times. Demand forecasting models can mitigate this problem.

This approach, however, lacks engagement.

So how do we engage our audience? We’ll want to use a hook during our introduction that introduces our conflict and characters. In our shoe store example, our conflict is the lack of stock. The characters could be customers, employees, or the organization as a whole. Here’s a potential introduction:

I don’t buy shoes very often but they’re something I’m willing to spend money on. Last year, my beloved pair of Clarks desert boots reached the end of their life. It was unfortunate timing as they’re my go-to shoe and I had big weekend plans.

So I ran to the nearest “Shoes R Us.” I arrived and was immediately greeted by a helpful assistant named Paul. I told Paul what I was looking for: “brown Clark desert boots size 10.” He went to the back and returned empty-handed. I received the dreaded “Sorry we’re out of stock but we have….” I told Paul sorry but I’m not interested.

Image by David Mark from Pixabay

I knew I didn’t want any of the alternative options, so I left the store and was able to find what I wanted at a competitor down the street.

Shoes R Us lost out on a sale of a shoe that costs over $120.

As the customer, I lost out on time and had to reevaluate my loyalty to the company.

Paul lost out on commission.

This situation can lead to decreased customer loyalty, employee turnover, and lost revenue.

Now, this problem isn’t unique. Given the vast stocks that we must have at every “Shoes R Us” location, it will certainly happen from time to time.

So how can we solve this issue?

Luckily, we can create machine learning models that will help us predict the demand for each style and size at all of our retail locations across the country.

Here’s how it works…….

After hooking your audience during an introduction, you’d then go on to explain the issue more in-depth. This is a more functional part of the story where you can introduce your methodology, data, and model. You can expand on the issue, highlight your knowledge, and offer ideas on what can be done to solve the problem.

Finally, it’s important to end your story on a strong note. During your conclusion, you can relate it back to what was discussed during your introduction. One of my favorite techniques is to highlight how things may be different for various stakeholders. So using our previous example I could mention how customers like myself would have increased loyalty, how employees like Paul will make more money (and likely be happier in the process), and the company will see an increase in sales.

Your story has first highlighted the problem, you then showed how your knowledge can resolve the issue, and finished by describing how the situation will change. This can be incredibly motivating. It’s likely your audience wants this problem solved. The information you’ve presented is the bridge between the current problematic situation (which you highlighted during the introduction) and an ideal situation (which you introduced during your conclusion). You are giving them the power to eliminate this problem.

Excellent storytellers are made, not born

The most successful data professionals combine world-class analytical abilities with expert communication skills. These individuals are well-known for their talent to turn insights into action. How do they do it? Through stories. Storytelling is a do-it-all tool. It helps you simplify complex ideas, engage an audience, and motivate action.

Image by Roché Oosthuizen from Pixabay

So do your communication skills match your analytical skills?

Developing storytelling abilities in a foreign language, such as English, is challenging, but it’s a skill that can be developed by anyone. It will allow you to rapidly advance in the field. It opens doors to international opportunities, can lead to promotions, and makes you an invaluable member of your organization.

You can learn more about this topic from my talk about essential communication skills for data professionals:

You can find me on LinkedIn and on my website

Subscribe to our weekly newsletter and join our Slack.
We'll keep you informed about our events, articles, courses, and everything else happening in the Club.

DataTalks.Club. Hosted on GitHub Pages. We use cookies.