Data Storytelling: The Way to Stakeholders’ Hearts(Part 3)
Motivating stakeholders to act on your recommendations
Stories change people while statistics give them something to argue about.” — Bernie Siegel, author and surgeon
Why this article?
In this article, you’ll learn how to write compelling data stories using an example that I provided.
So I have a big announcement. Drum rollssssss.
This will be the last article of this series. Yes! I’m glad you’ve come this far and I’m sure you’ve learnt a lot. But this lot that I claim that you’ve learnt won’t be complete without this last article, right?
So let’s dive right into this loaded article.
But hold on! If you’ve not read the other articles on the topics: What are data stories and why & when you should tell data stories, you should definitely check them out.
Now let’s begin and don’t forget to grab your coffee😋
How to Tell Compelling Data Stories
Take note that it’s not easy to craft data stories. It will take time to craft a story that will get to the emotions of your stakeholders but when you eventually do it at the appropriate time and when you do it well, the outcome will be mind-blowing.
I’ll be dividing this into three phases: the pre-data storytelling phase, the data storytelling phase, and the post-data storytelling phase. We’ll go through this step by step so you can understand what it means to craft data stories and get results. Emphasis on “get results.”
Pre-data storytelling Phase
Identify your audience
Audience?🤔 Okay, maybe we should say stakeholders. If you’ve read my first article, you’ll know why I’m deliberating on the right one to use. In this context, I’m referring to stakeholder but I’ll use audience most times because that’s what you’re familiar with.
You’re not an analyst for designing dashboards or writing data stories. You’re an analyst for solving business problems with data. And one thing you should understand is, you can’t successfully solve a problem if you’re trying to solve it for everyone. This is why you need to have a specific audience that you’re targeting. Different audiences require different approaches to solving a problem.
The CEO of a company is more interested in the overall performance of his company while a manager is more interested in the performance of his team alone. See? Two different audiences. Two different interests which require you to use different approaches to tackle them.
So before you even ask for the available data, your question should be, “who am I solving this problem for?”
Ask questions
The next step is to ask questions to your audience and other stakeholders. While asking questions, understand the business problem(s), the impact that the problem(s) have had on the business, the available data, the expected outcome, the expected deliverable, and many other questions that can help you to understand both your data and the business.
You need answers to these questions to know what to highlight in your story and what to not include in the story.
Prioritize business problems
This is optional if only one business problem was mentioned. But if during your course of asking questions, multiple problems were mentioned, then it’s important to prioritize these problems. Trying to solve all problems with one story will only be overwhelming for both you and your audience.
How can you prioritize the problems?
This is where domain knowledge is important. An understanding of the industry will help you to know the problem that has a greater impact on the business. You can also evaluate the impact of these problems on the business by asking questions.
Analyze the data and extract insights
At this point, you’re at the most critical stage that determines the direction of your data story. With an understanding of the business, you’ll clean and explore the data for insights. It’s not uncommon to extract several insights from one dataset. But after having extracted these several insights you need to pick one central insight that you’ll run with in your story.
You could pick this central insight by asking these questions to each insight that you find:
- What is the impact of the problem(insight) on the company?
- What is the expected outcome?
- Are there activities that stakeholders are already putting in place to solve this problem?
- Is the effectiveness of the solution measurable and how can it be measured?
Take note that Insights are different from observations. Your insight(s) should be the Why of your abnormal observation(s). Take note of the word, “abnormal” and abnormal can refer to a good or bad thing.
Finally! You’ve analyzed your data, extracted the insights and even selected the central insight to use in your story. It’s time to go into the data storytelling phase.
But wait. Hope your coffee hasn’t finished yet. If it has, go and refill, I’ll wait because this part requires all attention.
Data storytelling Phase
This is the phase where you begin to craft your story after you have your central insight. Finding data stories crafted in a business setting is really difficult because of data privacy issues.
So to drive home my point, I’ll be taking you through you a data project done outside a business setting. This will enable you to understand deeply how to tell compelling data stories.
Project Preamble:
The project was done to solve the problem: which webpage should be improved? for an e-commerce company that wants to improve the sales of holidays on their website.
So let’s go right into it, shall we?
Set background
First of all, don’t tell your audience how you analyzed your data. In a business setting, they are not interested in how you got your insights but in what your insights are.
Start with showing what is normal in your data, that is, the observations that don’t show anything wrong or extraordinary right. This will make them comfortable before you introduce a conflict.
In the context of the sample data project, the bounce rate was above the target. Hence, the need to start with it to make my audience comfortable.
You could also add more normal observations to build your story.
Here, I went further to show the high session duration that also implied progress for the business.
Introduce a hook
This is where you introduce an abnormal observation. Notice that in the first stage, you showed them normal observations. Showing these normal observations first makes them appreciate the abnormal observation that you’ll introduce later.
This is where your data story gets interesting because your audience will want to know the reason for the abnormal observation.
Notice how the abnormal observation below could make the stakeholders very uncomfortable because the most important metric that determines the revenue of the business is below the target.
So now you have their attention already. What’s next?
Go deeper
You’ve uncovered your hook but don’t stop there. You need to build up on this hook to further stir emotions in your target audience. In this stage, you’ll show visualizations that contribute to this anomaly.
What numbers are influencing it? What factors could be its significant contributors? When all these come together, they make your narrative stronger and more convincing. Your stakeholders get afraid of or excited about what this could do to their business.
Going back to our sample data project, the recurrence of poor metrics of the bottom landing pages was shown. If you look closely, you’ll notice that at least two landing pages had recurring poor performances in the three metrics.
Please take note that you could choose different methods of revealing your observations to your audience. It’s not advised to have different charts in one place but because I wanted my audience to compare these charts, placing them side by side felt like the best method for this use case.
Reveal central insight
At this point, you’ve peeled the layers of your hook to the core. This core is your central insight; the big problem or breakthrough for the business. Depending on this insight, you could choose to emphasize it further or just reveal it and move forward.
Because this sample data project was done outside a business setting, it was difficult to figure out the central insight. Having a low website conversion rate could be due to several things and it is inadvisable to present all these factors to the audience without figuring out the exact one that’s causing it.
Central insights are usually gotten from interactions with stakeholders while asking questions.
State Solutions
At this stage, your audience will most likely be asking how this problem can be solved. It’s your duty to tell them what to do. This is where you make recommendations to them. But be careful here. Doing it wrongly could overwhelm them and they could dump your solution because they don’t know where to begin.
You could take two approaches here and it depends on your kind of recommendation(s). I’ll categorize these into two: stand-alone recommendations and complementary recommendations.
Stand-alone recommendations are different actions your stakeholders could act on. They are multiple but each of them can single-handedly solve this problem. Complementary recommendations are the ones that can’t solve the problem all alone. They have to be applied together to work effectively to solve the problem.
Again, you need to have the domain knowledge to identify stand-alone recommendations and complementary recommendations. For stand-alone recommendations, don’t make the mistake of listing all of them and expect the audience to choose the right one.
It’s your responsibility to state each of these recommendations and evaluate the degree of impact that each of them will have on the business problem. Then pinpoint the best recommendation backing it with the fact that it has the highest degree of impact on solving the problem.
For example, to improve the publicity of a sporting shoe business, there could be two possible solutions. One is running ads and another could be hosting a marathon competition for people to participate in.
To pick the right one, you could predict that the ads would improve the revenue by 30% while the marathon competition could improve the revenue by 50%. Obviously, they would go for the latter because it has a higher impact.
The complementary recommendations will require you to list all of them but don’t just list them in bullet points. Create a roadmap for it showing how long it will take for each of the recommendations to be carried out and if possible include the person that will work on effecting each solution (your recommendation).
Ensure that you’ve confirmed these timelines from the persons responsible so there won’t be a conflict between the recommended timeline and the exact time it will be carried out.
In the sample project, the recommendations were complementary. A roadmap wasn’t used because it was outside a business setting. Just as stated before, to create a roadmap that drives results, you need to interact with stakeholders to understand the best timing.
However, here’s a sample created for the sake of this article so you could understand what I mean by complementary recommendations on a roadmap.
Show consequences
Yes, you’ve informed them of what to do. But you still need to make them act fast. You can do this by showing them what will happen if they don’t act. This is where you apply predictive analytics.
To what degree will this problem excavate if it’s not acted on fast? You should answer this question with a visualization to easily drive home your point.
To ensure that my target stakeholders act on my recommendations in the sample project, I forecasted the conversion rate if they don’t act. If they see that the already low conversion rate would go lower in two months, it would make them want to act immediately.
Great! You’ve told your story and that’s it? Nope. Not yet. Let’s go over to the post-data storytelling phase.
Post-data storytelling Phase
Remember when I made emphasis on getting results after telling your stories? Even though your stories could get you results if told well, it’s advisable to still take further steps to guarantee results.
This step is a Follow-Up.
Make the move to follow up on your recommendations by measuring the impact they had on the business. It could be easier if your story revolves around the functionality of the product offered by the business. At least, you could easily measure the revenue or the user growth.
But what if your data story revolves around improving the operations of teams in the company? If this is the case, then you need to constantly follow up with the stakeholders and quantify the impact that your recommendations had on the team when they applied them.
Making efforts to follow up on your recommendations puts you in the spotlight of showing everyone in the company that the data team is valuable. You don’t have to beg them to see your value. You make them experience that value and they’ll testify.
Conclusion
Wow! That was a really long ride and I’m so glad you followed me through. We are still in that era where data teams have to earn their respect in workplaces.
The most effective way to do this is to show your company the value you offer by making them act on your recommendations and get good results. You can easily do this with a data story.
In this LINK is the sample data story that I worked on. I only showed snippets of it in this article to aid your understanding so you can follow the link to see the details of the story.
Remember, to make your stakeholders act, you need to trigger their emotions. And to trigger their emotions, you need a data story.
Cheers to soon being that highly valued data professional in your organization!
Below are the other articles in this series. Enjoy🤗
