Is this what it means to catch a Stakeholder's cold?
Sometimes, the data is already there; we only need to reimagine how we approach it to make a positive change.
You’ve probably heard someone say that when designing data stories and visualizations, you should ‘tailor it to your audience.’
But in practice, what does that mean?
There are many ways of implementing it: You can choose a format that suits them, a type of visual communication, or a more or less engaging message. All of them are valid, and your choices can work to your advantage. But to me, the most impactful choice you can make is to change how you translate the final information of your analysis, considering your audience’s decision process.
I’ll give you an example.
From mortality rate to life-saving opportunities
My final project for the masters in Data & Design at Elisava was to support the Brazilian Ministry of Health in communicating the importance of implementing Breastfeeding Initiatives in maternities. Breastfeeding is the most impactful and direct way to prevent neonatal mortality, in other words, to help babies survive their first 28 days of life. This is directly connected to the Sustainable Development Goals, and although Brazil has reached some of the established goals, the pandemic has impacted these numbers.
Of the 7 thousand maternities in the country, less than 30% have implemented at least one of the four initiatives (QualiNeo, Baby Kangaroo, Baby Friendly, and Human Milk Bank). More specifically, of the 3,168 maternities in Brazil that have given birth to, on average, at least 100 babies in a year (justifying the investment), only 12% have at least one initiative implemented.
You can watch the explanatory video narrated by the amazing Ane Guerra to quickly understand the project.
So you see, this is not an IF problem but rather a WHERE. Logistics and investment must be considered: too many maternities and insufficient resources.
We also translated the data visually into what we call data portraits, which convey the nuances and details of each maternity.
There are many layers to this 12-month project, and you can check the final result in English and in Portuguese here. But back to our topic.
Before we started working with the health technicians, selecting which maternity to invest next was based on multiple factors, but mostly, it was a reactive process - if the maternity reached out and had an interest it would be considered for the program. We can say it was arbitrary, apart from the obvious choices of big maternities or the ones with very high mortality rates (rates are calculated as neonatal deaths/for every thousand births).
And this is where shifting the metrics helped change the decision making process. Instead of only looking at neonatal mortality rates, we focused on the causes of death (according to the International Classification of Diseases) that have a higher chance of prevention when babies receive breastfeeding on the first day of life.
And by looking at the percentage they represented in each maternity, we had a more detailed understanding of the number of babies we could save on a yearly basis. We called them investment opportunities.
The sum of these deaths represent 37.5% of all neonatal deaths, and they could be prevented if initiatives that promote breastfeeding were implemented. All states have a high percentage, but Alagoas, Tocantins and Amapá have the highest potential impact. Or the highest opportunity.
Neonatal mortality continues to be an important metric. But neonatal death can be associated with many causes and one initiative can’t solve all matters. When we focus on the potential impact the initiates we know can have and change how we approach it, our decision-making process can become more effective. Once we had that for the country, we could now look into each maternity that fit the criteria with the tool we built for the project:
This is the exploratory dashboard we built for the Ministry of Health. Each dot is a maternity. The horizontal axis represents the neonatal mortality rate, and the vertical axis can be changed to the metric they wish to understand and compare. By dragging and selecting the smaller charts on the left, we can choose the range of maternities we wish to visualize in the main chart. In this example, the opportunity size is selected, but we are also cross-filtering births, opportunity, initiatives implemented, and neonatal mortality rate, represented in the green selection. These selections give us the final result below.
There are less than 150 from all the 3,187 maternities that:
have no initiatives implemented;
have more than a 100 births/per year on average;
have between 35 and almost 70% of opportunity (percentage of babies that die before reaching 28 days and could be saved).
And if you select the ones with a mortality rate over 8.0 (considered high), that number goes down to less than a 100 maternities. I know you can’t always make those associations, but in this case, the change was very impactful and clear.
In this scenario I would even take a step forward when possible: if we were able to assess how much it cost to implement each of these initiatives we can combine it to our final decision. I know not everything translates directly into money. But $$ is a universal language after all and in places where it is limited, it can be a represent a big shift, but unfortunately, these things aren’t always transparent.
Bringing it closer to your audience’s language
Bad dataviz is not just about bad design choices. If the metrics selected to communicate don’t support the decision making process, it doesn’t really matter how incredibly beautiful or engaging it is. Sometimes, only improving how the information is designed or how the story is told won’t solve the issue. This is where domain expertise, statistical and problem framing knowledge play a significant role into making a positive impact in the way data is communicated.
Overall, tailoring something to your audience means understanding the questions and the decision-making process, and challenging yourself to propose new perspectives to the same data. The data was already there, our role was to shift the approach to build a common language with our audience.
I was once in Kenya for an event and a researcher said that ‘if your stakeholder sneezes, you better catch that cold’. Although I don’t agree 100% to that sentence, I believe this is an example of what he meant.
Knowing how to visually explain your metrics is important. But combining them with the metrics that are already present in your decision-makers day to day can directly impact lives.
If you have a challenge on how to bring your data closer to your audience, we would love to help. See you in a couple of weeks!