December 9, 2024
December 10, 2024
We’ve previously discussed the impact of data storytelling in today’s crowded media landscape, and it has only grown in value. To capture journalists’ attention, we need to give them information that can make their coverage stand out, by showing them the story versus telling them.
But, how do you bring a compelling story to life? I recently caught up with Adam Slater, the founder of Clarafy Research, for his take. For nearly two decades, Adam led reputation and public affairs insights programs for global brands at top agencies, such as Purple Strategies and ClearPath Strategies, and polling firms. He now specializes in public opinion research and insights for communicators, strategists, public affairs professionals, and organizations looking to better understand audiences and issues through data storytelling.
What follows is an edited version of our discussion.
Tell us about yourself and Clarafy? What led you to research and data?
My love for this field goes back to collecting baseball cards as a kid. I was a long-suffering Mets fan from New Jersey, but obsessed with memorizing players’ statistics and knowing the World Series champions each year. That love of statistics grew into politics and election results as I got older, which eventually led to polling data.
Survey data fascinates me – in a way, it levels the playing field and provides an effective truth serum against anecdotes or outdated beliefs. What I do at Clarafy is help a variety of folks better understand their audience and issues through public opinion research tools (e.g., surveys and focus groups). This helps stakeholders:
- Measure campaign progress,
- Understand messages and communications approaches; and
- Pinpoint audiences that need to hear those messages.
<split-lines>"Survey data fascinates me – in a way, it levels the playing field and provides an effective truth serum against anecdotes or outdated beliefs."<split-lines>
What advice would you give to someone who is new to analyzing survey data and data storytelling?
I would break it down into three lessons:
First, try to be heavily involved in the question development process, or at least deeply understand the questions before you get data back. Thinking about the questions ahead of time helps you determine ideal headlines or provide insight that helps build the strategy.
Next, layer your analysis. Don’t overthink it at first; just review the data and see what is most surprising and most expected. Get those numbers on paper – then, start to layer how those numbers might support one another: Which might be most impactful to your audience [your headlines]? What data points help build your story [your supporting bullets]? What data doesn’t fit the story [your throw-aways]? It’s OK to discard some data points – sometimes expected questions don’t fit the final narrative and that is fine.
Third, put yourself in your readers’ shoes: Why should they care? Why is this valuable to them or worth reading? Then, go back and punch up your headlines and analysis to optimize the impact for your specific reader.
Are there any common pitfalls to avoid or best practices to keep in mind? What is the most common mistake you see people make?
In my view, the goal of a survey is to simulate a real conversation with each survey participant. Survey writers would be well-served to put themselves in their respondents’ shoes with each survey they put in the field.
A lot of my work focuses on emotions – and many polls for public release or internal strategy don’t focus enough on understanding how people approach an issue emotionally. Most only strive to ‘understand’ or ‘persuade’ with logic. This is common in political campaigns. Some do a good job of connecting issues to emotions; others believe that merely raising an issue will be sufficient to change their vote. With so much competing for our attention, we have to find ways to better connect with our voters, customers, or stakeholders.
<split-lines>“Even if we don’t realize it, emotions are central to everything we do and every decision we make – and if we fail to understand them, our survey data will likely suffer, as will our specific issue or cause.”<split-lines>
What are some general tips and tricks you would recommend for maximizing the value of survey data?
When approaching any survey, you need to be thinking about the scientific method you learned in grade school and go through those steps. Define the problem, conduct background research, form a hypothesis, etc.
The hypothesis is incredibly important – and it’s often underutilized. Many times, you’ll define the problem, and conduct background research to get smart on drafting questions, but skip that hypothesis generation, which can undercut the quality of your insights and the value of the survey.
What’s your take on using Generative AI for data analysis?
I think Generative AI is a helpful first step. It can expose you to new ideas, bring new considerations to light, and generally help with brainstorming. It can be particularly valuable to remote workers or solo business owners as a brainstorm tool to define a problem or begin to solve an issue.
But, while helpful in some areas, there are still some pitfalls when it comes to opinion research. I’ve tested it out with both questionnaire development and analysis – it’s done better on question development, providing several interesting topics and lines of questioning I hadn’t considered.
On the analysis side, GenAI can provide a summary of inputs, but human thinking is truly unique when it comes to turning data into strategy. Discerning the insights you need is critical and while the tools are helpful in a lot of areas, they are reliant on inputs – so you risk a ‘garbage in, garbage out’ approach without human refinement.
<split-lines>"GenAI can provide a summary of inputs, but human thinking is truly unique when it comes to turning data into strategy."<split-lines>
And, stay tuned for more content and insights from Mission North's Data Team!
