It’s a familiar scene: you’re in a meeting, and someone projects a dashboard with graphs of your marketing campaign outcomes. You can see a general upward or downward trend, but it’s hard to know where that data is coming from, why these patterns are occurring, and what to do next.
Aaron Maass, our SVP of Data + Insights, is the latest guest on Lea Pica’s Present Beyond Measure podcast. The conversation is jam packed with stories and incisive advice from his 20 years of industry experience on how to discover insights from data, communicate analysis to stakeholders, and effectively manage data science teams.
When sharing data with stakeholders, presenting information visually is just as important as gathering and analyzing it. Aaron calls data visualization “a means to an end” that allows analysts to see how patterns emerge beyond what a simple data table could provide.
One his best data visualization tips is to start by examining data through the lens of its consumer – in this case, the client or other stakeholder – and consider how it might be interpreted from their perspective. Then, analysts can layer their opinions on top of the visualization to address any business issues the data may be revealing.
Getting inside the heads of stakeholders to form insights and develop visualizations is a challenge that is often compounded by major misconceptions about data analysis. In Aaron’s words:
“The perception is that data can be some sort of silver bullet to solve problems. You see this manifested in ads for business intelligence software exclaiming that you can get all the insights you need with the push of a button or in job descriptions looking for data officers to uncover groundbreaking insights to revolutionize the business.
Sounds great, but it’s not realistic, and it does the industry a disservice. The reality is that analysis and using data in effective ways isn’t easy. It takes time, people, resources, investment…it’s not something that happens overnight.”
To maximize return on investment for data analysis, the first step for stakeholders is being realistic about what data can do. Step 2, Aaron says, is asking questions of analysts: “It’s the job of the analyst not just to create these visualizations and present them, but also to help stakeholders understand where the data comes from and what it might mean.”
As a stakeholder, learning more about the data gathering and analysis process provides the context that leads to actually using the data to solve problems. Aaron suggests asking the following questions:
- What’s happening? (where did the data come from, how was it collected, which tools were used?)
- What does it mean? (why is it significant, what could be done with this information?)
- What do I do next? (what’s the next action step to take as a result of the data?)
As for getting meaningful takeaways from data that will be actionable for business, collaboration is key and stakeholders should expect to participate in the process of discovering insights. As Aaron says:
“At the end of the day, we’re the experts in the tools and technologies and data, but we’re not the experts in their business. That’s where the teamwork and collaboration comes in. This is where the stakeholder has to tell us their pain points, problems they’d love to solve, what would add value, what are their goals for this quarter and this year, and then we can figure out how to solve those problems with the data.”
Two-way communication helps avoid the trap of being distracted by “interesting” data and visualizations that command attention but don’t answer a business issue. This is why Hero’s Data + Insights team always works iteratively with clients to examine patterns, provide ideas and analysis, and then create strategic roadmaps to improve the customer experience together.
Want the full scoop? This and more in the full podcast:
- Learn how to train data practitioners to find insights
- Get tips for effective data visualization
- Hear an amazing story of when Aaron’s team discovered a pattern in website clicks that saved a client $5 million in manufacturing costs