By Mercy Koehler

Researching your audience doesn’t have to be overwhelming. During the 2024 NPA Summit, industry experts shared practical, time-saving methods for understanding and engaging with your community. Journalists Patrick Boehler and Becky Pallack led a discussion that introduced lean, empathy-driven approaches to audience research, emphasizing efficiency and intentionality. Meanwhile, Columbia professor Jonathan Soma demonstrated how to responsibly leverage AI tools for organizing and analyzing data, providing practical tips for integrating technology into your research workflow. Together, these sessions create a helpful framework for quickly understanding and organizing audience research in an effective and empathetic way.

Strategies for lean research 

Before beginning audience research, we must first understand who or what it is we are researching. This task can be tiresome, especially when working with a large community. In their session, “Empathy Without Exhaustion: Lean strategies for understanding your community,” Boehler and Pallack outlined five principles of lean research that help make this process easier. Lean research is an approach that can improve the practice of data collection involving people and communities.

• Focus on critical questions only: What is it that you really need to know about your audience to understand them better?

• Embrace "good enough" data: Sometimes a simple answer from your audience is enough information to go off.

• Prioritize actionable insights: What about your audience is something that you can actually improve upon? What changes can you make to increase engagement from this audience?

• Learn and iterate rapidly:  Repeat and improve upon your research methodology as you understand your audience more.

• Make research a team habit: Everything is easier and faster in a team environment, splitting research tasks with your team can shorten research time exponentially, and it is important to get insights on data from someone other than yourself.

Using these principles can give you an outline of how to prioritize your audience research in order to make the process a little less daunting. Boehler and Pallack also created three lean strategies for understanding your audience in the best and most efficient way possible. Those strategies are:

• Empathy Interviews: A well-known research method, empathy interviews are a quick and humanistic way to understand a member of your community. Pallack pointed out that these interviews can be done anywhere, and do not have to be in a professional setting. 

• Lightweight Personas: What preexisting information do you already know or have about this audience that can help you help them?

• Jobs to Be Done (JTBD): This is a helpful way to outline a person’s specific needs and understand what type of problems they have.

With the proper tips, becoming an empathetic researcher can be a quicker task than it may seem, but it is important that the proper methodology is made, so that we can improve the experience of our audience. As Boehler and Pallack said, “Research can only be lean if it is intentional.” It is important to understand what you will do before you try to do it because, as they said, exhaustion = lack of clarity. Boehler and Pallack finished their masterclass with some final key advice.

• Focus ruthlessly on what you do not know

• Start tiny and build momentum

• Embrace and acknowledge imperfect data

• Make research collaborative and habitual

Using AI for audience research, responsibly

Now that we know how to conduct lean and empathetic research, how can we use artificial intelligence technology to organize and understand that research? In his session “How to use AI tools to streamline and accelerate audience research,” self-titled “puzzle-piecer” and professor at Columbia University, Jonathan Soma gave a few caveats on using AI for audience research, before diving into tools you can use. 

It’s imperative to eliminate the misconception that AI has human understanding, said Soma, emphasizing that AI does not understand feelings or concepts, only patterns and statistical probabilities. While AI might illuminate patterns that describe what or how our audience is feeling, we must still add an empathetic perspective to understand why. 

Also, errors and biases are expected from AI, said Soma. There is not one AI platform that will be perfect all the time, he said, it is important to understand that AI has been proven to be biased before. For example, this outline from Bloomberg shows distinct racial biases against Black individuals when Open AI was asked to analyze resumes for a financial analyst role based on a person’s name, regardless of qualifications. 

Soma advised using AI primarily for “error-resilient tasks.” For example, asking AI to complete tasks with a known output —  like changing the gender of someone throughout an AI-generated story or asking for a fix to a snippet of faulty code. 

So, how do we use AI in audience research given these caveats? Soma gave examples of AI tools that can be used to streamline tedious research tasks and extract useful data from survey answers or lengthy documents. Soma personally recommended Claude for Sheets, an AI browser extension, that you can use to analyze data in a spreadsheet within minutes. 

In one example, Soma demonstrated how to use Claude to extract names and emails from a spreadsheet of open-ended form responses. In another, he showed how Claude could analyze whether a block of text was about a specific topic, e.g. “=CLAUDEEXTRACT(“Is this piece of legislation about retirement? Answer ‘Yes’ or ‘No’...”

When done correctly, Soma’s tips allow you to add new columns with specific data points in moments. This can significantly cut down time while researching and analyzing data about your audience and aid in answering specific questions you may have about your data. Even still, Soma gave a firm reminder to always spot-check the AI outputs! 

For more information on this subject, Soma’s YouTube channel has a playlist titled, “Practical AI for Investigative Journalism,” in which he has six videos with detailed explanations of how to utilize AI for researching your audience.