My ESOMAR YES Experience Started with an Everyday Problem

Over the years, ESOMAR has become a strong supporter of young professionals and their ideas. To encourage industry involvement and sharing, the Young ESOMAR Society (or YES!) was launched and with it, the YES Award. At this year’s Congress in Berlin we were thrilled to have 5 brilliant finalists from around the globe, all as first-time speakers and most as first-time delegates – thanks again to Research Now SSI for sponsoring their attendance! They pitched their best and brightest. They did an amazing job. And one decided to write about their experience. As YES Coordinator, and organizer of the pitch competition, we are happy to introduce Selina Pietsch, from Factworks, on her experience at ESOMAR Congress 2018…

By Selina Pietsch

It’s the first day of the ESOMAR Congress 2018 in Berlin. The room is filled with over 1,000 people. Next up on the program is “Young ESOMAR Society (YES) Award finalists”. It’s our turn – five young researchers from around the world will tell their story, including myself. One after another we take the stage. We try to captivate the audience with our innovative research ideas, condensed into pitches of just 60 seconds. Next up, it’s me. I’m very excited – and nervous. This is my first time on stage. I see the timer counting down the seconds and suddenly it’s over before it even began. I did it! Endorphins rush through my body.

Believe me, what felt like the blink of an eye and eventually became a 60 second talk and 3 slides took a long time to prepare. And it all started with an idea. Let me share with you where my idea came from and how it developed until it was ready for the ESOMAR stage.

Observe & Identify a Pain Point

In May 2017, I started working at Factworks; an international market research and analytics company. One of my first tasks there was coding open-ended answers. This meant I read and categorized online survey responses. I quickly understood how powerful open-ended responses were. Getting people’s thoughts without forcing them to choose from fixed response options is essential to getting the full picture in research. However, I also noticed rather quickly a big drawback of using open-ended questions: analyzing them is an extremely time-consuming business.

Find Something You Identify With

The first thing that came to my mind was: isn’t there a way to make coding open-ended responses more efficient? I knew that “Text Analytics” was gaining more attention. Different text analysis solutions had been tried out at Factworks. However, none had been convincing enough to replace humans.

Fortunately, during my studies I specialized in quantitative marketing and information systems. The latter turned out to be very useful when I started looking for ways to better automate open-end coding. I knew that a lot of other industries already benefited from the various techniques in this field. I thought: shouldn’t there be a better way to use Text Analytics for coding open-ends?

To understand how to make coding open-ends more efficient, I had to detect current weaknesses. One of them is that open ends are very different from typical Text Analytics use cases as open ends are short and sweet, just like posts on social media. There it was – the foundation for my idea was laid. I started doing a lot of research into different Text Analytics techniques in other areas and industries that I’d learned about during my studies.

One technique was Topic Modeling. This technique discovers topics in text documents without any prior labeling. Different Topic Models have been developed for specific use cases and text characteristics. Some have even been optimized for shorter social media posts, like survey open-ends. I really believed I’d found something meaningful here. In my eyes, Topic Modeling had the potential to revolutionize open-end coding.

I obviously wanted to confirm this intuition using data. So I undertook a comparative study taking Topic Modeling algorithms from different domains and applying them to survey data.

Find Something Relevant

I was able to show that Text Analytics and especially Topic Modeling had enormous potential for market research and for the analysis of open-ended survey responses. Further, the models more specifically designed for social media outperformed the traditional LDA (Latent Dirichlet Allocation), often considered a state-of-the-art Topic Model. These findings were very valuable in informing the way we look at open-ended data. They also fitted well with other evolving areas of Text Analytics at Factworks. It also saved time compared to manual coding and removed human bias – algorithms are more objective than humans.

I wanted to share this great news! This is where the ESOMAR YES competition came in. A stage for young researchers with innovative ideas seemed like the perfect platform to reach the right audience and spread the word.

A Once-in-a-Lifetime Experience

My ESOMAR Congress 2018 was a fantastic opportunity and an unforgettable experience. Not only because I won the YES award, but because of all the people I got to know during the event and their reaction to my pitch. The feedback showed just how many people were able to relate to my topic and felt it was relevant for their work too. This was how I knew that all the work I’d put into this had absolutely been worth it!

All the ideas presented during this competition were unique and deserved to be heard. Each of us had a different motivation behind wanting to present our idea and could probably tell a different story as to how and why the idea came up.

For me, there were 3 main motivating factors that proved to me that I had made the right choice:

  1. The idea I focused on was a real pain point – a daily struggle that many can empathize with
  2. It was a topic I was passionate about and could personally identify with
  3. It was relevant and ultimately translated into a real value-add and efficiency gain – not only for me and my company, but also for a wider audience

By Selina Pietsch, Factworks

Sound like fun? Feel like being a part of the community? Or know someone who would? Be sure to check out the YES website and get yourself and/or colleagues 35 and under involved! Speaking, competing, programme committees, becoming published…it’s all possible with YES!