Interview: The Science Behind Brandwatch Search With Aykut Firat
By Phill AgnewSep 28
Combining high-quality mobile survey technology, a robust polling methodology, and expert data analysis,
our bulletins will be essential reading to get the pulse of the nation
Published August 13th 2020
MyHealthTeams creates social networks for communities of people facing chronic conditions, including autism, breast cancer, multiple sclerosis, and more.
There are 37 of these networks which span 13 countries. The aim is to connect people with similar conditions to share tips, get referrals, and, ultimately, improve health outcomes.
Beth Schneider has been in market research for over 35 years. She joined MyHealthTeams in San Francisco four years ago, where she now focuses on both survey research and social listening analysis.
The data she has access to is fascinating. Social networks from MyHealthTeams generate a wealth of data which the team can learn from, but getting to those insights was challenging.
Before using Brandwatch, Beth spent a lot of time doing manual analysis to try to identify conversation themes. To go deeper, find the unknowns, and to save valuable time, the team wanted a new technology.
“The key thing we were looking for was a new approach to easily uncover actionable insights and quantify sentiment among the thousands of conversations happening daily on our social network,” says Beth.
The team chose a solution that now lives in Brandwatch Consumer Research, known as the Data Upload API. MyHealthTeams’ networks are closed, so social listening platforms can’t analyze the data as standard – instead, the team uploaded the data they wanted to analyze. With the Data Upload API, users can import any text-based data that they have permission to analyze such as support emails, chat logs, customer feedback, and surveys. This data can be stored securely and privately in their Brandwatch account.
Beth uploaded 250,000 anonymized verbatim mentions and was able to start analyzing the data quickly, finding all sorts of insights.
“By analyzing the data that we uploaded into the platform, we were able to readily discover unexpected trends, data-backed insights, and even influencers. The visualizations are amazing. They’re easy to generate, easy to understand and, importantly, make it easy to discover new ‘a-has’ and areas for further exploration. And they are a succinct way of summarizing important trends.” – Beth Schneider, Director of Research, MyHealthTeams
The reports Beth’s team were able to generate allowed them to confirm key assumptions – they were now backed up with clear, data-backed connections.
But the team were also keen to find out what they didn’t know about the conversations going on in their networks. With so many mentions, doing this kind of research manually is incredibly difficult.
Beth loved how the platform’s AI-powered text analysis gave the team more of a ‘birds-eye view’ on all mentions, so they could find themes, conversational peaks, and areas to further explore. This allowed them to discover new topics they hadn’t anticipated and identify influential figures such as doctors who were being talked about by members of the community.
“The platform is like a beacon,” says Beth. “It doesn’t give you the answers, but indicates where a researcher should take a look.”
The ability to filter by sentiment is really important to the team and, because of the high specificity of the conversations on MyHealthTeams’ networks, it’s something that requires a flexible tool to get right. It’s not just keywords that might connote a particular sentiment, but the context in which they appear. Brandwatch’s machine learning technology was critical for segmenting the data in this way.
Beth is an experienced researcher, and spent the time ‘teaching’ the platform to recognize how mentions fell into particular sentiment categories. Once the data was trained, her team could monitor shifting trends over time.
“The tool is smart, and it made us feel smart!” Beth says. “It learned quickly and eventually we were able to automate sentiment analysis very accurately.”
Beth says there was a real sense of excitement among the team regarding the tool’s power to uncover deep and meaningful insights.
“The Data Upload API had a huge impact on our work,” she says.
Special thanks to Beth Schneider for sharing her experiences with the Brandwatch blog.
Combining high-quality mobile survey technology, a robust polling methodology, and expert data analysis, our bulletins will be essential reading to get the pulse of the nation.