by In the lead up to our Qual360 2024 tour, we are re-living the best of 2023. One of the highest rated presentations by our audience was Natasha Hritzuk from Warner Bros Discovery talking about her recent segmentation study, combining qual and quant.
Building knowledge triangulations by bridging qual and data analytics: A Warner Bros Discovery segmentation case study
- Bringing together behavioral information with primary data and qual
- Recognizing and leveraging the compatibility of qual and analytics
- Developing a collaborative environment: Facilitating cooperation between data analytics and consumer insights
Natasha Hritzuk, VP Consumer Insights , Warner Bros Discovery
You can find all Qual360 conference recordings for free in our video section!
Video transcript below (This is an AI generated summary. There may be inaccuracies.)
Building knowledge triangulations by bridging qual and data analytics
Natash Hritzuk presents a case study of a challenge faced while working at Warner Brothers. The challenge was related to segmentation of their consumer base, which was fragmented and not tied to first-party data. So it was important to implement a nationally representative segmentation approach, which involved fusing the segmentation onto first-party data and Nielsen and augmenting it using qualitative research. This also included creating bridge points between the two data sets, which were missing in their initial data. The benefits of triangulation are highlighted, for example greater accuracy and more reliable decision-making. businesses are encouraged to adopt this approach to better understand their customers and markets.
Warner Brothers has multiple media platforms, including CNN and streaming services, and that the segmentation of their consumer base was fragmented and not tied to first party data. The speaker explains that each segmentation was based on either Nielsen viewing data or first party data, but these segments were not similar to human segmentations and were not tied back to the first party data. Many of these segments were very thin and did not reflect human subjects.
It was critical to discover a segmentation that represented the US viewing audience, rather than just the Warner Brothers Discovery audience. The task was to create a segmentation that was nationally representative, capturing 360-degree views such as psychographics around viewing behaviors, triggers, needs, and goals. The speaker highlights the importance of these psychographics in understanding the consumers and their viewing behaviors, which is crucial for creating a fully fleshed view of who these people are.
The research team employed a little bit of a set of barriers and obstacles in working with data science, as they had limited knowledge of the other person’s understanding of the task. However, the team transformed the title segmentation into a consumer segmentation approach, which involved fusing the segmentation onto the first party data and Nielsen, and then augmenting them using qualitative research. They employed workshops in the final stage, where the consumers’ viewing behavior was exposed to pick 500 titles using Nielsen. The upshot of this was to see how the consumers responded to the exposure, and how the information gathered could be used to flag content that might not be apparent in their historical viewing behavior.
Natasha highlights the problem of trying to understand the coverage of a set of 12 entertainment segments. The team spent about two months trying to land on a model, but ultimately decided to move on. She mentions that the executive team saw segmentation as a core priority for the year, and that eventually they did a massive survey to determine eight segments. Initially k-means clustering was used as well as including first party data, but ultimately there was a realization that the media ecosystem was limited and that information beyond the screen was critical.
One of the main challenges was to create a complete picture from two different data sets when only a few common elements exist. Again the importance of bridge points between the two sets, which were missing in their initial data was highlighted. This led to a low match rate of only 20 matches, meaning that 80% of the predictability of the model was incorrect.
The team realized that they needed more variables and connection points beyond what was provided by first-party data. They turned to qualitative research to capture more details about the audience, including their habits, preferences, and activities beyond what they watched. They also dug deeper into the psychographics of the audience to understand their interests and motivations. This allowed them to create a broader range of variables, including technology use, media consumption, and daily routines. These new variables, combined with the existing data, resulted in a match rate of 60%, marking an acceptable level of accuracy.
Further, the presentation highlights the importance of segmentation in identifying specific groups within a company’s own ecosystem, and the role of third-party data in understanding consumer behavior and preferences. Qualitative methods are used to refine and clarify key issues that stakeholders may be facing. Using segments as a framework allows the crystallizing of marketing, media, and content strategies, suggesting that stakeholders look at specific platforms, messaging, and context in order to reach their target audience.
Building knowledge triangulations in the entertainment industry is a key success factor. Warner Brothers is using Nielsen data to segment viewers and provide a handbook of research tools for continuous analysis. Starting with qualitative research will provide a comprehensive and accurate list of variables, resulting in a more efficient workable survey and possibly a better segmentation solution.
It is important to find those points of connection in the data, similar to dating apps where more texture is needed to improve the quality of the match. Hritzuk recommends starting with qualitative research before moving on to quantitative analysis, and major stakeholders benefited from exploratory work to ensure the accuracy of inputs. The most effective activity mentioned for stakeholders is use case generation, which provides a framework for building out applications of the segmentation solution.
