Taking a Data-Driven Approach to Thoughts and Feelings Around Diversity and Inclusion

vector illustration of abstract headshots

By James Cuthbertson, CRO, Relative Insight

Understanding and predicting pressures surrounding diversity and inclusion (D&I) can be tricky. Businesses must proactively revamp and rethink many different practices in the face of rising concern around these issues: internal company policies; inclusive cultural initiatives; external communications, marketing and advertising; and more. But what exactly does diversity, equality and inclusion mean to differing groups of individuals? Our recent research here on Relative Insight found that there is no “one size fits all” when it comes to understanding this complex issue.

For the most part, diversity and inclusion are sensitive topics rooted in feelings, rather than hard facts, so it’s easy to miss the mark when developing initiatives and communications that address these concepts. That’s why we decided to analyze over 300,000 words of open-text survey data collected from respondents in the U.S. and U.K. to understand opinions around D&I.

Historically, open-ended answers have been seen as a mess of unstructured data that is difficult, expensive and time-consuming to process. However, by using text analytics that blends AI and natural language processing, combined with a comparison-based methodology, we were able to reveal the trends, differences, and insights latent within language.

The insights we recovered from the survey were fed by granular data from individual responses to help us understand the intricacies and subtleties intrinsic to the perspectives of various groups. What we found was, not surprisingly, that feelings and needs surrounding these issues were, well, quite diverse – further illustrating the complexity of the topic.

We presented respondents with approximately 20 questions asking them to: think about brands and organizations that they thought were doing a good job from a D&I perspective; tell us about brands that they wouldn’t engage with on principle; and give us any advice they had for these brands. We then split, segmented and compared the responses from these answers by gender, sexuality, nationality and other characteristics to understand differences in the way each segment spoke.

A common theme from survey responses across all groups was a vagueness in language from traditionally non-marginalised groups, and precise, specific answers from those in more marginalized categories.

One of the comparisons we ran was by analyzing responses by gender. Those who identified as women strongly aligned the ideas of diversity and inclusion with their own experiences. They were more likely to use the word “women” in their responses and judged brands’ inclusivity based on how those brands portrayed women in their ad campaigns. Men rarely mentioned “men”, but rather gave examples of brands doing a good job based on CSR and corporate structures, rather than from a consumer perspective.

For businesses and brands, identifying these kinds of nuances are critical in helping to unlock how the personal definitions of ‘diversity and inclusion’ differ, which can help them create policies that are more human and relevant to employees and customers.

By understanding specific opinions, organisations can better understand where change needs to take place, and determine the direction for future commitments surrounding diversity, equality and inclusion initiative and action.

About the Author

James Cuthbertson is the CRO of Relative Insight. An innovator in text analytics, James loves language and has helped grow the company from a start-up into a commercial machine. With over 12 years-experience in the early stage technology market, James has grown, and developed sales and marketing functions focused on products in audience insights and data. www.relativeinsight.com