What’s Bubbling Up? Quantifying What You’ve Heard Qualitatively

moving at speed of biz car_featured

The following is an excerpt from Moving at the Speed of Business, a new report that explores a range of methods that market researchers can use to quickly and simply quantify information to build confidence and incite action across the company.

Being able to sort people, objects, concepts, and experiences into categories is an essential cognitive task. We couldn’t understand the world around us without it. When our objective is to understand not how we see the world, but how somebody else does, however, our own frameworks and preconceived notions get in the way.

Whether engaging in passive social listening or active questioning, how do we suspend our assumptions in order to let new, unexpected information emerge? Text analytics is one of the best examples of how to dynamically explore and synthesize vast amounts of open-ended content. Even “minor” themes, which might normally be considered outliers if they were discovered from a handful of observations or focus groups, can be easily quantified and visualized. Additionally, most software programs allow analysts to move in between capturing themes that organically “bubble up” from the data and applying a pre-determined lexicon to the data. For large data sets, in particular, computer-aided exploration and categorization is light years faster than hand coding.

For example, we conducted a project with a large grocery retailer to understand how shoppers’ emotions were connected to their stores and their competition’s. We imported about 10,000 verbatims from member-generated surveys, discussion boards, brainstorms, etc. in several of the client’s private communities into a text analytics tool called Luminoso. What we saw wasn’t just a word cloud. Luminoso noted closely associated topics through both physical proximity and color. So in this case, click on “pumpkin” and not surprisingly, “Halloween” appears in the brightest text, and in close proximity to it. Other related terms were shown in orange text, and by clicking on any word, we could drill down to the verbatim in a highly structured and focused way.

However, even without our intervention, a really interesting pattern emerged. A cluster of food-related terms appeared in one quadrant of the screen, and another cluster of shopping-related terms — “coupons” and “bags” and “shop” and “checkout” — in another quadrant. And the telling thing is that, even before we began looking for emotional language like “happy” or “delicious” or “excited,” what became immediately apparent is that when people were talking about our client’s brands — Stores 3 and 4 — it was in very process-oriented, transactional terms. They were talking about coupons and bags. But when they talked about food — the product itself — it was in relation to our client’s competitors — Stores 1 and 2.

Being able to dynamically explore months and months of open-ended comments, and then see the surprising connections in a visually compelling format, piqued our (and our client’s) curiosity. It was instantly engaging. The fact that we could “play” with the words highlighted in the visualization, dragging them around and seeing what other connections we could find, helped us validate our findings as we were exploring the data. The connections we made felt immediately tangible and “real.”

This synthesized analytic process – where exploration and validation are woven together and seamlessly executed – saves time and instills confidence in the themes as they arise.


Leave a Reply

Your email address will not be published. Required fields are marked *