Image courtesy: Flickr/Enokson
Market research companies are getting hip to the opportunity of using semantic technology to deliver meaningful results to clients from the surveys they conduct for them. Global research agency Millward Brown, whose clients include a large number of Fortune 500 brands ranging from CPG to services to insurance, certainly does: It’s just signed on to use OpenAmplify’s semantic platform to help with analyzing and finding strategic value from survey participant’s open-ended comments.
Bringing a semantic platform into play here came about through the efforts of the Millward Brown’s global innovation team, where Sherri Stevens is director. Stevens hopes Millward Brown is an early adopter in the space to leverage semantic technology’s possibilities, but she’d be surprised if it isn’t taken up in a big way by the research industry at large. “To be able to use technology to understand sentiments and topics and consumer intentions based on their own comments to us is really important – it’s a step forward in surveys,” she says. “Obviously, asking people direct questions has been in surveys since forever, but this takes things beyond coding to something that can help one find the meaning and quantify sentiment and all that across multiple consumers.”
The real voice of the consumer comes out in open-ended queries, vs. just having them check off answers in boxes. But having to rely solely on hard-coding or manually interpreting results lacks scalability, and adds time and costs to the process of trying to unearth the core kernels of information that’s important for researchers’ clients to know. “That’s what working with OpenAmplify is designed to help us do, because it can do that in a scalable fashion and we can combine data from multiple sources,” she says. Stevens points out that the platform’s ability not only to dive deeper into a study’s results but to pull other information together from other sources in a structured manner that can be merged into these studies adds more value to Millward Brown’s communications with clients about brand perceptions, issues, and more.
Where OpenAmplify makes a big difference, she says, is its NLP expertise that helps you learn something rather than just count the numbers of times specific words pop up. Understanding whether the subject or object of a sentence is being referred to and the sentiments consumers are expressing – and the strength of that sentiment – is, as she puts it, “cool… And it’s exciting to see that semantic analysis has come far enough in automating the ability to tease apart those kinds of [sentiment] differences that research analysts can highlight for clients’ benefits.”
None of this means the human element gets taken out of the equation, as research starts with building a report that fits into what the brand or marketing manager’s concerns are to begin with. “We’re not just take OpenAmplify’s output and popping that into a Powerpoint – we don’t do that with any of our data,” she says. “It has to have meaning and usefulness and tie back into a strategy and relevant opportunities, but using the OpenAmplify system helps us find those key verbatim, key ideas, that people want to tell us about.”
She’s looking forward to some innovations from the technology in visualization to better show the relationships between words, for instance, and she wouldn’t mind it if a semantic platform like OpenAmplify could somehow make more hay out of consumers’ shorter commentary. After all, everyone is getting used to writing in Twitter-speak these days, and it’s not surprising that consumers usually like to get through surveys fast. “The truth is that the OpenAmplify system works better with dense responses rather than two-word responses,” she says.
Millward Brown has been very pleased with the pilots it’s run so far, Stevens says, noting that the firm’s been getting a lot of interest from brand clients since announcing the relationship. “We’re really trying to do a better job of providing meaningful, forward-looking results for clients so they can manage their brand better,” she says. “We think this helps; [insight into] good open-ended answers and tracking them with other studies and making sure it’s fully integrated into other information we know, so we can find out what people really are trying to say rather than just what we want to ask.”