Flourish in a Flash: How AI and Social Listening Can Influence Gift Cards
Updated: Feb 13, 2020
Register for The Flourish Conference today to learn more about this and other topics influencing the future of branded currency
Please note: Transcripts are computer generated.
We'd like to thank our sponsor Home Chef for sponsoring this episode of Flourish in a Flash. Do you find that a day at the office is ending later than expected? Don't have the interest or time to get lost in the grocery aisles? Have you exhausted your nearby takeout options? Why not let Home Chef show you how quick and easy home cooking can be. Home Chef offers flavorful meals anyone can cook and the recipe options never get stale. There's always something new and always something for everyone. Use Home Chef gift cards to reward your employees and customers with an easier home-cooked meal solution. Bulk discounts on gift cards are available, so visit HomeChef.com/giftcards for more information.
You're listening to the Flourish in a Flash podcast with the Flourish team, Des, Holly, Kristen, and Erica.
Welcome, everybody, to another episode of Flourish in a Flash. I am Kristen Thiry, co-founder of K+H Connection and the Flourish Conference. I am very excited this week to have two special guests joining us, Derek and Vina from Scio Motus, and we are going to be doing a study with them that we’ll be presenting at Flourish. So without further ado, I would like to introduce you to Vina Rathbone Falvey and Derek Huyser. Welcome to the podcast today, you guys.
Thanks. Thanks for having us.
Of course, thanks for joining. So maybe we could just start with a brief introduction. I know a lot of our listeners have not met you before or heard about you before, so if you could give us a brief introduction of yourselves as well as tell us a little bit more about Scio Motus.
Sure, I’m happy to. My name is Derek Huyser. I’m a partner in Scio Motus. Scio Motus is a qualitative research firm that focuses on techniques that are often used for quantitative research. We found ways to apply those to qualitative research at scale.
And I'm Vina Rathbone Falvey, another partner at Scio Motus. We have a third partner here in Boise and Derek is in Chicago.
Great. What we’re basically doing in partnering with you guys is using your research and using your methodology to do some social listening and attribute some personality or psychological attributes to those personality types. Could you explain a little bit about what is social listening and how do you go about doing that?
Sure, I’ll take this one. Social listening is a little bit different from social media monitoring or something that you would do if you were managing a brand online. Social listening is when you do a broader scrape of the internet based on a topic that is interesting or relevant to your business. That could be full industry, that could be a topic, a trend… But you look at everything that people are saying, even if they’re not talking directly about your brand. We have some big data aggregator tools and visualization tools that do this. We do this for brands who want to know about their competition, who want to know what’s changing in their industry, and what they should be excited about or worried about. It’s pretty cool to just leverage a ton of social data from the internet and see trends from it.
Yeah, and one of the things I love about what you guys are doing with social listening, especially for brands who are interested in having their profile assessed, is that none of this is secret sauce data. It’s all publicly available data—what they’re talking to their customers about and what their customers are saying about them—all publicly available on the social media platforms. So that’s one of the things I really think is so interesting about it.
To Derek’s point earlier about making it quantitative, we consider ourselves qualitative researchers because we are really interested in what people say, what they think, how they feel, what they like or dislike—those more qualitative factors. But when we use social listening, we can do that at a major quantitative scale. So if you wanted to say, “What do people think about how Burger King tastes?” you could do a query to say, “What are people saying about the way Burger King tastes?” And you could pull data from a period of time—maybe a year, maybe a month, it’s up to you. And then see that I get a quantitative scale, like a large percentage of people feel this way, a smaller percentage of people say this, these are some themes that come up... These can be really directional for a brand and a communications team.
Can you expand on that a little bit? How are brands using the social listening? And what are some examples, maybe from retailers or nonretailers, of how people are using social listening?
I can give one quick example. We had a client who was loosely in the beverage industry, and they were worried about a new trend in the market that they thought might dislodge their standing in the marketplace. We were able to use social listening tools to understand that trend, and there certainly is a positive trend that in the long term has some potential to be competition for them. But we were also able to apply some quantitative measures to it and look at the scale of that trend and the scale of the conversations. And while they were extremely positive in sentiment, there were just a really small number of them happening compared to the overall conversation about that segment of the beverage industry. So while it was something that they need to pay attention to, it was not an existential threat to their business, which was what their concern was. But the social listening allowed us to understand that in fact, while it’s a positive trend, it hasn’t reached a critical mass yet that they need to really be concerned about.
That's a great example. I think that’s really helpful to understand. One of the interesting things that you guys do beyond social listening is that you take everything that you’re hearing from the social listening and you apply AI to that to generate those personality insights. Can you explain a little bit more about that process, what you do, how you apply AI, and what the output is of that?
Yes, this is the fun stuff that we get to do. We are really interested in understanding people, and one way to understand a person ends on a bigger scale—a group of people or an audience—is through their personality. There’s some science going back to the 1940s about correlating people’s personalities at the words that they use to their language usage. Some geniuses at IBM were able to write these rules for connecting language to predict personality into a machine learning algorithm powered by IBM Watson. So we didn't do that; we just learned how to use it. It was just really cool, exciting technology. I don't know if any of you have taken a personality test, but if you have, it’s kind of fun to get some insight into how your mind works or into your preferences or why you do things. We’re able to do that now with machine learning at a bigger scale and to look at a group of people and say, what collective personality traits might an audience have in common? And how can we, knowing those personality traits, make more meaningful connections to the audience for our brand? How this ties into social listening is we start there with recruiting. Let’s say that we’re looking for a group. Derek, help me come up with an example of who we might be recruiting a group of…
Let’s do endurance athletes.
Yeah, we looked recently at endurance athletes who are interested in marathons or ultramarathons. We would start with the social listening and say, who is identifying themselves as a person that is interested in (endurance running) based on their language on social media? We write that query of people saying, “I ran a marathon” or “I want to run an ultramarathon” or “I’m here running a marathon right now.” And we find those people who are active online and we collect and anonymize. I’m very quick to say that we anonymize and protect user data. I never reveal the name that it’s associated with to pull their user text. And then, as I mentioned before, that language that people are using can help predict their personality. So we run that through Watson, which tells us in return their personality traits, and we look across those personality traits. Let’s say we recruited an audience of 30 or 50 people. We would say, what do these people have in common? And then we study more about those personality traits to see what sort of consumer behaviors might be predicted by those personality traits. I should probably mention that we use a particular personality model for this called the Big Five. The Big Five consists of five personality traits: openness, conscientiousness, extraversion, agreeableness, and neuroticism, and all five of those major characteristics have sub-traits. Let’s say that the trait of openness has other things related to it like intelligence or artistic interest or imagination. And the way that Big Five determines personality is that they rank all of those traits on a scale. So maybe you’d be very high in openness—you really like new ideas, you’re very open to change, you’re openminded. And you could also be very low in artistic interests, like if art is not really your thing and creativity is not where you spend your time or energy. So then, with all of these 30 or so traits that you look across, you rank high or low across all of them, and that gives us an idea about a person’s personality.
That’s really interesting. I was wondering one thing while you were describing how you develop a query set and you’re looking for those ultrarunners or distance runners or marathon runners, and you develop the language that you’re going to look for. Once you identify those people, do you then look at their full language of everything that they say on their profile or just what they say about that particular topic?
We can do both. If we’re doing personality insights, I'll pull everything that they say publicly. Again, I don’t take any private data like emails or any profiles that have security on them. We don't see any of that stuff. So we take that for personality insights. If I was looking at trends—let's say trends in ultrarunning—I would only pullwhat that person and many other people say about ultrarunning, for example. So we can do both, and they both have a place and a purpose.
That’s great. So you tie that to the Big Five personalities and you’re able to come up with these psychological attributes—high in extraversion, for example. I don't know all the different traits, but high and low in the different traits. What do you use those traits for? How do you apply that back to the brands and give them additional insights into their consumers?
Great question. We use a lot of academic research about consumer behavior. We weren’t the first people to think of this, and there’s been quite a bit of academic research done about this sort of thing over the past five or ten years in big universities like Cambridge and Stanford. They’re doing tons of research studies about how personality traits can predict consumer behaviors. That can be things like shopping behaviors, like people who are likely to join the loyalty program versus people who aren’t likely to be motivated by price, and how that correlates with personality (because it does). Or it can be something like if somebody is a dog person or a cat person—their personality traits—there’s academic research that connects those traits to a multitude of consumer behaviors. So we comb through that for a lot of insights, and then brands put it all together. So we say, here's a personality profile of your audience or audiences and here’s what we know from research about their consumer behaviors. And then we work with them about messages that will resonate with them, programs that could resonate with them, and maybe things to avoid, too. For example, this audience is very motivated by loyalty, not price, so you need to reach or address them or do programs that work with that.
I think it’s also worth noting that the ability to speak to people on a personal level is very innate in humans. So once the data is presented, it becomes much easier to sit down and think through someone that you know who fits that profile, and the difference in how you would speak to them and message them versus someone who’s very different. And I think that while we use a lot of big data, we use artificial intelligence in this research to come up with these personalities and to marry that with these research-based insights. There is a lot of intuition that we’re able to spark in markers when they see this information because they can relate to something that’s very innate in their nature. That brings out a lot of possibilities, both from a messaging standpoint as being a said from an activation standpoint. It really creates a wide range of options for how to leverage this information.
I think this is so exciting—this approach of really applying the social listening and the personality assessments after you’ve put it through IBM Watson and you’ve used the AI and the machine learning to do that analysis. Because really, one of the things that you guys know is such a challenge in our industry—in the gift card industry, in the branded currency industry—is that our currency is anonymous. We don’t know who those customers are, who are getting gift cards and buying gift cards and receiving gift cards. And that’s for a reason; we purposely mask that information a lot of times to avoid some escheatment regulations and some other things. So what you guys are able to do—and you’ve already started the initial phases of this for us—is take a look at those gift card recipients and see what their personality traits are based on the language that they’re using and the fact that they’ve promoted that they received the gift card and expressed thanks or gratitude or whatever it was to the person who gave them that gift card on social media. I think what’s going to be really interesting is now looking at the buyer side and seeing if we can pull some data and get the personality assessments of the buyers in the gift card industry. It’s going to give us access to a wealth of data that we’ve never had before. I think any sort of personality assessments or things that we’ve associated to gift card buyers or recipients in the past has all been self-reported, and we all know we’re not great at self-reporting. Any sort of self-reported data is still a little bit biased and still not 100% accurate when you’re trying to recall something, but being able to analyze the language that those gift card buyers and recipients have used really gives us deeper insights. Vina, you mentioned before that if you’ve ever taken a personality assessment, sometimes there are those a-ha moments, even for you personally. Like, Oh yeah, that is why I do that thing! It gives us insights that, on the surface, we would have never otherwise really known about these segments of customers.
It's often really complementary too to traditional qualitative or quantitative research. You do a survey or you do a focus group and you get some ideas about who your customers are demographically. Like, Okay, I think it's women in their 30s. I just say that because I'm a woman in my 30s. But that’s a pretty big group across, even your state… And really, is that every single woman in her 30s? Because that’s also a pretty wide range in someone’s life. And often we find that it might be women in a certain range of their lives, but really it’s certain different personality traits that are driving that. That really helps brands when they know that to be very efficient with their messaging with their targeting, so they’re not wasting marketing resources. So we like to complement additional research, too.
Absolutely. So one of the things that we’ll be doing at Flourish is partnering with a particular brand to do this kind of assessment with that brand’s customers as well. We’ll be marrying the data that we’re seeing with that brand and the personality assessment in the psychological behaviors behind that brand’s customers, as well as what we know about those gift card recipients and buyers. We’re going to marry the two together to really help hone in on what that messaging should be, what the language should be, and how we should be talking to those customers. Can you expand a little bit more on what you’re hoping to get out of this study and what the takeaways will be?
Sure. Well, we’re really excited to do this. We’re really excited for Flourish. I think the goal would be to take the brand that we’ll work with and give them a new perspective on their audiences and how they look and think about their audiences. It’s hard to say without knowing all the details yet, but whether they have a really good idea of who their audience is or whether they don’t, it’s interesting to confirm or expand upon that or change that just based on wherever we start with social listening. Yes, it is this group of people, but you also have this big group of people over here who also really love your brand and talk about it a lot. And then we’ll be looking at the personality traits of those audiences, whether they’re established or new or both, and seeing how those match with the brand, how those can evolve with the brand, how those affect or how the brand affects those. We will look at what the relationships are there and then we hope to give that brand some good insights on these audiences and their consumer behaviors driven by their personalities. These are great ways to talk to them, to reach them, to connect (audio cuts) with meaningful for this audience in a really specific way.
One of the things I’m always excited about is to see where there’s a correlation in the personalities that allow the brand to activate new customers that look very much like the existing ones, and not from a demographic perspective but from a personality standpoint. I think that personality, in our view, is very predictive of future behavior. This is why we get really excited when a brand can uncover what it was that drove those customers to them and created that loyalty in the first place, because that's such a strong indicator that they can replicate that in new customer groups. Arming the brand with that knowledge and devising strategies to help them go activate those customers around the personality instead of the demographics is something we’re really excited about.
Yeah, and really regardless of how well this brand is targeting their customers and honing in on those different brand pillars, I think this study will still shed additional light because we’ve never really done the cross-reference of a particular brand and the gift card recipients and buyers. I think that’s where we’ll certainly help this particular brand target those customers better, not just overall, but also for their gift card programs specifically. And I think that’s what we’re excited to see—those results and how they overlap or how they diverge, and how we can be better about targeting those customers. I think this is going to be so exciting and we are just so grateful to you guys for partnering with us on this study. We will be announcing our brand partner here shortly, so definitely continue to listen to the podcast and subscribe to our newsletter to get all the great Flourish updates. Derek and Vina, thank you so much for joining us today. And if you guys want to hear more about this study and get all the results and the great data that they’re going to be putting together, please join us at Flourish March 16th through 18th in Chicago. For more information, you can go to FlourishCon.com. Derek and Vina, thank you again so much for joining us today.
Great, thanks for having us.
Thank you, guys.
Flourish in a Flash is produced by K+H Connection, a branded currency consulting firm. You can learn more about K+H at KHConnection.com. And you can always find out more about Flourish and the Flourish Conference at FlourishCon.com or follow us on all of our socials. On Twitter, Facebook, and LinkedIn, it’s @FlourishCon, and on Instagram, it's @Flourish_Con.