Welcome to "AigoraCast", conversations with industry experts on how new technologies are transforming sensory and consumer science!
AigoraCast is available on Apple Podcasts, Stitcher, Google Podcasts, Spotify, PodCast Republic, Pandora, and Amazon Music. Remember to subscribe, and please leave a positive review if you like what you hear!
Karen Graves is the Senior Director of Sensory & Consumer Sciences at Bell Flavors & Fragrances. She is passionate about integrating sensory properties and consumer needs to guide and validate creative development and innovation.
Prior to joining Bell Flavors & Fragrances, Karen spent more than a dozen years at Kraft Foods as a Sensory Manager. At Kraft, Karen managed the Sensory Evaluation Center and led Sensory and Consumer Research for Philadelphia, DiGiorno, Jell-O, Velveeta, and Kraft Macaroni and Cheese.
Transcript (Semi-automated, forgive typos!)
John: So, Karen, thanks a lot for being on the show today.
Karen: Yeah, thanks, John. Thanks for inviting me today.
John: It's great. So we're talking about a bunch of things here before the show, and I think the thing that maybe was most interesting to me was your thoughts on us getting started with AI and sensory. Okay, I know that this is something in AI, I mean, you can talk about how it's been going on for various forms for two hundred years. But it's become really a hot topic of the last decade, you know, throughout our culture. And something, of course, I'm very passionate about, something that we've been getting increasingly, I mean, that's the whole point of Aigora is to help sensory and consumer scientists prepare for AI. But you've been on your own journey. And I think it would be interesting to hear your thoughts on, you know, what do you think sensory scientists should be keeping in mind as they begin their own AI journeys?
Karen: Right. And that's one thing to me. I really like to keep up to date on the newest technology and how we can really use it and leverage it in sensory and consumer research to make our jobs easier, right? Either to make our jobs easier or to get closer to the consumer to have better guidance for the projects. And, you know, when I was on LinkedIn and started to see the Aigora posts on LinkedIn about AI and what they could do, I started to learn more about it. So I know, you know, you and I had some conversations over the phone and then I joined the AigoraPlus network where then I could see some of this in action and it was really impressive. One thing that I thought was really impressive from the start was the data mining going into the historical data and able to sort and categorize a ton of data really in a quick manner, and then that you could take that and really use have that historical data be useful for you as a sensory scientist and your project teams.
John: Yes. Yeah, I definitely agree with that, because if you think about it, I mean, pretty much any large company has huge amounts of historical data. I'm sure at Bell you all have voluminous quantities. And it is interesting to me, looking back to my early years at Institute for Perception running projects, how, you know, other than the knowledge it's already in the heads of the experts, it would seem like each project would start off with almost a blank slate in terms of the information from the past that was being used. So how do you see that changing going into the future as we start to use our historical data a little more proactively?
Karen: So I think setting up the infrastructure first is the critical part. Still that you're organizing and establishing a framework on how you want your data organized and captured. So I think that would be the first step. And then from they're making sure you have the right programming and coding to get what you need to help in thinking about how you're going to use that data. What's the measurement or the criteria that you're looking for?
John: Right. And speaking of coding, actually, I know you have some interesting thoughts on coding. We were talking a little bit about, you know, black boxes versus the need to see what, you know, what's actually going on inside analysis. Maybe we can hear a little bit of your thoughts on coding within sensory science. Do you think it's necessary that sensory scientist need to learn to code or what are your thoughts on, how comfortable do we need to be?
Karen: I think with that regard, there's a couple levels of I see, right? Like you could have a person who needs to know like that basic understanding of the coding and that can kind of look from it from afar and understand like what's the general gist of what's happening with that code sequence? But then you may have a sensory scientist who really is interested and really gets into that type of data and they are more of like a sensory data programmer when they take some of the aspects of the data science and apply it to sensory. So, there could be a couple of ways. But I think in general, like your average sensory scientist, just needs to have that basic understanding. And for somebody like me, I learned MS-DOS programming, you know, growing up in elementary school. So I feel like that's some of the coding I see in R brings me back to sitting at the old giant cardboard box, refrigerator box computer doing the MS-DOS. And then also you think of SAS programming too, so.
John: Right. Yeah, that's right. I think that when you've got these kind of interpreted or, you know, scripting languages where essentially they're just instructions, it's not like Fortran or you have to write some code and then you compile it. Instead you have something being run line by line. I definitely agree with you that it's good. And I would encourage really all sensory scientists to try to at least get some basic idea of what are the things that can happen in code, right? Like if then, okay, we've got to some choices happening here or here's a loop. Something's happening several times, you know, and then what's the part where the data are being read? What's the part where the analysis is happening? What's the part where, you know, results are being reported out? I do think it's really important that you should be able to read code at that level. And then, you know, if you want to go deeper, you can, but, once upon a time, I felt like we all need to learn to code. I don't really feel that way. I liked coding, but yeah.
Karen: And I would say to you can encourage some experimentation. I think you and I did that over a Zoom call once, where we had an immediate need with Covid at Bell Flavors & Fragrances, we shifted from a triangle test to Tetrad. And because Tetrad more powerful, you need less people and with less people in the office, we're doing Tetrad. And I remember we had a Zoom call you were showing me in R, tetrad and there was several functions within R and a couple ways you could go. So I think to me that's also like don't be really scared to experiment and see the different ways you could take that coding, because within Tetrad we ultimately are using the confidence interval. But there are other ways that you could take that Tetrad coding in R.
John: Yes, definitely. And just as we talk about R I do think that people are curious about doing sensory in R. It would be good for people to know, the sensory package by Rune Kristiansen is got really nice out of the box tools that you can use if you've got 16 out of 31 correct in a tetrad test. You want to know, is that significant what's the effect size, in a few lines of code you can have the answer to that question. So I do think it's a lot of that stuff is pretty is pretty plug and play compared to what it used to be, say, 10 or 15 years ago. So something that I really enjoyed about our kind of precall were your thoughts on kind of being courageous, because I think it's easy to get intimidated by all this technology that's coming. And so what are some of your ideas about, you know, as sensory professionals, how can we start to use some of these tools and maybe move forward even when, I think it is normal to feel a little bit of intimidation? So what are your thoughts on how to handle that?
Karen: Yeah, absolutely. And I think your sensory scientists, we all feel comfortable with a certain set of tools or a certain set of data analysis tools, right? Not only from a methodology approach, but from a data analysis. And so that's where, you know, learning about these alternative tools and the AI tools certainly just I would say, don't be intimidated, go for it and start small. That would be my first suggestion, because in our case, you know, you don't have to start off. It's not this whole big programming code. You can just take a simple code and then you can create a template for that. So then you can use that for the future. So it's not, I know I was personally intimidated because then when I started looking at the R software and senseR, you know, you find books, you find YouTube tutorials. There's a ton of information out there and I think connecting with the right experts helped me. So that's why I like our conversations were extremely beneficial because it helped me kind of take this like just plethora of just information overload about a software and really help it make it be relevant to what we do here at Bell Flavors & Fragrances.
John: Right. Now, that's a really good answer, I would say that increasingly that's what I find, that instead of trying to albeit the expression swallow the elephant, I think that maybe that's the or eat the whole elephant or whatever. And how do you an elephant, one bite at a time. Yeah, that's what we need is to have a small problem that we care about, like a simple question, like what's the P value for a difference test. Can we get that code running? And then once that's running, maybe you can build on make a little table, but to have specific things you're trying to do and build it out very, very valuable.
Karen: And then like what we did in our end too is then so we built the code along with you for the Dee-Prime confidence interval for Tetrad. But then we are also looking at other ways of analyzing the data. You know, we have some internal calculators or through the data analysis on our data collection software, and then we can kind of compare and contrast a bit. Right? So because as scientists, we want to see what that data output looks like and feel confident with the one that we're choosing moving forward. And I think as sensory and consumer scientists, we need to go through that exercise to just, you know just so that we truly understand the nature of that output.
John: Yeah, I definitely agree with that as a scientist for sure. And so I think that, you know, kind of related to that maybe. What are your thoughts on, we're talking a little bit about black boxes, you know, kind of before this call. You know, on the one hand, I think that you have to, you can't learn everything, right? But on the other hand, you do have to have some idea of what's happening in an analysis or in a program. So what are some of the things you look for, at what point do you start to feel comfortable that you can trust something that might have been built by somebody else?
Karen: Yeah, that's a good question. I don't need to see, like, the formal equations for me personally. I need to understand, are they operating off a binomial model? A beta binomial model? A thurstonian model? Those to me are the frameworks on how something is being analyzed kind of in that box or how that decisions being made. So to me and I know if you talk to my colleagues here at Bell Flavors and Fragrances, I'm known for drawing the two standard distributions for that thurstonian model. So everyone I've done some internal training. So everyone understands that Dee-prime metric and the visual of the thurstone because that's a very important part of how our data is analyzed.
John: Yeah, absolutely. So at a theoretical level, you need to understand what's going on as far as the nuts and bolts. I think that's kind of a reasonable approach for sure. Okay, so now we're kind of looking into the future a little bit here. Well, actually, I do want to talk to you a little bit about how you adjusted to Covid, because that is something that you mentioned, you switched from primal to Tetrad, what are some of the other adjustments you had to make? And what are the changes that you see yourself keeping even when hopefully the Covid crisis is resolved to some extent in the future.
Karen: It's been an exciting journey here at Bell Flavors & Fragrances. And we've had a team committed to keeping the work and we were not shut down at all. So we just kept going so even as the schools were closing, the sensory department was deemed essential function. We did scale back, instead of doing 8 to 10 panels a week, we did scale back to more or less like two to four, especially during that March, April time frame, because we had panelists coming in at different shifts. So we adjust the calendar in a bit and we adjust where the panel was being conducted. At first we were doing it a in a more controlled way with it was more near our sensory desk area and we were paper ballots. Nothing was shared. We are wiping everything down. People had to bring their own pens. We weren't using the booths and we're still currently not using the booths. And you have to learn and adapt. And that's what we're all doing right now during Covid and after about a couple about four to six weeks of that approach, we realized it wasn't working because we started to get congregation of people. So less social distancing. So what we are doing now, we've been doing basically since May we hand deliver samples to people's office areas that are in odor free areas of the building so that they can do them in their own safe environment. Right? Because we all feel comfortable coming into our office space or our desk space as each individual can clean their area so that they feel comfortable with that sanitary conditions and the handwashing and all of that. And then they end their data on their own computer instead of a shared computer now.
John: Okay, I see. That's interesting. So you adjust the methodology to increase the power and then you made a really a lot of physical adjustments. It's interesting that it was a lot of the methodology had to be adapted physically because of Covid.
Karen: Right. And then our next step as we're planning ahead, because Covid is going to be around for a while. Once more people get into the office on a regular basis and we'll have more of our full panel. Then we'll move into panels in a larger conference room so that we can accommodate maybe eight people at once and then accommodate social distancing that way.
John: Right. And hopefully we'll see some medical advances that start to at least make it, maybe less serious of a problem. We'll see what happens I guess. It's hard to predict the future, as Yogi Berra said that you can predict anything except the future, something like that. So that's good. Alright, so what are the things, you know, as you look forward to the future in sensory, what are the things you're most excited about? The kinds of changes or new technologies, things that you think are offering the most promise over the next couple of years here?
Karen: Well, I think it's, to me, the exciting part is just sensory having that increase seat at the table, driving decisions, influencing project teams. I think we're getting even better at that right? Throughout the years, we as a sensory as a whole, these are things that I think we're all getting better at. And that's to me what I'm really excited about, like not only at conferences, not only we talking about advancing methods, but we're advancing our influencing in our negotiation skills because that's a huge part of what we do. And then as well as writing reports that are meaningful and tell that you know give that data storytelling influence are our key stakeholders. So that's one thing. And then, of course, you know, it will be interesting to see technology developing and adapting with, you know, more and more people, you know, being glued to not only just their phone, but multiple devices. You know, you've got people not only watching TV, Netflix or whatever, but they're also multitasking on their phone to an extent. And seeing how those behaviors are changing and seeing how we can get consumer data that way. You know, I find that whole consumer research using Alexa, it's really, really neat and I taking that further and really getting into consumers’ homes without using a lot of people resources, because traditionally you have a whole team doing an ethnography. And so here they can use technology, you know, to replace people. And it is becoming more of a standard, right? You're just you're having more Zoom's and you're having more interactions with technology. So people are more and more comfortable with it.
John: Yes, definitely. And I would say what's interesting about a lot of this is while it is true, we won't have a set like, right now, you know, it's a really resource intensive if you're going to send interviewers to all of these people's homes. Right? And now I think in the very near future, in fact, you mentioned the Amazon devices, I won't say her name because I don't set them off everywhere. But the, you know, the Amazon smart speakers or Google increasingly allows us to have a presence in people's homes. In fact, I don't know if you saw it, but Amazon just announced that the new echo shows are going to have motion tracking capabilities. So we're, you know, having some big launch that we're going to be attending to try to learn about that. So if you have any thoughts about how we can might use motion tracking in people's kitchens, it seems like there should be something we should be able to do with that, it's, interesting.
Karen: Yeah, I agree to me right now, routines and change in routine is a huge opportunity area for people to figure out and to learn more about. And that could be part of it, because that's a huge part of our routines as motion and how we move around the house. And, you know, like that act of sanitizing your hand as how much is that happening at home would be interesting.
John: That is interesting. Well, you know, we’ll have to look into this, a motion activated survey would be quite interesting because you could at certain times a day notice that someone is moving in their kitchen and say, excuse me, do you mind, whether they've already registered, they're in the survey, right? They're not gonna be surprised if it's some random person. You know, on the panel, like, for example, with P&K, we've got the panel. And so then the person is enrolled in the survey and they come into the kitchen and preparing their dinner. And then the device says, hello, I notice that you're here. Do you mind if you do the survey now? Right? That's sort of thing. It's very exciting. Okay, so that's great. What are the other things you kind of keeping your eye or the new methodologies? Like what are the kind of big picture questions that you all are thinking about there at Bell Flavors and Fragrances, you know, as you're kind of moving forward?
Karen: You know, I think, you know, big picture, you know, always keeping up with the trends and understanding what's that new thing that new flavor of fragrance that's appealing and also how we can deliver an excellent sensory experience against that trend.
John: Right. Now, that's very interesting. And there are some very good tools now. You know, Sai Kodur from Spoonshot was just on AigoraCast. They have a great tool for, they monitor I believe is twenty three thousand different sources of data, things like and not just I mean there's obvious sources like Instagram and Facebook and stuff, but interesting sources like food tech prospectuses or food science journal or abstracts, you know, this kind of thing to try to get an idea of what's coming. So I think that's one idea that I think is really promising for using new technologies to track trends. But as far as actually delivering those experiences, do you see yourself kind of embracing machine learning more? Trying to, you have some target and then you're going to try to use machine learning for the formulation side? Or do you see it more in terms of trying to understand the consumer experience? Like, you know, what are you more interested in? The formulation side of things or the consumer experience side of things?
Karen: For me, I'm more interested in the consumer experience side of things. And then also like how does the fragrance and flavor like what's that whole affective in flavor experience that's happening. So and how can we capture that in a Bell Flavor & Fragrances to have, you know, to mimic that in somebody's home.
John: I see. So you have some sensory experience that you're trying to create and then the question is how do we produce that?
Karen: Yeah.
John: Yeah. And that's really, really fascinating. Okay, well, this has been great. So, Karen, we are actually almost out of time, so we'll need to wrap it up here in a little bit. But I would like to ask your advice, because as someone who has been, I think in a lot of really dynamic situations, I know when you were you were at Kraft, you all were doing a lot of exciting research. What advice do you have for someone who's kind of just coming into sensory now and what do you think that they should be focusing on, you know, as they begin their careers?
Karen: So my advice for a young professional is really to learn like the multi aspect of sensory science, right? So you can be great at the methods. Right? You can really know your methods in the data analysis, but also understand how the other disciplines fit into sensory and how your decisions impact your projects. And so that's where I would encourage folks, you know, even things like this, like the AI in data science. Right? Like that's a tool that you can use to make their job more seamless, but then also understand the business side and how sensory impacts the top line and bottom line growth of your projects and of your company so that you can always bring it back to some kind of value. How your time and how your panels impact the financial side because we learn a lot about science. We learn a lot about food science. But I would encourage people to take something maybe that's a little bit out of your comfort zone, like shadow a salesperson or shadow a marketing or even a finance person, and just learn in general what they do and what things they worry and care about so that when you make your recommendations and revise your reports, then you can target your key stakeholders.
John: Yes. I think that’s extremely important. And I do see that as something we've got a lot better at in sensory is understanding, how what we, not just conducting sensory as a kind of academic exercise, but understanding that we are part of a business. Business has goals, and maybe some of that matches up with the design based thinking, right? That one of the general principles of design is that for something to be valuable, it has to be useful. So it will be the same, you know, it's interesting to see those ideas come into our field. Okay, Karen, this has been great. If someone wants to get in touch with you? What would be a good way for them to reach out?
Karen: So the best way is to find me on LinkedIn. So feel free to send me a message or reach out.
John: Okay, that sounds great. Any last comments, bit of thoughts?
Karen: No, just thanks for having me, John. This has been great.
John: Oh, thanks a lot, Karen. This has been wonderful. Thank you
Karen: You're welcome.
John: Okay, that's it. Hope you enjoyed this conversation. If you did, please help us grow our audience by telling your friend about AigoraCast and leaving us a positive review on iTunes. Thanks
That's it for now. If you'd like to receive email updates from Aigora, including weekly video recaps of our blog activity, click on the button below to join our email list. Thanks for stopping by!