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, and Pandora. Remember to subscribe, and please leave a positive review if you like what you hear!
Janavi Kumar is a Digital Senior Sensory Scientist at General Mills, where she focuses on experimentation with digital tools and new technologies to enhance the way we collect and utilize sensory and consumer data for decision making.
Prior to her current role, Janavi served as a Sensory Scientist at General Mills in strategic product and consumer research related to innovation, renovation, and quality. She has led sensory research for a variety of different product platforms and initiatives in her role with the Product Guidance and Insights group.
Transcript (Semi-automated, forgive typos!)
John: So, Janavi, thanks a lot for being on the show today.
Janavi: Thank you for having me, John. Happy to be here.
John: Oh, great. Well, something that I think is really interesting about your kind of I guess your current role is that the best of my knowledge you are that I have seen so far the first and maybe the only digital sensory scientist at a major CPG company. So can you talk to us about what does that role mean and then talk a little bit of how is it that you came into this role?
Janavi: Yeah, absolutely. I can share a little bit about that and then, you know how I came to be on sort of this digital sensory scientist role. So I'm originally from India and I moved to Iowa State University for college and my undergrad at the time, my focus was actually genetics. And I do clearly remember when I came across sensory science as a field in the food science, it having a really big impact on me. And now I do hear this from a lot of people, that sensory science is something that you kind of stumble upon. You know, you don't go into college thinking, hey, I want to be a sensory scientist. And yeah, for me, it was just kind of the cross disciplinary nature of the field that was just so interesting and compelling to me. I took on my grad school. I kind of just did a hard 180 and decided to go all in. In grad school as I moved to Kansas State to do my master's in food science with an emphasis on sensory and consumer behavior. So I supported research in the Sensory Analysis Center for a year, which is a really great experience. And then in my second year, I worked on a USDA grant related to developing strategies to prevent obesity among adolescents in low income communities.
John: Were you in Dolores' lab?
Janavi: I was for the first year. Yes.
John: Okay.
Janavi: So that was a really great experience to work under the chambers lab and instruction of first year and then the second year I moved to the Department of Human Nutrition. And so that's where, it was kind of more on the behavioral science kind of thing. So really great experienced overall. And in terms of my work experience since then, since grad school, I've spent the last six years as a product guidance and insight scientists at General Mills. You know working on different product categories and brands and very recently stepped into the newly created role of digital sensory scientists, which I'm just so excited and passionate about. And I'm happy to share more about it. We chatted a little bit about it, but you know, thinking about how it came to be prior to the role. I was just very generally outward focused and very interested in this topic of how digital technologies may impact the work of a sensory scientist. And it's very stemmed from sort of a recognition and probably more personal urgency that a feeling that my role, just like many others, might be disrupted, right?
John: Right. Yes.
Janavi: How do we as sensory scientists kind of up our game in the digital and tech space, right? And I remember having those early conversations with you as well last year as you were starting up Aigora, right? And it was just really great to connect the dots with thought leaders like yourself in the space. And internally at General Mills. I was really fortunate that the organization was really forward thinking about certain business roles that would be needed to kind of advance our capabilities in this area. So it was just a confluence of factors. We talk internally in General Mills about like the three overlapping circles of organizational needs, skill and passion, and those kind of aligned fortuitously for, you know to love for the creation of this role and then for me to step into it. So very grateful to be kind of in this role and in the space.
John: Yeah. It's really, I think very impressive interview. I mean you, you were one of the first people to reach out when I started Aigora. Like when my wife and I founded Aigora. Actually it's kind of interesting, you have a background in genetics because my wife is a professor of behavioral genetics and it does seem to be culturally a very I mean, it happens too often to be a coincidence. I oftentimes, by the people coming from genetics background, come into data science. And actually, interestingly, I think while of course, there are statistical approaches to analyzing genetics, you know kind of genomic data. It seems like some of the predictive tools and some of these computational tools that we've been talking about are I think better suited to larger data sets or complex data sets where there might be a lot of interaction. So, it would interesting for me to hear a little bit more about, you know, when you were doing your undergraduate in genetics and then you took the sensory course, what were the things like when you were studying genetics? What is it that you were interested in? And then what was it about sensory that kind of caught your eye and got you to kind of move from what I would see as kind of already a data science perspective into a sensory perspective and now coming back to this kind of data science perspective again. So can we kind of go back to that moment in time? I'd like to hear more of that first sensory class and what was really appealed to you about it
Janavi: Yeah, man, that's like going back way back. I would say that for me, I just had an inherent kind of interest in technology and science. Right? Which kind of keep me going into genetics in general as a topic. And I will say probably in undergrad, we didn't touch upon some of those really advanced data science measures. But I'm sure I believe that it might have influenced my thinking at the time. For me, it was really the confluence of all of those different or just the way sensory science touched on so many different disciplines. Right? Think about the psychographics aspect of sensory, just the physiological, the chemistry, the food science, and just so many different aspects of it that really appeal to me. And I loved all the different tangents that people could get into once they get into sensory science. It's just a cross disciplinary field. And I do think that there's a lot of stats in sensory as well. And I do feel like that kind of that overarching sort of concept of the field really did appeal to me. So I wouldn't say we've got into very technical data science principles in my undergrad. From what I recall, but I certainly think it did influence my choices in life, if you know what I mean. Yeah.
John: Definitely. That is a culture. I mean, you've got really kind of computational tools that are going to be a computational way of thinking about things in genetics. I'm interested in science and technology like you said then through sensory, I mean, I love sensory. To me, sensory is the science of the experience of life.
Janavi: I love that. Yes, totally is.
John: That's why you have all these different disciplines coming together, right? I mean, my background, I mean, I have a double major in math and philosophy that was my undergraduate.
Janavi: And I did not know that. I think I knew about math, but not philosophy.
John: Philosophy as well. Yeah. And then, of course, I've always been around a lot of psychologists. My postdoc was in a psychology department. I married a psychologist who is now behavioral geneticist. It's definitely all these ideas coming together. Okay, so let's talk a little bit then about how you are, you know, what are the things that are interesting to you in terms of the kind of new technologies that you're thinking about? You know, the technologies you think are going to be important over the next few years when it comes to understanding the science.
Janavi: Absolutely. You know, I'm happy to share a little bit more about the role of a digital sensory scientist.
John: Oh, yes. Please do.
Janavi: Kind of the technologies that we will be looking at. So when we talk about digital sensory scientists, like it's a dedicated role within General Mills, so embedded within our sensory organization to kind of do the upfront exploration and vetting of sort of new digital tools and technologies in sensory and consumer learning. So when you think about data capture, visualization methods as well as measurement, so it really does though come with a strong focus on business value creation and the overall process.
John: Right.
Janavi: What is that process to pilot tight end to business use cases and really integrate these into large, complex organizations? So I think a small sample of this is some of the smart speaker work that we worked on together, right?
John: Yes.
Janavi: Technologies to pilots in our organization to see how they can uniquely help us answer specific business questions, right?
John: Right.
Janavi: And it's also about really evaluating. So just because you can do something right, should you and how is it different and better than today's capability? What does it really unlock? And sometimes we may not be at a right time to invest in it. So thinking holistically about how those tools add value and what stage they are in. So that's one part of my role. And the other is kind of developing, partnering to develop our internal data strategy. So that's including the leveraging connected data sources, automated reporting tools, and certainly a partnership with other technical partners in our organization. And, you know, that's more than anybody in terms of the challenges we see. It is very specific to the FMCG space. But documentation really hasn't been our strong suit. Our data is just subject to a lot of variance from test to test. And sometimes standards are not consistent over time. So we know the ability to like mine and reuse that data sort of diminishes when we don't really have great documentation or standards. So that's really important for us internally and definitely all sensory scientists and organizations to pay attention to. And then the third piece is really about change management so that when you think about new things in general, right? New tools, new applications, new processes like introducing anything new comes at a cost. And so there's going to be a time, personal investment to scale up and learn things and learn things. And while learning and advancing new capabilities is really important, like being really realistic in the cost of doing so. Righ? So there's that kind of the the few areas we focusing on. And then to your earlier question on what new technologies. Just like maybe taking a step back and highlighting some of the market shifts that we're seeing. And this might be specific to CPG or FMCG. In the past decade, you've seen this. It's just so important in terms of the company's role, innovation, right? The role that innovation plays in that company's growth and the ability to sort of succeed long term. And the pace sort of at which this needs to happen is just an unprecedented need for experimentation, testing a new product development. And that does impact our world. Right? The world of sensory scientists in the organization like this. More experiments, shorter timelines, and we get pushed for product insights much, much faster. And this is where I think automation comes in, right? For expediting things or processes that might have been more manual in the past. And one great example is something you've looked at and many companies looking at is automating analysis, report writing, right? We're just very standard things that we do. And the value proposition is that the sensory scientist is freed up to work on the most value added things, activities, custom insights, storytelling, kind of to influence decision making. Right? Which is the real kind of outcome there. And then touch a little bit about this to data management getting better at data collection and storage.
John: Right. And I would say that's related to that reducing the product life cycle or testing life cycle.
Janavi: Absolutely. Yeah.
John: That you want. I mean, the thing is, all this stuff really works together. If your data are well organized and you have automated reporting, you have the ability to pull essentially to pull new combinations of data from your historical data and run simulated experiments. And if you have good predictive modeling, you can reduce it. For example, if you're thinking about, okay, there's 10 things we could do. If you based on your historical data, have trained to predictive model to predict the outcome, what you can do is you can get the probability of success for 10 things and focus on just the two or three things that are most likely to succeed rather than having to, you know, keep going back. You can reduce the number of iterations, right? When it comes to product optimization, how often are we in the situation where we have an idea what we change we want to make. But there's a difference between, like the ideal sensory profile that we're trying to make and like, actually making the product that delivers that sensory experience. Right? And so sometimes it takes multiple attempts to land where you're trying to land. And if you can reduce the number of attempts you have to make because you can do more simulations, you can make more predictions, you know, I think that there's definitely a business value in that.
Janavi: Yeah, and that's absolutely true and it starts with the data, right?
John: Right. Yeah.
Janavi: Like, are you in a place where you can mind your data is in the right structure? Do you have the right processes to set that up? Right? So I think absolutely, I think that's like the ultimate goal, right? Are we able to reuse our information, can we can we develop smarter experiments? Can we develop more informed experiments? I think that's right. We definitely want to go. And, you know, especially in an organization that's so focused on new product innovation and we are running a lot of new experiments. Right? So generating so much data and moving so fast, we still need to be paying attention to documentation, data standards management, because ultimately the reuse of that information will provide so much value. Right? For using predictive models, for more informed experiments. So I do think that’s something that we are starting to think about a little bit more in the sensory community and super important. So I love that you added that as well. And then I think one thing, you touched on this, too, when you talk to me, but, thinking about the outcomes, right? So it's an evergreen topic, I think, but getting to closer to consumer truths, right? And sensory and consumer science methods have been evolving more in the direction of more like behavior based learning. And in recognition of the fact that they're definitely better predictors of purchase than liking. And kind of this need to or this, yeah, a recognition to take into account consumers decision making processes in the design of sensory and consumer experiments, right studies. And then there's two wins like you could take the new measurements or even how we kind of adapt what's out there. Right? And in to research to just maybe encourage more intuitive responses in our usual methodology. So like system one approaches, for example, like thinking fast and slow from Daniel Kahneman talks about the use of how do you trigger more intuitive responses with the use of heuristic approaches. So it has been changing the wording of a question like how much do you like this product to how successful do you think this product would be when launched. Right? Might trigger a different more intuitive response. So there's that angle of it. And then there's also sort of then your measurement technologies that are being used. Right. So. Implicit testing, the response latency area, facial coding, all methods that aim to get us closer to sort of those intuitive system, one response.
John: Eye tracking.
Janavi: Eye tracking, yes. So that's a whole other you know space. I think we are starting to look at more in the sensory community for sure and continue to look at that.
John: Yeah, that's interesting, you know, because it is when you're really speaking my language here, because I see that, like for me, the four areas are automatic processes, managing your data, advanced analytics, like better especially predictive modeling, but also better search techniques optimization, that kind of thing, and new data sources. And the new the new data sources could be, like you said, coming from better questionnaire design, right? Asking questions that actually measure things that are going to be more predictive of the things you care about, like repeat purchase behavior. And I do think that we as a field really do need to take a hard look at liking, because I think there are much richer measurements we could be making. Right? Well, that Danielle Van Hout who was on AigoraCast, you might know. She was at Unilever for 25 years. I believe she was the head of sensory for Unilever when she was there. She has championed a method called the degree of satisfaction difference, which is a kind of satisfaction based measure. And their research indicates that the satisfaction is really a better thing to be measuring than liking. That when people are, that really repeat purchase behavior to a large extent is driven by people, a product meeting or exceeding expectation that's really an important thing to measure. So I think there are we definitely should be looking at the questions we're asking. And then we should also be looking at what are the new ways to measure things like you mentioned eye tracking or smart speakers, which I think you have there, I think smart speakers have a great future for particular applications.
Janavi: Yeah. You know, I was going to say that too when I think the other kind of big area that sensory scientist maybe should be paying attention to is I mean, already happening here, you' know taking context into consideration a lot more. Right? Thinking about, yeah, we often talk about context effects and bias in sensory research, which is a very real thing. But the paradigm shift is really the validity of your research out in the wild. Right? And we see that huge shift from central location testing methods to more in context, more at home and covid-19 has definitely accelerated this trend, right? More and more in home testing and certainly there's Web based and mobile data collection that has really helped speed up the insights in this setting. But what are the ways we can enhance our measurement capabilities in this space, right? As you mentioned, voice surveys, you know get us kind of a more nuanced view about what might be relevant insights to capture in the moment of a consumer's experience with the product versus traditional retrospective data collection. So definitely, as you mentioned, an emerging space in like an application to a specific sort of business question. So, yeah, that is, I think, absolutely something that is interesting for sure. There's also, like I know we've talked about this, but discussion of like stimulated and immersive environments. And I do think that although it has its limitations, it's definitely I think when you think about immersive, certainly has a lot of relevance in creating those brand experiences right? And even how sensory cues fit into those experiences one way to to approach that problem. So, yeah, I just think that generally in context, as a space, as an area where innovation is a real thing and sensory scientists are definitely looking into that space now.
John: Yeah, I would add to that. Something come to my attention recently, sonic environments. I think there's a lot of emphasis on visual immersive environments and not enough attention paid to kind of sonic immersive environments, because sound actually is, you know, that you never turn off your sense of sound. Did you know that? Even when you're sleeping, you're hearing things. And so it's this kind of always on system, it's in the background we tend to take for granted, but it has a huge effect on our mood. And also, interestingly, there's been some recent on AigoraCast, for example, Steve Keller, who's the head of Sonic Strategy for Pandora and they have research that they've conducted with Charles Spense at Oxford showing that they can manipulate the taste of something by changing the sounds that you're hearing while you're tasting it, right? And with the wearables and the 5G coming, I can definitely see a time when we like in the near future, have speakers, small speakers in our ears all the time. And then we will be living in an augmented environment. And it could be, you know, open, you know, General Mills bar in the future that a little sound will start to play that will make the bar taste better. You know? And the thing is like that's something, there'll be something specifically designed, a sonic environment that you're creating for your consumers that exactly complements the like taste experience or the texture experience, you know. Like I think you're going to have this cross modal stuff is going to become more and more important, you know, as wearables become a bigger deal. So I think it's super exciting time.
Janavi: I love that. I love the connection. I mean, just in terms of connecting those two brand experiences. Right? What's the connection to like maybe a brand on General Mills and the optimal texture of the product and how to create that multisensory experience to really wild the consumer, so I really do love that.
John: Yeah. And I think that's really an interesting thing that's coming. You know, there's I mean, I do want to get, I know you have some really interesting ideas on coding, and I don't want, I do want to mention one thing, which is today I found out and we're still trying to figure out how to use this, but there's a new capability when it comes to certain Amazon devices where you can have your skills track movement. And we're trying to figure out how use that.
Janavi: Is that come through with the email like to Alexa developers?
John: Yes.
Janavi: Someone did send that to me, too. So I was like, oh, this opens up some new opportunities. So that's great. Yeah, isn't that fascinating?
John: No. And you know, that there is emotion recognition is coming there as well. You can already manipulate the way that the Amazon, but I won't say her name because it sets off everyone here, but the Amazon device that you can already control the voice inflection, the emotion, the way that you deliver the survey. You will in the future be able to and I believe Amazon can already do this. I still think it's generally have emotion coding where you'll know not just what the person says, but the tone of voice in which they say it. So if someone, you know, says, do you like it? And they're like, yes versus just do you like it? Yes. There's going to be a difference there. And of course, how quickly do they say, yes, that's another thing. So I think there's a lot of like metadata.
Janavi: I think that integration with emotion coding would be just amazing and just even thinking about we do I mean, the integration of Alexa with some of the video devices as well. Right? So that, again, I think there's more opportunities there to collect data in different ways. And ccould you integrate the emotion technology to the video feedback, too. Right? And so you get this whole other layer of information. Yeah, I think that is just so cool and a huge space and something that we should all be looking at. So, yeah, thanks for bringing it up because I did see that email come through to like this is cool.
John: It's cool. There's all this stuff like I think you're also going to have these interactive consumer experiences that combined with scientific research, where you were going to have a situation where you've got some you know, maybe you're cooking something and you're taking the consumer through the process of cooking. It's an interactive experience in the kitchen that has value anyway. Right? But along the way, you're asking questions, you know. But those questions are scientifically designed. It's not just the random question. A few years ago, Apothic Winery came out with a chatbot that would walk you through a wine tasting and ask you questions along the way. But it seemed like it was kind of incidental. I think that what we need to take to the next level, we need it to be those questions are like designed with the same care that a normal consumer research survey would design its questions. So, you know, I'm really excited about all this stuff. I think that the future is extremely bright for using these new technologies.
Janavi: Agreed. Yeah.
John: So, okay, before we run out of time, because I can talk to you for hours, I mean you and I have talked quite a bit about is whether or not it's important for sensory scientists to learn to code. And I actually think your answer on this topic is very interesting. So it would be good, I think, for our listeners to hear your thoughts on this.
Janavi: Yeah, we have chatted quite a bit about this. I will say honestly, you know, I felt a lot of pressure to learn how to code or at least need to code in order to be kind of digitally savvy and proficient in this space. And, you know, there are a lot of resources out there for someone who wants to learn. If they're passionate, they recognize that everyone's strong suit. And it does give people anxiety. Give me anxiety. I do believe that, you know, engineering and programming are very important skills. But to me, like, it's only in the right context. Right? And for the type of person who is willing to put in sort of the necessary blood, sweat and tears to succeed.
John: A lot of tears.
Janavi: Yeah a lot of tears. You know, are you going to be coding for a large portion of your job to justify the massive upskilling? I think that's one way to look at it. Another way to look at it is, coding is not the goal, right? It's a tool for solving problems. Is there a different means to the outcome? Right? Maybe I need to be able to run pieces of code or scripts. Understand my inputs and my outputs. Right? Maybe then be able to know enough to adapt that code and it's less about writing it from scratch. Right. So we're really the subject matter experts, means we need to be able to communicate our needs, our questions, and then maybe it's just about understanding who to reach out to. Resourcefulness, organization. How can you translate your needs to someone who might be able to help you,right? So I think that's just my perspective overall, having thought about this for quite a while.
John: Yeah, well, I think what you really have, my eyes to on that topic is that it isn't a yes-no situation, it isn't binary, right? You know, it's not the code or don't code is a continuum. And and I think there's a kind of you might you might call a managerial level of awareness of coding where you have the ability to take a section of code, read the comments and skim it and have an idea what it does and be able to run it without necessarily. That's not the same thing as being the person who writes the code, right? I think you can understand code at a high level. And that's a good skill because you should be able to run it and you should be able to make small changes to it at the top. Like, you know, like if you're going to really integrate with your data science team. they give you script and they say, look, when you get a new file, change the file name here and it should just run and contact us if there's problems. Right? And then you can run it and you have like some level of capability there. Yeah, proficiency, which is not the same thing as being able to write it. And I think that was really actually did point because that's its ability to interact with code without necessarily that it's different from like having no words, but it's also not having to go all the way. There's value in the middle, right? Yeah. And I think that was really a good way to think about it. One other thing I would say on that point is I think it's important to know it's possible to have an idea of what you can ask from the data science team, like, what can they do? You know what I mean? Like an understanding, for example, looping the idea that as soon as you can specify something, it can be put in a loop and it can be done many times. And understanding coding will help you to avoid repetition. And then you can get your data science team.
Janavi: Yeah going back to my ninth grade computer science class, looping, for loops.
John: Okay, so we are out of time here, Janavi, and it's been a pleasure having you on here. I would like to just kind of get your quick advice, if you have just one thing to say to someone who just finished up the degree in sensory science, what advice would you give to that person?
Janavi: Yeah, I'd say this thing probably embrace a learning mindset. Right? There's just so much change coming our way. And I think our ability to pick up a new skill set, adapt a new way of learning and just adapt to sort of our evolving environment, I think is what's going to determine our success that we should continue and engage and learn.
John: Adapt. Yeah, that's great. That's perfect. I think the show title, adapt to learning mindset or maybe a learning mindset. That's really, really good.
Janavi: Yeah. Sure.
John: Okay, and if someone wants to reach out to you, suppose they want to maybe apply for a job at General Mills or they just want to connect with you?
Janavi: Yeah, absolutely. I think LinkedIn is the best. Just my name. I think it's pretty unique, Janavi Kumar. You can just search for me and connect with me, and I just be really happy to chat with anyone who is interested in this space. So feel free to reach out.
John: Okay, I will put the link in the show notes as well.
Janavi: Perfect.
John: This has been great.
Janavi: Thanks you so much for having me, John.
John: My pleasure
Janavi: Take care.
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!