• John Ennis

Betina Piqueras-Fiszman - No Questions Wasted


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, and PodCast Republic. Remember to subscribe, and please leave a positive review if you like what you hear!


Dr. Betina Piqueras-Fiszman is Associate Professor at Wageningen University. Her research focuses on the exploration of novel ways to understand food perception, preferences, and motivations. Betina is particularly interested in the connection between how the mind works at an implicit level and what we observe explicitly. She uses a multidisciplinary approach to investigate the link between consumers’ associations, perception, and motivation, and teaches on these topics. She has published around 50 scientific peer-reviewed papers and book chapters and actively contributes to various international academic symposia, courses, and other public events. She is co-author of the book “The Perfect Meal” (which won the 2015 Prose Award) and co-edited the book “Multisensory Flavor Perception”. She also collaborates frequently with industry in a wide range of private and European projects. Betina won the 2017 Food Quality and Preference Young Researcher award and currently sits on the editorial boards of Food Quality and Preference and Appetite.


LinkedIn

University Page




Transcript (Semi-automated, forgive typos!)



John: So, Betina, thanks a lot for being on the show today.


Betina: Hi, John. Thanks to you.

John: Okay. This is actually the first AigoraCast episode since the whole coronavirus outbreak has happened. This first one that we've recorded. So I think we really interesting to start off with your thoughts on how sensory is going to adapt, and I know no one really has the answer to this, but it would be good to hear your thoughts. How the sensory is going to adapt to a world in which it might be challenging in the near future to get groups of, say, 10- 20-30 people to get. So much of what we do is often dependent on getting a group of people together in a room. And I think that's going to be something we are going to be able to do for the near future. So, if we could just start kind of start with your thoughts on this topic. How do you see sensory adapting to the short term reality?


Betina: Yeah, I think as any other field, we should adapt depending on many factors, economic factors, but in this case, it's pandemic, so, of course, we cannot conduct research or people cannot collect data from people coming together to a central location, at least in Europe that's the situation for most of the companies doing such research. So alternatives have to be found. So that's relying mainly on at home use testing or, you know, people collecting the samples from somewhere or the sample is being delivered. Sometimes it could be focus groups in which many people from their homes, you know, discuss these samples via online, or it could be completely online surveys in which people could see the images if visual samples are present. So that's something that we are seeing currently happening in Europe. So, yeah, I think that, another identity views also to, for larger projects we have perhaps is still rely on already collected data from people and then tried to predict consumers responses in the near future based on the data already collected without having to collect more data.


John: Yeah, this is interesting. A bunch of things I have to say in response to that. So, okay let's talk about historical data because this is obviously a topic I'm really interested in, I mean, I'm sure you're aware of my enthusiasm for graph databases. Get your data all organized, connect it, try to get essentially run simulated experiments where you are pulling data from your database, maybe across many studies. However, one concern I have is I wonder if the data that was collected before the pandemic may be different in some qualitative way from the way that people are, I mean, I guess, what are your thoughts on this? I don't know if anyone knows the answer, but people going to be affected by the pandemic in some fundamental way that changes their, you know, the sorts of things they're interested in? How do you see that playing out?


Betina: Yeah. In terms of the sensory data, I think that's a little bit less variable, although we're talking about the people who are suffering from the virus. Then from what we all heard, this sensory is a bit impaired but in terms of how they think about products. We might be facing the same situation as opposed or a situation in which people, you know when they go again to buy some products, they don't rely so much on more superficial factors, but more on needed factor. So in that sense, in terms of what we might observe at the supermarkets, that might have a bigger influence. But in terms of sensory data per say, I think that we can rely so far depends on the projects, on the data that we already have collected. And they should be able to predict or to fill in some of the gaps.


John: Yeah, I know that is really good point. I mean, there's this kind of range, too, you know, I mean, I'm expose with sensory you know we are talking about this before the call. There's this, you know, sensory consumer science is a little bit of an ill-defined term because not clear where the boundary is. And as you get more and more into the kind of consumer science, then that's where you might start to see some of these effects where, like right now, like I'll take any toilet paper I could find. Now it's like, okay, if I could find a box of toilet paper on Amazon, I would order it.


Betina: We're adapting, I mean the people we are seeing in the supermarkets now, we were inspired to watch and to share these experiences because how people behave, for instance, in the Netherlands, it's very different in terms of, you know, hoarding for some food. It's very different from the behaviors that I observe here. In terms of what shelves are being emptied or what brands are being, you know, consumed, so yeah, people have very different behaviors in these things compared to the regular days, like hoarding of course to their needs.


John: That's fascinating because actually maybe the data itself about what is getting hoarded by country is actually very valuable for understanding cross-cultural differences. That some research is in order. Like if you can figure this out, you know, you can figure out, okay, what are some important differences between countries? And then that information itself can be really valuable as the economies come back after the pandemic is over.


Betina: Yeah, definitely. I read a post and he was sharing that in Italy, the macaroni, you know, the great one was I don't know if he was lifting this shelves or nobody, it was a completely different behavior from that is normally observed in normal circumstances. So the great one that absorb more sauce and all these things that is usually very popular there during these days was left on the shelves. Yeah. So people are shifting a little bit of they're behaviors. And I have no clue whether those are more expensive than the other one. But that's very interesting to compare across culturally.


John: Yeah. I saw a picture of a United States supermarket where the meat case was totally cleared out except for the plant based stuff. The plant based stuff was still there. All the real meat had been bought.


Betina: Yeah. To be honest, it's difficult to be able I mean, people say, well, what we're observing right now really reflect the preferences of people. But of course, there's many other factors. I mean, if that is ten times more expensive than usual meat, then people would still not buy it even in these circumstances or perhaps even more under these circumstances. It's then it really reflects how complex people are. It's not only about preferences. It's about many other things. So many times using sensory and consumer science to really predict how people are going to behave in a supermarket is very difficult. And we can see that even nowadays.


John: Yeah, that's really interesting. Well, let's just takes back to your kind of research project in general, which I really admire, which is looking at many sources of data to try to get maybe better predictive models for human behavior. So maybe it'll be interesting to hear you talk, what do you think are kind of the key signals? I mean, maybe independent of the whole crisis, but just in general, what if you found any research has the signals that are most predictive of people's behavior maybe most informative. You know as you think about all the different ways you can collect data? What you find it to be the most helpful?


Betina: I think that's a question I think a lot about also because I have students and students ask me a lot of questions, too. It's really difficult to answer because, you know, we're normally looking at very, very narrow area. So all of us are in little bubbles or research questions that we want to answer. So it really depends on what exact research question you want to answer. But what I have come across is that many researchers, when there's new technologies or new ways of collecting responses, it being responses or you know using virtual reality. You know, there's not a best method or a best, you know, response that is better predictive. I mean, it would help predict things in different ways. So as long as we are aware of that, I think that there is no best method. And then we also see it now e in the conferences when people are actually comparing non-technology based methods. Each of them provide different insights. So I think that it's all a matter of combining and knowing a little bit the background of how consumers behave. And for me, most may be circumstances where we want to gather as much information or insights as possible from consumers. Sometimes it's better not to use as many technologies as possible or as many methodologies as possible, but sometimes less is more. Also, talking about these post-war situations. We still able to collect data but in a more minimalistic way. And sometimes we can do that if we know how to, so not because those methods are out there, all of them, we have to use them all. It's about optimizing.


John: And what are some of the criteria that you use when you're trying to figure out which methods to use? What are some of the factors that guide your decision making?


Betina: So I guess that's a little bit of the benefit of being in the academic world, that we can explore a bit more the fundamentals. So if a research question is about, you know, knowing how the body reacts to certain stimulations, then we know what kind of tools are there to keep us exactly that answer. It's a matter of attention. We would know what to, you know, which technologies to use in a project. But many times I have the feeling that we collect a lot of data and then you know, imagine attracting people in a supermarket, knowing where they looked at or what they picked, what they read. At some point, we also why are you wearing EEG helmets or why their heart rate is being tracked. So at the end it seems so relevant to have all that information together. And how to put it together is the question that many researchers are facing. So it is very difficult to answer that question of which one is the best or which one is more predictive. I mean, I think we should slice it at some point and trying to model everything together when there are so many things being correlated. It's a bit very, very tricky. Yeah.


John: And I think it's kind of challenge for our age now. We collect data in many ways and there's even more ways coming online like, actually I'm starting to do an Alexa based surveys, which is really interesting to collect data in the moment. Handsfree, you know, someone brushing their teeth, shaving, shampooing or cleaning something, you know. You can ask a few key questions right at the moment of evaluation, right? Someone drink something, you don't have to put down the beverage to answer the question.


Betina: Yeah and a home testing is fascinating because you always know that at home there will be some, you know, things that are not completely under control. But then there's the trade of you know in a natural environment, for me, the key thing about all these field is to be able to collect data in a way that is most natural for consumers. So perhaps we are very much relying on scales, okay from one to nine, how intensive is this but perhaps we should just step back and think, how do consumers think about this in their everyday life? You know, perhaps ask some key questions and then it's comes back to the less is more. Not all the consumers have to get all the same questions. And although what we're seeing about personalized advertisements and, you know, what we see in Google or in social media are tailored according to what we have seen before or what we decided to see, what we decided not to see. So perhaps sensory and consumer science should move a little bit in that direction. So that's why I was saying that I did before that I don't see the field moving so fast, you know, so it is good and it's bad in some way because we could harness this development in technologies and use it for our benefit. We do not have to ask all the questions to everybody. So personalization would be also positive and I'm just wearing these green hat right now, so I would have no clue how data would look like or how to process it. But it would be an idea on how the world is moving and how we adapt to the world.


John: Right. Well, I think the pure market research community has been in that realm for a while with adaptive choice based conjoint among other technologies, right? You don't necessarily need to ask, you know, once you can get an idea. Yeah someone doesn't like something that's in this area. We don't need to keep asking questions about that area. We can focus on the areas where they seem to have more interest. And I also see that with chatbots coming in and this is another advantage. Right now, the technology with Alexa is not ready for in-depth open ends, but it's getting better all the time. Eventually, what will happen is you'll have questions and then dynamically the survey will figure out what questions to ask next as a function of the question to the answers, right? I mean, right now you could just say, okay, if someone gives a high value for sweetness, then you make sure you ask an open end which is we noticed that you gave a high value sweetness. Can you tell us more about why, you know, you can dig deeper into things that are interesting. Yeah. So that's the idea of personalized surveys. That's very interesting concept.


Betina: Yeah. That data comes back to the notion of how technology can help us is by going deep, digging deeper into what we already have and not necessarily collecting more and more information because at the end it will be stored somewhere. And of course, we can find patterns within the data that we have already observed. But yeah, sometimes what we already have, it's really good enough and also good in terms of recruiting people because as we were discussing earlier, we would already be able to detect some clusters or consumers that are of specific interest. And then not everybody has to be targeted or, you know, we can personalized surveys depending on which clusters those people are. So just sweetness or those that you know specific fatty diet or I don't know.


John: Well, this is really good, Betina. This goes back to like, why I do these podcasts at all, because these big ideas that come out of these conversations really helped me. This is a big idea that you have a lot of historical data in a product category, you could look at the data and figure out, okay, what questions do we, I think it is a big, the way people usually ask these, you know, when they will look at historical data, they often ask questions like, okay, which questions weren't informative, which questions, you know, basically column by column, which variables were correlated with other variables. But there's something deeper here, which is that you don't have to take out the whole column. You may be only maybe it's the case that some people need to get that question and some people don't, right? And so you could do something and it's a lot more precise in your analysis, which is try to look at your historical data and ask, okay, what subset of questions could we ask? But now it isn't just column by column. Now it's like person, you're looking at the interaction between the people and the questions and you're saying, okay, what did we not need to ask? And you could get this kind of personalized survey like, that's very interesting to look at your historical data and suppose we asked a big subset of these questions, but it was not just taking a whole question, but you take out for some people the question could be about the same information, right? Okay, I will work on that. That's really interesting.


Betina: Yeah. But also, you know, some chunks of the question could also be completely irrelevant for some parts of the population. So you have an historical data you can really tailor the structure of the questionnaire.


John: Right. And then you could design your new surveys going forward so that you could put the questions at a diagnostic early on, right? And so as much as you can, you don't, I mean, obviously, there's a natural flow to evaluating a product, but because I mean, you don't ask people that. Anyway, there's an order in which it makes sense to ask you a question, but you can still front load the questionnaire with the questions that are going to help you to move people through a personalized survey more effectively. Which reduces the survey length. It gives you more like basically you're increasing. I mean, the I guess the goal of all this is to increase the signal to noise ratio. You're trying to get as much information as possible.


Betina: Yeah. And also the consumers are feeling that it's more personal to them and more relevant is like, you know, going into Google or social media and finding all these ads that are not relevant for you. It's also fatiguing to see all these information and all these questions you have to answer, which are of no relevance to you. So then that would also make them feel, oh okay, these people, you know, know what I like and they're asking you a relevant questions that would make products better for me. I'm not for, okay, you know, the entire population.


John: Right. So you getting a better engagement.


Betina: Yeah it's a win-win situation.


John: Definitely. Now, sometimes you need to know this is kind of interesting that sometimes you actually need to drill into the reasons people don't like things also, right? That it's interesting that like the least useful information is in the middle, that like if someone like something, you're gonna dig into that. But if you also really don't like it, you begin to understand why, because sometimes that tells you, you know, I guess this area of max diff research, you know, if he does ask you what's your favorite? Like, which of these four items would you pick? Then you don't get nearly as good information as if you ask which would be your first choice, which would be your last choice that you kind of need that range of information.


Betina: Yeah, what you're saying right now about why people don't like things. I mean, I think that if we think about the process of perception that we're most likely they stored in our brain. And as long as something comes across our daily routine that does not match what we were expecting, then we did take the difference. Otherwise, we would not notice. And that's also link to this advertising that we find something that bothers us, we would notice and then we would do something about it. But we could use these predictive models to than design questions in which people would pay attention to all of the items and not have indeed noise in the questionnaire, so to speak, or items that are not really relevant for them. That would make the data that they provide much more valuable and not many attributes, because we all know that consumers, you know, if they have to prove they know quite quite long questionnaire and there's many attributes that they, you know, a bit irrelevant for them they would just say whatever. And then that makes our data not a good quality. But, yeah, a bit less analytical methods would be good. Going back to the point of making methods that are more natural for people about that, I also observed in class, for instance, my students had to do a optimization of some samples of some orange juice and they had to make themselves with some people and make them taste their innovative orange juices and then what I saw was that clearly people drink much less from the sample that they don't like, but some samples are barely tasted. So that is an indication of, you know, which samples are more successful than others without having to ask any questions, you know. But sometimes less is more. We get much more insights from just seeing how people behave towards samples as food and then adapting our methods to capture that. The thing where the use of technology or a thing with the use of more quality methods, which is rather fun nowadays.


John: Right. Well, another way technology helps us is it allows us to collect more of this kind of you might call exhaust data or metadata. You know, where you have, there is information but it wasn't being captured in the survey like the information, like if you have sensors on the glasses that are somehow measuring how much is drunk? So you can capture that information easily. Yes, that's good. And of course, reaction time. I really think reaction time is not paid enough attention to how quickly someone gives you a response is also informative and that is captured if you wanted to be captured by your computer.


Betina: Yeah, that's true. And also linking this, there's a lot of papers comparing implicit explicit methods. And of course, the conclusion is both of them provide useful information. But I think the battles will things back to the fact of if we have knowledge on how perception works and how our brain works. We know that too, because system one, system two but when we're shopping. Of course, system one is the first one reactants and system two is, you know, comes in later to taking one system one hasn't.


John: To make up the story sometimes to defend the decision.


Betina: Yeah, that's right. But then at the very end, how we shop and how we make decisions are most of them involving the two. So it's not that we buy impulsively. We do everything by impulse. You know, implicit data reaction tends to give us some information that help, for instance, help prone our people to follow their, you know impulses. But then that does not mean that they do not reflect. And then do not buy it. So then we really need the two types of information and then we can get categories to use it to get to categorize people in terms of, you know, people might be impulsive, but then that does not mean that they buy everything by impulse. So there's different ways with the methods we have to dig in further and then answer questions that are more helping us understand how people behave.


John: That very interesting. You know, my post-doctoral research was along those lines because when people categorize items like suppose you're learning to play tennis and you're figuring out when to hit a forehand, when to hit a backhand. You're learning these categories, right? And, of course, you know, categorizations are so much of what we do in life in general. Yeah. So you have a rule based system, which is you learn, okay, if the ball is coming over at this angle and I'm positioned here like you have rules that guide the decision and then you also have just procedure of learning, which is more involve in line with system one. But the two systems do work together and actually sometimes they compete, but eventually they work, the two of them combined to give to guide behavior. So I think sometimes is, yeah, there is almost an overemphasis on system one nowadays where there's this idea that people have like they have no self reflection. They just go in and they animalistic do whatever they feel like doing. Yeah.


Betina: Yeah. That's very funny because actually the verb hoarding which is this kind of impulsive buying and panic whatever. It's called in the Netherland in Dutch hamstring. It calls hamstring and there's all these explanation of what they're giving in terms of why the term is going to hamster. But that's very funny. But you mention it because that's exactly what animals do. They cannot reflect that much.


John: Right. Like back to buying all the meat, when I said the meat and not the plant based meat. That people are in line like now a moment of panic and they're like buying the thing that they're comfortable with that they know that maybe at some level they feel like, okay, this is the thing we you know, we've always eaten. Yeah. That's maybe more of a system one reaction.


Betina: Yeah. Habitual foods. I mean, and this goes back to the question of what dependent variables are out there that help us predict things. Well, we already know which products consumers have a habit of buying. Then we know that those products will likely be chosen regarding the liking of these people. So then we do not need to collect data from other products or put these people in other habitual choice paradigms, you know, we already have that information from them because we might have their purchase data or they might have declared it at some other point. So we have this historical data of some people. We know that they will not change to Brand B if they have brand A or, you know, several decades. So then these people, you know, to address these people than others should be used or perhaps these people are not really worth addressing or, you know, but 21 days to create a happy bases. But yeah, that would help in targeting, you know, strategies and, you know, when should make the most of this product or data.


John: Yeah. That's fascinating. Okay, Betina, I could talk to you all day but actually we're pretty much out of time. So with the time we have left, I'll just ask the parting question that I always ask, which is the advice for the young sensory scientist. Someone who just graduated with a degree of food science or sensory science, or they coming into the field from, you know, psychology or whatever. What advice do you have to people over the next two years?


Betina: Yeah, I would give them the same advice as I give my own students on that one that I would have liked to help myself is that just be curious. Of course, that's not something that can be learned. But I think that if you're curious, you're going to, you know, open your boundaries, because what you will experience, where these students will experience in their everyday work will be much broader than whatever niche they studied at school. And by that, I'm not only talking about, you know, knowledge, but also self skills, too. It's not only about studying and getting all A's, or whatever. It's all about communication, reading all about other domain. So just marketing, psychology, business. They really need to prepare to be able to shift quickly. So my advice is really to, you know, broaden your domains and be curious and talk to many people like you.


John: Yeah, well, that's good.


Betina: That helps a lot.


John: Yeah. I mean, honestly, just networking is a good way to talk to people, finding out what they're doing. I find that to be credible


Betina: Yeah. Sometimes it's more what happens, you know, after hours in the conference and not so much during the sessions. It's as important I think.


John: I would say that is one of my concerns. If we go into this world where we're not getting together for a while, that you know, we do need to have those conversations. I think about the relationships I've made over the years, like for example meeting you and Thierry in conferences. Another person who I have a really good relationship with, Michelle Niedziela, I don't know if you know Michelle, you'd really like her if you haven't met her yet. I'll introduce you if we ever all in the same place again. But she's neuroscientist to using sensory who does a lot of these implicit measures. We met because we had posters next to each other and so we start talking, you know, like that sort of stuff I think is really important. And I hope it survives into this new post-war period, as you describe it.


Betina: Yeah. The best ideas come up from these types of moments so far. Cherish with them and make the most of them.


John: The idea of the personalized survey and no question wasted, I think that's and like the idea that you can use your historical data to design personalized service in the future, that's really a good idea. So I work on that .So okay, just wrap things up here Betina, how can people get in touch with you? What's the best way to connect with you if they, you know heard something interesting. They want to talk to you. Maybe they want to be a student in your lab. How should they find you?


Betina: Yeah sure in the Wageningen University website and otherwise also through LinkedIn, Twitter, I should check more often, but otherwise, yeah, I guess there's plenty of ways.


John: Okay so we'll put the links in the show notes for the how to find you in Wageningen online and then your LinkedIn. And then people can just send you a message. Okay. Well this has been great. Thank you so much, Betina.


Betina: Thank you, John. I hope to see you soon then.


John: Hopefully Eurosense. Hopefully that happens.


Betina: Yeah, absolutely. See you soon. Bye.


John: Okay. That's it. Hope you enjoyed this conversation. If you did, please help us grow our audience by telling a 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!



©2020 Aigora