Martin Kern - A Comprehensive View
Welcome to "AigoraCast", conversations with industry experts on how new technologies are transforming sensory and consumer science!
Martin Kern holds a PhD in agricultural science, is a certified enologist, and has been a Managing Director of SAM International since 2005. Today, SAM is a member of Eurofins Scientific with more than 15 of its own facilities, where it forms an International Business Line providing Sensory and Consumer Research to its customers worldwide with a global network of about 120 partners in more than 65 countries.
Martin provides valuable consulting in the Fast Moving Consumer Goods industry, Food and Non-Food, supporting product success through consumer satisfaction. Together with his team, Martin has developed new methods and approaches and contributed to the further development of Sensory and Consumer Science. Some examples of his work include the Ideal Sensory Trajectory, Context Setting Research, and Value Scoring. Martin is also a member of the management board of the European Sensory Network.
“It all started with learning how to make good wine: my professional background is in wine making and enology. Today I am passionate about professional Sensory and Consumer Research. I take great pride in being the voice of the consumer, listening to and understanding them, translating findings into “actionable intelligence” for the “Fast Moving Consumer Goods-Industry.” Martin Kern
Transcript (Semi-automated, forgive typos!)
John: Martin, thanks a lot for being on the show today.
Martin: Yeah, thank you for inviting me.
John: It's a pleasure. So, Martin, I think your background is fascinating and I think it would be a good starting point for people who haven't had a chance to meet you in person. I find you an extremely engaging person with a wide-ranging background. So maybe you could take us through your history how you got to be where you are now. Leading one of the really largest testing groups for sensory consumer science in the world.
Martin: Yeah, a pleasure to do that. I was raised as a son of a farming family in southern Germany with cattle breeding, viticulture, and arable farming. And I studied viticulture and enology at the universities of Geisenheim and Giessen and I received a doctorate in agricultural science. Upon completing my doctorate, I started my career, this was very interesting as an R&D for plant construction, specializing in food processing. Making, for example, apple processing continuously nor from apple to apple trees, integrating new machines and new processes. But then subsequently, and together with a partner, I had the opportunity to go to Spain and to found together with my partner, a sparkling wine estate there. And this I built up for seven years. This was another very good school, learning a different until that time, a very different culture to me. All the Spanish ways to do things are different.
Martin: And then after seven years of doing that, we sold that business because of health reasons of my partner. And then I became the member of the management team at the second-largest spirits producer in Germany. And there I was responsible for R&D and production. And this was another very challenging position with completely different drivers and yeah, things to be done and do. And during that time, my former partner in Spain, offered me then to enter at SAM, so I was able to realize my passion for sensory and consumer research taking the reins of SAM, sensory and marketing international at the time since 2005. And as a managing director and later also as a shareholder, I developed SAM together with the management team into one of the leading providers in the field. And today, as you said, we are a member of Eurofins Scientific with more than 15 own facilities. Before my international business line within Eurofins providing sensory and consumer research to our customers worldwide with a global network of more than 120 partners around the world.
John: That takes us to something we were talking about before the show, which I think is really very interesting. If we can just kind of jump into this, which is that with technology and with people getting more and more comfortable on Zoom, you were telling me that last year you were able to conduct a study in the Democratic Republic of the Congo, I believe is that the name is the Democratic Republic of Congo? Is that correct?
Martin: Yes, that's correct.
John: So that is and you did it without even having to send a staff member to the Congo? Is that is that correct?
John: So can you take us through that? Because that's very fascinating. That's, I think, really exciting because now we're opening up really new markets, new parts of the world to sensory and consumer research. You know, people are going to benefit there from the improved products that will follow. So can you take us through, like how that happened? And then the mechanics, how did you manage to successfully execute this test?
Martin: Yeah, the thing was that the company there was requiring us to go there and to make some provision, of course. And it was exactly before the pandemic started really to be, yeah, to come to a lockdown. And then it was clear for us that we are not able to travel, we cannot ensure that we go to supervision. But there was the problem. The prototypes for the research already were produced. And this was a huge amount of money, of course, for the company to be prepared during months, they prepared prototypes and everything was ready to be tested. And then the company came back to us and said, look, guys, what can we do? What we can do? We can change the protocol. We start with smaller sessions. Sessions of two persons, and we do a trial. We do a blind session, let's say without any consumer involvement. We just try it with the staff there and we control everything. We are videoconferencing and that we did, we consulted with our client, we said, look, we changed the protocol to something where we don't need to be on-site, we can control everything, every step, how to prepare the products. We look via camera, via zoom technology. We look there, we see how to measure the temperature, how to position the products, how to let's say poor in glasses and serve the products. And then the first session we did with the staff who later served the products? They behaved like a consumer coming into the session and this was even more than a training for them, how to later do the servings and everything.
Martin: And then the first session we did, we did not with five or seven consumers. We did with two consumers. And then there was a learning. Then after the session, we made debrief to the servers, to the staff there. And then they did the next session with five consumers and more and more. They get acquainted and familiar with the entire process. And then we completed a study. And this was great because we saved the prototypes. We could do the research. We saved travel expenses and quite a number of travel expenses because it would be very expensive and time-consuming to travel there.
Martin: So and in the end, we get very consistent results. And it was a very good research, by the way.
John: That's fascinating. So this was still a central location at that time? It wasn't before everything was...
Martin: It was a mix. It was a sensory central location test in a very controlled tent outside the normal buildings.
John: I see. It's kind of a hybrid.
Martin: Yes. Hybrid.
John: Yeah. And were the products then, where they profiled at European locations, or were they, I assume there were some sensory profiling as well?
Martin: Yeah, the sensory profiling was done with a panel with a descriptive panel on by the client has done that in headquarters.
John: I see. Well, this is honestly to me, Martin, that's fascinating for a number of reasons. For one thing, I think that there are real pockets of the world, I mean, especially Africa to some extent, parts of Asia where sensory science is still a relatively new discipline. And so I think the fact that you're bringing that to the heart of Africa is really very exciting and the fact that technology helped to make that happen. Yeah, I mean, I think that's phenomenal, right? Because all this knowledge is coming now in through Zoom into the room, and you're simultaneously training the staff and collecting data with, yeah, I mean to me, I think that's fascinating. That's a big game-changer for the field.
Martin: I mean, this was the first study we did. It was in March, April last year, and we did since then, several studies in the same way. So we call it Supervision 4.0
John: Supervision? Okay. Yeah. You know, when I was at the Institute for Perception, a lot of my time was spent traveling around the country. One year I was on a 104 flights traveling around supervising tests. I have to say it would have saved me a tremendous amount of time if we were using Supervision 4.0. So, that's I think, a very good model, regardless of where the tests are happening. That's really great. Okay, so let's move on then, because as we mentioned at the top, that you've been at SAM for more than 15 years. And I think it would be really good to get your perspective on how you see the role of technology in the field evolving. What have you seen based on how things were when you started, you know, into the present and then into the future, you know, for the tools that you think are the most useful? So what has been the kind of arc of technology that you've witnessed?
Martin: This is quite a big change. When I started, we did all the research with paper and pencil, right? Yeah, and we had a big scanner and the scanner was the machine, which was running day and night, and then you have to control, yeah, the scanned questionnaires and you have to complete and some of the questions were not really completely correct filled by the consumer. So you have to look back and to correct some mistakes which were done by the scanner. I mean, this is how I started. And then later the scanning machines were faster. And then after a while, it was 27-28, we started with digitizing, but this digitizing did not work very well and the first time we did really digitizing was 2011-2012. And I remember exactly the first online web-based questionnaire we had. Consumers were coming. The first question was answered, and then we had to wait five and a half minutes until the next question came. So there was a learning that you need to have a certain closeness to the silver because otherwise, it takes too long until the response comes back. And today, we modify the questionnaire five minutes before the test, and during the test, we have an adaptive questionnaire design. It's no problem.
John: Right. Yeah, well, that leads us to something that you and I have definitely shared passion for, which is a chatbot is in some sense an adaptive questionnaire. And of course, smart-speaker surveys, I think, are the next kind of evolution of surveys. So what do you see as kind of the most exciting technologies for data collection? I mean, we can talk about chatbot or smart speaker surveys or you talk about maybe some things, some of the more neurobiological type measurements. But what do you think are the additional measurements that technology helps us to make that are going to be the most interesting and useful in the next few years?
Martin: I mean, chatbot is for sure something that helps to be, yeah, to be not so rigid and not so strict. It allows a certain, let's say, qualitative elements in quantitative research. And this is one thing what I see will help to get a more realistic or let's say a more natural way of collecting data. Smart-speaker might be even more interesting for all the things that relates to body care and home care, where you need your hands to apply to make the normal way of product usage. And if you think about lotions, facial creams or hair treatment, or shower gel, all body-related things where you can be a smart speaker and have a direct feedback from the panelist, from the descriptive panelists on the one hand, but also from the consumer. And so you get more response at that moment when a product is applied.
Martin: I think in some product very important and can allow that you come to a direct response which currently is only possible by memory. Although, memories a few opinions, but it's memory and then it can fade out and I think as well there will be other opportunities, for example, not directly in the way we collect data. This is one thing, but let's say what I see coming is that we have new links combining different data sets and using bigger data sets. I think what we have done so far, the objective and the hedonic data set to link together and to make a model or several model to get something predictive, what would be the overall liking if we combine or we make an ideal profile? But I think today, instrumental analysis will be one of the thing where we can more use and this will help a lot. And also, we have more sources for qualitative data. If you only think about social listening, which we can combine as well those data as well, qualitative and quantitative at the same time. So it will help to make the research we do more robust and more, let's say, looking more left and right, and not only on the poor sensory perception. We will have more things we can integrate in a research to get more robust, less biased, more meaningful results in order to understand how market performance of products is built. And I mean, we both know or everybody knows that the sensory performance of products are not easily the sole explication of product success and of market performance.
John: However, I do think the sensory experience is very important for repeat purchase, especially when it comes to fulfilling a brand promise. I mean that to me, that's where I think the sensory, like marketing, sells the product for the first time. But then to some extent, the sensory is what helps to get people to buy it the second time. If the sensory experience delivers on whatever promise was made about marketing for the first purchase, I mean, how do you conceptualize sensory? How do you see that sensory fitting together with a larger kind of market research enterprise?
Martin: I think sensory explains very well overall acceptability and preference pattern. I mean, understanding the differences between cultures and markets and understand segmentation of the target groups because humans perceive products differently and therefore, they are likely different, have different cultural experiences, and that makes them liking or disliking products. Again, for sure, maintain acceptance level and capacity with sensory and consumer research. This is something what sensory and consumer research is really has its strength began as well anchor perception-based claiming. This is also very important and sensory characterization, we can make a driver analysis and modeling of ideal profiles, develop uniqueness of products, and we can as well use sensory and consumer research for sensory-driven new product development. But what I think is when it comes to repurchase as an entire building of repurchase, I think even there, the sensory part is not more than 50 % and claiming and prize and as well the psychological structure of the consumer is playing a big role in the point of sale. And this is why I think we cannot ignore brand and packaging and brand design and packaging design and claiming from the model, which explains market performance. To believe that sensory performance and the overall liking is the only thing explaining the repurchase is a mistake.
John: I definitely agree with that. So you mentioned in there the fact, there are so many subtleties and complexities when it comes to the sort of data that we deal with. Really driven by differences in people, differences in cultures, this kind of thing. But I think a lot of I would say classical data scientists if we can use that term or traditional data scientists have trouble analyzing consumer data or have trouble analyzing sensory and consumer data, because what looks like a lot of noise if you don't understand all these subtleties. You'll often hear data scientists say, a liking distributions look very flat because they don't understand that different people want different things. So you look at the distribution for a single product, it'll look quite flat. Now, if you understand the differences in the segments, then those liking distributions can start to tighten up. What do you think, Martin, when it comes to predictive modeling when it comes to AI within sensory and consumer science, where do you think AI could be the most helpful? And where do you think it's a mistake to try to build predictive models? What can AI really predict within a sensory, in your opinion?
Martin: I think there is a huge opportunity to go broader, to integrate other data than only liking related or product-related. I mean, the world is more complex, it is really not one single driving factor. I think we have at the moment certain paradigm changes and changes in the structure of how consumers access products. Make a few examples.
Martin: Health orientation. Health orientations of consumer is growing. I'm pretty sure that this is playing a much bigger role in the future than it has done in the past and this is because not only because of pandemic at the moment with covid-19 is an issue. This has accelerated the orientation for healthiness. But we have to state that the increase of the average age across the world as people get older automatically the attention to health increases. And this is one factor where people and consumers get more attentive to what products they consume. So this is one thing. The other thing is we have globalization taking place and the overriding cultural development we experience is enormous. So today, everything is available at any time anywhere.
Martin: And this is changing consumer behavior. Also, on the other hand, the offer is exploding as well as places where we can make our purchases. We can make them online. We can specialized jobs. We have the retail. We have huge superficial places. So this clearly has an impact as well. And then the third thing perhaps is the climate change. So environmental sustainability is gaining traction in the mind of consumers, affecting the behavior of what products they choose. We assume that market development gets a rational bias towards more sustainable products, even if something has to be sacrificed for it, such as paying a higher price or even a lesser sensory experience. And the last thing is the nutritional know-how, especially looking at macro nutritionals, which are salt, sugar, fat, protein, and also fibers. They fight against obesity and the offering in the shelf. I think I mean, when I was in Brazil at Pangborn when I was there, I saw there is a TV show where they had really obese people doing efforts in decreasing their weight and making competition across weeks and that in Brazil.
John: So this is not the American show? This was a Brazilian show?
John: Okay. In America, we have shows like that too.
Martin: Yeah, and I mean, it requires not only changed the recipes of the offer over the next year significantly but also develop new products with a completely new composition and other focus areas and I believe if you take this, the health orientation, the globalization, the climate change, the nutritional know-how and the change of what composition our food should have in order to avoid that we get obese and overweight. This is a huge challenge coming to the industry, how to manage that and how to get how to design the offer and what to provide to the consumers. This is for me with a classical approach of market research and sensory and consumer research not manageable. The complexity is enormous. We will have the need to have more tools to tackle the complexity, and this is the core business of artificial intelligence. Manage this complexity and manage real big databases.
John: Yes, well, I definitely I mean, I would say right now the biggest, what I see as the most important thing for sensory and consumer science groups to do is to get their data organized, that it's knowledge management is the biggest issue for us right now that we can talk about AI and machine learning and really, there's a lot of nice ideas. And to some extent, we can execute on those, but we're not going to be able to realize the vision that you're talking about until we get our data organized. Right? It's all these different data sets and different people's computers, different formats, different naming. That's a big problem, right? The same thing in two places at two different names. So we're going to have to get that sorted out. And I think that you know, our clients are at different stages on that journey, companies worldwide or on that different points along that journey. But I think it's the single most important thing and it may not be the most exciting topic, but until we identify these different sources of data and we connect them to each other and we learn to be able to easily pull the subsets of data that we're interested in, we're not going to be able to develop the models that you rightly have pointed out that we need. Right? I mean, it's going to be time-consuming. Whenever there's any idea we want to investigate, it's just going to take way too long to get the data together.
Martin: I fully agree. I recommend to our clients that they start doing a database well organized, well-defined. And in this database, not putting only the results of the research they are doing but putting at that time all available data they have, because you don't know what it is good for. Maybe at that time when you do the research, it's not so interesting to put the instrumental measurements they are in and to do let's say the market data they are in. But if you have them available, it's a low effort to get them. Whatever you can get for the data you have in research collected, complete it. Do it in database and even after one year, you have created datasets where you can get additional meaningful information out of it. And even if you look for the research we do day-to-day, I am pretty sure that we only get, I don't know, 30, 40, 50 percent of the insights which are in the data, a part of them is never touched because it's not the objective of the research. It's not nobody has thought about it. You did not really put in questions that there might be more behind. So and I think is another opportunity for AI because AI can deliver automatic processes to look if something in the data is in, which is something, yeah, what needs to be considered. Let's say some structure of the data is in some additional finding, some segments are in and all those can be very interesting to look at. But if you do not look at the moment, without AI, you do not do the effort. You don't find it. But I can ensure that everything is checked and everything is coming out.
John: Right and that could be either in the form of some rule-based process that you can specify or perhaps through pattern recognition. Like if you can identify patterns, right? You can also identify anomalies, anomalies, or the opposite. So we're on the same page with that 100 percent. I would actually even say that if you think about a large company and all the research they've ever done in a product category if you were to pull all of the data together, you've got all that body of knowledge. My guess is that the total insights that have been extracted from that total body of knowledge, not just one test or another, is probably less than five percent. I bet if you were to put it all together, right? And do analysis across all of it, I think there's an enormous wealth of information in there.
Martin: Yes, I fully agree with that.
John: Yeah. And I think two ideas that are very important are virtual pilots where you have a business question and you say, what data have we already collected that could help us answer this question, combined data you've already collected to analyze it as if it's a new study but actually it's existing data. And the other one, of course, is simulations, have created a simulator where you've got in a product category, a simulator informed by your historical data that can support computer-aided design. So I think that there are huge opportunities there and I'm really excited to hear you talk about that.
Martin: Just another example. In 2015, we integrated into our research toolbox, Bayesian network. So by using Bayesian network is looking at all the data and all the products and all the consumers at once and you can revisit all the data you have very easily with the network, and the results are because it's a completely independent and very different way of looking to the data. You get additional information for free. They are there. I mean, just you just have to apply another statistical approach, which is by using a network and you can get from the same data much more meaningful information, which by the way, we saw that several times. You don't get out with classical statistics. It's impossible to get there. And this makes it so very interesting and meaningful. And if, you know, this is an even more sexy thing, if you know that you will be able applying Bayesian network, then you can design your questionnaire accordingly. You can anchor specific questions, which you can later use as a target to be investigated. What defines this target? What is the driver to get to this target? If you then made with bayesian network different splits, then you get even more meaningful information which are in the data. They are already there. You just have to revisit them. And by using network, you don't need to get back to another primary data generation. You can use the existing data.
John: Now I think it's one of the best applications of looking at the historical database that most large companies have, sometimes hundreds or thousands of sensory profiling tests that they run and those can be used to form these Bayesian networks. I think that definitely just needs to be done. Everybody should be doing that.
Martin: Yeah. But let's say there is the speed of our day-to-day work and jumping from one project to another one and having no time to sit down and to reflect. So it's the business of the time. Yeah. And which perhaps does not allow that somebody is saying, hey, stop, let's have a second look and let's see what we get in additional information. And there I think is the big advantage what makes or we can do that, R&D is getting a real big acceleration of being faster. Not in the individual project, maybe the individual project will take the same time or even longer with AI because you'll get a more complex issue to solve. But the R&D itself gets faster because the insights are coming more robust, less biased with a higher level of prediction or even prescriptive, where we can have a lot of different scenarios comparing the best and even playing with different scenarios to get the best solution and this would help to be in the R&D and in the new product development, very focused on something, what can be very successful from the very beginning and I see there is the biggest advantage. Managing complexity to allow R&D to be more focused and more, let's say, success-relevant performing work.
John: Well, completely agree, and I think that's a really good place for us to wrap this up. So, Martin, this has been a really, to me, fascinating conversation. I appreciate you taking the time to talk with us. Well, how can people contact you following the show if someone has a question or they want to hire SAM, or what would be the best way for them to reach out to you?
Martin: The best ways either LinkedIn because in LinkedIn I think I have a big network there and I am present there. So this is one way to get in contact with me. Also, I am in the more German-speaking social network, which is Xing. Or you can access me with my company email address which is email@example.com that I am available as well and via our home page where you see my email address as well.
John: Wonderful. Okay, any last advice for our listeners? We usually like to wrap up with advice for a young sensory scientist, what would you say to someone who is just starting out in the field? What recommendations would you give them?
Martin: Make experience, especially in companies commercializing products, fast-moving consumer goods, in order to understand the part of market performance because it's very different if you go as a young researcher to a research company and you are asked to consult companies from the fast-moving consumer goods industry, having no idea how market performance is built up, how marketing is doing, how R&D is working, and the entire selling process because this is playing a big part, and this is the best school for later providing a very good consulting service to companies. When I build up my company in Spain, it was from the scratch. We had no production, no clients, nothing. And I did, I mean, if you have time, I can tell the story of my product success with sensory and consumer research. It is a very nice story because when I started this business in Spain, it was producing sparkling wine there. And of course, as a young entrepreneur, I had the idea of producing the best sparkling wine of Europe, of the world. And my partner at that time say, look, please do a sensory and consumer test to an ideal profile and I would like to see whether your product produced by your thoughts and your all enthusiasm and what you dedicate into this product is really as good. And we did a product optimization, and we achieved well with 12 well-selected competitor products, and my product is a prototype for the target market Germany. We achieved a fantastic result. We are significantly the best-rated product representing the ideal profile for the market. Bingo. Then with that in my backpack, I go to the retail and say, look, I have the best product, much better than all that you have in the shelf. Let's take that chance. You will have kind of a best seller. So it took us four years, 4 entire years until the purchase that allowed us to have an entrance in his market when I had this talk, I was really encouraged about this huge opportunity. I counted the number of outlets of the retailer. More than one hundred and fifty, and I expected the purchase order with about one hundred and fifty pallets, each with four hundred and eighty bottles as initials. Fantastic. I celebrated already the imagination of what is to come. Sadly, it was very different. The first order where 72 bottles placed a secondary offering in the middle of the aisle close to the checkout. Well, I thought never mind, my products being so good after one to two weeks, the big order will come because the 72 bottles are gone and sold and it's a self seller. But even worse, after three months, the purchaser informed me that he had to sell it on the price and that he will not place further orders for this product as it does not perform. What a disappointment after all the years and efforts. The moral of the story is that we did not take into account all the vital aspects that have been speaking about before. It's the pure product alone is not enough. The labeling, the entire image, all the things which are relevant for purchase and repurchase is much more than the product allows the product alone. And this was the learning process I had.
John: There's no better teacher than life, is there?
John: Experience is the best teacher by far. Yeah, I'm sad that happened to you, but it is a great story. It's great. And it's a very good place to end the show. Martin, thank you very much for being on the show today.
Martin: You're very welcome and thank you for giving me the opportunity to have this discussion with you.
John: It's been great. Thank you. 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!