• John Ennis

Amanda Grzeda - Freedom to Deliver Business Value


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!


Amanda is an R&D director for the Sensory & Consumer Product Insights team at PepsiCo, leading research for global snacks. While collaborating daily with fellow sensory professionals, product developers, and business leaders representing diverse product portfolios from around the world, she ensures that consumer empathy is top of mind for key business decisions.


Note: Amanda Grzeda is an employee of PepsiCo, Inc. The views expressed in this presentation are those of the author and do not necessarily reflect the position or policy of PepsiCo, Inc.


Amanda's Contact Information:

LinkedIn




Transcript (Semi-automated, forgive typos!)

John: I like to share a small personal story, which is that I actually did know Amanda until last year at SSP. What happened is I was in a session and I thought, wow, the speaker is really good. I have to meet her. And then that night I saw her on the walk to the Rock 'n Roll Hall of Fame. So I ran up and introduced myself. And we've been friends ever since. So, Amanda, it's really a pleasure to have you on the show.


Amanda: Yeah. Thanks so much for having me, John. You flatter me.


John: You deserve it. So it's good. Okay. So actually, it's kind of interesting that you're on the show today, Amanda, because I had your colleague Frank Rossi just on a previous episode. So in that call with Frank, he and I talked about how the skills that are required to support sensory scientist because he's a statistician, he's supporting sensory scientist. He has seen that those skills have changed and that what he looks for now and he makes hires is very different actually from the skills that he had when he was hired. So I think it would be interesting to hear your perspective about how the skills have changed for the sensory science side. What do you see now that is really needed to excel as a sensory scientist compared to how things might have been when you started your career?


Amanda: Yeah, it's an interesting question. I guess in general, I think that sensory science itself is very, you know, the solid foundation that we all have to have. And it's consistent across generations of sensory scientists. You know, we're always continuing to evolve and learn new methods and techniques. But the foundational science is pretty consistent. Where I'm seeing bigger differences is really just the that add on top skills that we sensory scientists use in a more business environment. So if you think about it historically, some of those that have been in the game a little bit longer, we may have spent, you know, quite a time on data analysis, you know, questionnaire building, things like that have become more automated in today's world, whereas maybe the younger folks come straight in and they see the need to focus directly on interpretation. And, you know, translating those results to my business partners. So I think it's really on us that have been around awhile to make sure we're spending the time honing those skills of communicating our data with impact as opposed to saying, oh, I just get to save all this time because I don't have to do the analysis anymore. It's really like, where are you going to spend those moments going forward? It's just a little bit different focus of time, I think.


John: Right. I mean, that is really interesting. So maybe we just talk a little bit about things that you see that have changed, that have freed up this extra time. So like, you know, where do you see automation is a hot topic right now? Right? And actually it's interesting, a lot of fear that people are going to lose their jobs due to automation. But actually, I think what you're seeing is the opposite is true, that instead what will happen is time will be saved to do other maybe more interesting things. So what are kind of the key pieces when you think about automation, what are the key things that have been automated or you know, the changes in the processes that are changing sensory science?


Amanda: Yeah, I think what we've done an excellent job at is automating the things that had little value. So almost the must-have attributes of what a sensory scientists are or have begun to be automated. So, you know, things like, you know, building questionnaires detailed in word and then copying over into a software system, things like data analysis where it's really by rote things that were just really repetitive in doing. And of course, those things are incredibly important to make sure that all of our data is completely accurate. But we don't have to spend that time there anymore. I think the systems that automate those processes for us are really freeing us up to add value, add more value to our business partners, add more value to the R&D partners we support, the marketing partners we support, you know, our other business partners that are trying to really understand what do sensory scientists bring to the table. And so because I think it's a huge opportunity for us to not just gonna be, you know, working away at our desks, running tests and submitting reports, but actually shifting our time to okay great, I got the data right away. So now I'm going to spend time interpreting and I'm going to look for themes across multiple tests. And I'm going to lift insights for, you know, it just kind of focusing our attention to things that deliver business impact, which I think is the long term health of our science, is we have to make sure that we're adding value or delivering value. And so I don't think the solution is oh great, we save time, let's run 50 more tests. I think it's different. I think it's we're running the right amount of tests. Let's use that time to make sure that the data we're creating is a landing and is really delivering that consumer centricity so that our partners feel like we are the consumer in the room, that we're representing the opinions of all these folks that are giving us input on the products or designing.


John: That's really interesting. So it sounds like you're an advocate for sound quality over quantity. That we don't just do more, okay, great we can do things faster, so let's do more of them, but rather let's instead focus on the new resources we have available, the extra resources we now find ourselves with to focus on, for example, things like storytelling. Is that something that you would say or when you're thinking about, like really delivering value what are the key things that you're thinking about?


Amanda: Yeah, absolutely, so I think, you know, we've we've thought about storytelling as a as a critical component of this because we know not all of our partners receive the kind of data we create in the same way. So if you're presenting within your sensory team, you're gonna go to one level of detail, you know, on the methodology and the detail, how you collected the data, how you did the analysis. If you're presenting to R&D leadership, you're going to shift that message a little bit to get it to land with them. Hey, this is the process that we used. And here's the most efficient way we think and produce this product, because consumers like these two equally. And I understand what you're up against from ingredient cost standpoint. So I can I can be part of that conversation. If you're trying to cater a message to a senior leadership team, you're going to take a very different approach. And so I think taking the time, you know, even in that example is one study and we might write five reports based on the one data set, depending on who that audiences. And just even having that vision to understand the different messaging is required. I think there's a long way in getting the data, the message to land. You can't just assume that the one report you write is gonna be a perfect fit for everybody. You know, I think for most of those audiences, they might just, you know, say I have no idea what this is. Or even worse, I don't understand this so I'm going a question every little inch of everything you said. And then you have not really getting to the heart of the message. You're debating methodology and what you choose this city to test. You know, that's not at all where we want to be. We want to focus the message. Make it land. Make it resonate with whoever it is we're communicating with.


John: And so the tools for automation then, like, okay, first you mentioned the statistical tests. There's no point spending a lot of time running statistical tests when there are a lot of them are routine. What about the slide creation? I mean, then do you use tools to produce many, many slides that you then pick from depending on the deck you're building? Or what's the kind of process that you would recommend that someone who's trying to leverage automation follow in order to build all these different reports?


Amanda: Yeah, I think, you know, one of the first things that we did was we looked at, hey, what are all the typical pages that we create and let's create a consistent report flow. So from scientists to scientists, we don't want our business partners to be like,okay, Amanda's presenting now I have to figure out her style and absorb that versus someone else on our team. So I think that's a keeper step is kind of saying here's if were to include all of these pages, here's the order that they would they would come in and then I think it's yeah, it's curating. Here's the full report. Let me curate a message for this person. Let me curate a different message for this person and really depends on you know, you have to take time to think. What are their motivations? What are their key questions? What do they care most about? And then pull from that story to really kind of land the message with them specifically. So I think, you know, it's pretty straightforward at this point to sort of automate those pages. You know, at least this is the chart we want that the headline or the title of the chart, you know, things like that. And then really, it's so we don't have to spend time doing that. Then we can kind of add the interpretation, you know curate the pages into different kind of sub-decks.


John: I see. Yeah, I mean, it's fascinating because I do think that, like from data science and sensory like for me what data science brings that everyone can benefit from is an emphasis on processes that we're going to try to automate processes as much as we can. We're going to try to have a set of steps that we follow that's consistent. Right? And ideally, you would like to go back to some prior research and reproduce everything you did. Now, I think at some point, you know, the paths start to diverge. Probably the deck that you make, you know, for your I don't know, whoever you might executive you might present to. Maybe there's some custom stuff there that you just did by hand and that probably is okay. But at least until you get some point. It's all been more or less routine or at least this process that's followed. So maybe it'd be interesting, a couple directions would go this, Amanda. One would be if you could recommend different tools that you like that to our users. What do you, supposed someone is not in your position. They're not in the kind of mature position. Like there's more organizations and they're not as mature when it comes to data science and they're trying to get started. What would be the steps you would recommend for them getting started? Tools you'd recommend? Like how should someone who aspires to be like you start to move in the direction of this more kind of data scientific approach?


Amanda: Yeah. I mean, I think it starts with that internal analysis of, you know, look at you, look at your team, look at your colleagues and say, what is it that we do every day? You know, let's talk about our processes. What are the tasks that we do, you know, day in and day out. And then for those that are common across multiple people, number one is try to come together to say, hey, can we agree to do this in a similar fashion? That's the hardest part, is having that conversation of, hey, can you know, can we all agree to take the same approach? Can we all align on what we would prefer our standard report flow to look like? And once you have that, I think then you can say, okay, what are the ones that take up the most time or, you know, that are happened the most frequently and trying to figure out ways, you know, using technology to automate those things as much as possible or even just systematise even if you don't if you say it's not a matter of software, it's not a matter of, you know, a specific tool even. It's what are the things that we would process so we don't have to reinvent the wheel every single time. So sometimes it's a matter of just communicating across your team and your teammates to say, hey, I found a quick way to do this. We should all be doing this because it could save us all a bunch of time. So that's kind of it doesn't really require it's really the mindset of automation and not necessarily. Well, I can't do it because we don't have funding to do this, you know, to buy the software that's going to create this these templates for us. Yeah, that's not at all where we started. You know, it was really just look at what you can control and see, you know, how can you make the best of what you're currently doing?


John: Yeah, that's fascinating. It reminds me, I was actually teaching a course once with a client where I was talking in data science actually about scripts. To me, of course, a script means, you know, scripting language stuff like R, for example R scripts, we have R scripts. Well, there is someone in the room who thought I just met scripts in the sense of like writing down the steps. And actually that was more helpful to them than thinking about R scripts that literally if you have a script to follow, it doesn't have to be a piece of software. It could it's just instructions. And that if you have a shared Google document that says like to anybody who's working on it, this is how we do this sort of project. Here are the steps you follow. Like we have a script that we used to get ready for AigoraCast and it's helpful, you know, like just having a process that you can you can list. You know, I think it was, I can't remember which scientists said this that if you have a process. Oh. "If you think without writing, you only think you're thinking". Do you know who said that?


Amanda: I don't.


John: Yeah. It might be Richard Feynman, someone like that, said if you think without writing, you only think you're thinking that like at some point getting things out of your head and into a document, whether that document is an R program or whether it's, you know, a Google Drive document or even just beginning a piece of paper, that's already very serious progress. And I think that's the spirit, you know, of the data science is having a process. Okay. The thing is, my experience, whenever you get a large group of people together, you try to get them to agree on a standard way of doing it, you may or may not be successful. So, you know, do you have advice for that kind of change management piece if you've got, like, different? I mean, it can be really tricky in a global organization where you kind of have tribes where it's like, well, this is our tribe, does it? But do you have advice for like how to gather the troops and move them forward in a unified way? What's your advice there?


Amanda: Yeah. I mean, I think the first thing is to acknowledge how difficult it is. It is absolutely difficult to harmonize to a consistent approach. We've all grown up doing things our own way and we feel like there's really valid reasons for that. So really, it's in my head where we find that we can make a lot of progress is shifting the mindset from this works best for me. This makes me most efficient to thinking as it as a unit, thinking as a group. This makes the group more efficient. Therefore, I can flex a little bit because I see the benefit for everybody. And it does take time to shift people's vision from step A to step B. So it's really I think case studies are a great way to bring people along to say, "Hey, here's where we started, here's where we finish" and look what it brought to the group. You know, I had to change. I had to shift how I did things, but in the end, it was really successful. So I guess, you know, don't assume that harmonization is going to be easy. It's never easy. But find the things you know, find the common ground and, you know, write it down. We do that absolutely. You know, we have a document library of methods that we've harmonized to. Make sure that, you know when you're bringing on new people, even that they kind of, you know, enter the fold to say, hey, we've already solved that one. Go read this document. This is our process. And the other thing, too, is, you know, make sure that people understand these are not set in stone. So when you harmonize to something, when you agree on a process, it doesn't mean that that's it, that that's final, that that's never gonna change. We have a real strong spirit of continuous improvement. So, you know, hey, let's experience this. Let's call it, you know, is it safe to try? If we can all align to that, let's try it. Let's try it for six months. Let's revisit what's broken. What do we need to improve upon? You know, let's take our hindsight and say, hey, next time how might we do that a little bit differently? How might we adjust our process? So I think it's those having that bit of flex really helps people say, okay, I can live with this. You know, it's a consensus to the decision. You know, maybe not a 100 percent agreement and then see where you are in six months. I think a lot of people go, hey, that wasn't as painful as I thought it might be. And you find, hey, you know, we have the successful protocol now because we were able to all agree to give it a shot. And then we lived it for a bit and then other ones, you're right. You know, like, wow, that did not work how we anticipated it. We've got to make some changes. You know, sorry, guys, that one didn't pan out but that's okay because we're learning as we go. We're getting better.


John: Yeah right. Continuous improvement, that's so important. Yeah. That was the advice that got me through PhD was do anything, then make it better. That's the bit of advice my dad gave me, and I use that all the time. Just do anything and then make it better. So then, would you recommend regular meetings, team meetings to talk about processes? I mean, how does that actually get implemented? Do you schedule every six months we're going to review of our processes or like, how is that happening?


Amanda: I think, you know, we kind of do it more broadly. So we have tenuous improvement sessions where we say, okay, let's get together with the sole purpose of evaluating ourselves. And then we choose to pick the topics that are both compelling. So that could be this was a huge success. I really love this. Let's highlight it or this is a bit of a pinpoint. You know, I think we can improve something here. It doesn't feel quite right. So I think by being, you know, setting that time aside, totally separate from we have to work on our product team protocol or we have to work on this protocol. Let's get together and just focus on continuous improvement. We've tried a lot of really great success in, you know, sharing. Oh, here's where we see differences. You know, kind of more in an observational standpoint. This is different than this. A different than B. Then separately, you have the what is better. So it's kind of like breaking it into smaller chunks to figure out, you know, let's take the time to see how we're doing and then prioritize what are the things we want to fix most. Because you can't fix everything. There's too much going on. So we really have to prioritize and focus.


John: Yeah, it's fascinating. Sounds awesome, actually. Sounds like a great place to work. So if people and listeners, they want to work for PepsiCo, you can share your experience as you can. So now what about another topic that's on my mind a lot is why standardization is important is data management, because I think that in, you know, our field, I don't know. I mean, I have hypotheses as to why this is. But our data are very diverse. And usually once you get into the historical archives of the data set for a large company, you'll see every possible configuration of how the data might be stored and you'll see everything. It's amazing, actually. Once you really get into it, you'll see things you never thought you were going to see. Like different weird file formats of Excel that you didn't know existed or you know data. Yeah. Anyway, so maybe can you talk a little bit also about because there's that processes when it comes to analytics and reporting, but there's also the data management. So do you have any recommendations on how to get people together as far as how you're gonna manage your data?


Amanda: Yeah, I mean, I think with what we learned is it's incredibly important. It's critical that going in you have some sort of a question that you're trying to answer. And even that was kind of hard to get to. So I think, you know, with all of the hype and the excitement around data and big data, and we can get magical new insights from the data we already have without any specific question. You know, you go straight to the data management piece, you say, okay, let's form all of excels. Let's create a database. Let's, you know, let's form out the data. But you get to the data and you, for all the reasons you just mentioned, you realize this is way too big to manage. We just can't we can't jump into the deep end immediately. So what we found is ask a smaller question and really keep it lesser focused. So, alright, we're gonna collect all the data that relates to this very specific question for now. And we're going to see how much of the question can we answer? Do we feel really confident that we got the answer or do we feel like we're missing something? And so you can almost add things piecemeal by starting with a really focus question and then saying, okay, we know there's a whole bunch of data out there but we're gonna lock it down to this this little box. Let's fill that in. And then as you do that, you learn a little bit at a time about all the crazy data formats and how we're going to handle this data and how we're going to handle that data. But it kind of breaks off into smaller chunks so that I guess not only can you answer questions, critical business questions, but you can feel a little bit more confident of how do I wrangle this massive data? You know you can take one step forward as opposed to just standing there staring at this complex environment and having no idea how to make any progress at all.


John: And plus, you're motivated by a question you care about, right? That you have a reason to get into it rather than. Yeah. I mean, I think there is this urge sometimes that, okay, we'll just get all the data formatted the same way, which could be a tremendous amount of work. And then we'll hope that something good will happen versus what you've described. It's interesting. What you describe is actually kind of a local graph search the way if you did have a graph database. What you've described is a local graph search. That would be better. But you're doing it. You know that the you know, the data set level. So that isn't actually kind of interesting to me. Okay, so we actually don't have a whole bunch of time left. So there's something really gets important, Amanda would be if you could give your advice like so you now are about 20 years into your career so or 15 years? How long have you been at PepsiCo?


Amanda: Yeah. So I'm actually a chemical engineer by training. So I started here and was a product developer. I literally had no idea that food science or sensory science existed until I was working here at PepsiCo as a product developer. And then I saw the light and jumped over to the sensory science and haven't looked back. So yeah, it's about 10 years in sensory science and about 7 in product development.


John: I see. Yeah, That's fascinating because my father began his career in sensory as a product developer of Frito-Lay and actually we lived in Texas when I was a kid for a very short period time. So it is funny how these paths that we take into sensory. So, supposed that you have a young sensory scientist, what advice would you have for her or him about how, you know, what are the things they should be focusing on right now in light of the fact that automation is changing the way that we work?


Amanda: I think, you know, first is make sure you have a great fundamental grasp of of the science. So everything we do is built on, you know, make sure you understand all the details on why is that discrimination is that the one that you chose for this particular business question, you know, always just be buttoned up. You don't necessarily have to communicate all that, but you need to know it. And then second is leverage your teammates, leverage, leverage the automation that exists, and really focus your time on delivering results with business impact in mind. Make sure that every test we run, every test you run, every report you write, that method lands. And if you're asked to run a study that you don't really see the purpose, then ask that question, you know, what are we solving here? Because a lot of times if we can stop and ask some of those questions on why are we doing this? What do we expect to learn? We can even, you know, evolve the research a bit to make sure that it's more impactful than just blindly saying, yes, okay. Yep, we can run that. I know the perfect method. You know, really ask a couple follow up questions to make sure that the research is as robust as possible so that we're getting everything out of it when we do run the study.


John: Yeah, that's been a big change for me is thinking you know as a scientist, I think about, okay, what's the research question? But actually, what's the business question is an even more important thing to think about. What's the business question here? You know, and then also, you know, talking I mean, you are speaking to the either you are the decision maker or you're speaking with the decision makers but, you know, that was a lesson for me to make sure that you find the person who really owns the research and get their take on it. Because if it's easy, if you're several layers down to end up doing a lot of good work, that doesn't land at all because it didn't fit. So I really agree with that. So is there anything else you'd like to add, Amanda other bits of advice?


Amanda: I don't think so. I mean, I think what I've realized is kind of late is, you know, get involved in some of the external societies, you know, in our science, because I think you can really learn a lot from, you know today we talked about kind of general things. But, you know, by building your network of your professional network outside of wherever it is you work or study, I think you can really get a more broad perspective and it's super helpful to see your day to day to realize that you're not the only ones that are facing these challenges and everyone's trying to solve these issues at the same time. And you can get really great ideas from each other. So whether it's, you know, submitting papers to Pangborn or getting involved in the society of sensory professionals or sensometrics. You know, make sure that you're fostering that professional network because, you know, again, like, you can make friends like I've made with John and really learn a lot about, you know, things that you may not have thought about before or just bring home things to your current job to make your life a little bit easier. I think it's very important.


John: Yeah, I totally agree with that. I think sometimes people only think about the people outside the company when they're looking for a job. But actually, there's a lot of information, like a lot of benefits to networking outside your company that even if you're going to always stay where you are, you'll be a better, you know, sensory professional as a result.


Amanda: Absolutely.


John: Okay, that's great, Amanda. So where can people if they want to reach out to you to follow up purchases to connect with you, how should they connect with you?


Amanda: It's probably easiest just to look for me on LinkedIn. I have my semi-unique last name, so I think you can probably find me pretty easily there. And I'm happy to get to reach out. If you have questions, you know, feel free to let me know and I'd love to talk to you.


John: Okay. Sounds great. We'll put your LinkedIn like URL in the show notes so people will find you. Yeah. Given the millions of people who listen AigoraCast, I'm sure you'd be flooded with. It's true, though, actually, you know, we are actually doing pretty well. We're trending on Apple podcasts in science. We were like 178.


Amanda: Wow! Congratulations.


John: So. Yeah. So it's going all right. I mean, it is fun. And people contact me to say that they enjoy the show. So I'm sure they'll get a lot of this episode. So thank you very much, Amanda. It was extremely wise advice. And I hope that everyone listening take your advice because they will definitely do better.


Amanda: Awesome. Thank you so much for having me. I really appreciate the opportunity.


John: Alright. Thanks a lot. 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!



©2020 Aigora