Why Marketing Needs To Own Market Research With Molly Strawn-Carreño, Director of Brand Growth at AYTM | Ep. 429

0:01Welcome to the Marketing Millennials, the no BS marketing podcast. I'm Daniel Murray and join me for unfiltered conversations with the brains behind marketing's coolest companies. The one request I tell our guests, stories or it didn't happen. Get ready to turn the f*** up. Welcome back to another episode of the Market Millennials. Today, we have Molly, Director of Brand Growth at AYTM. We're going to talk a little bit about market research. It's an interesting topic, especially in the age of AI. So welcome to the podcast, Molly. Hi, Daniel. It's great to be here. Thanks so much for having me.

0:48I want to kick this off. And I know you spend years talking to researchers and marketers. So what is the biggest thing marketers think they know about their customer that's actually wrong? Ooh, I feel like there's a lot of things. Marketers tend to, especially experienced marketers, especially CMOs, people who have been in their industry or in their company for a very long time. They tend to know just at a gut instinct what works and what doesn't. But the thing that I think is very interesting about humans is who would have guessed that, you know, a worldwide pandemic hits. and people hoard toilet paper or buy bags of beans.

1:28There's just nothing that can actively predict what humans are going to do. So that experience is important, but it's also important to test those things to ensure that as your customer changes, as your company changes, as the times change, as everything changes around us, that you have an understanding of that person. And I think that that's what a lot of marketers could potentially get wrong is just get a little bit ahead of their assumptions and potentially leave things that are untested and spend a lot of time and money just for something to not be as effective as they would potentially like.

2:03The thing that I think is so funny is you see this online, the statement all the time that we are customer obsessed or I'm a customer obsessed marketer. In your definition, what does it actually take to become a customer obsessed marketer? Like what is the, from the research side of it? It means that you have an understanding of not just the, like, not just the way that your company helps them, but what their daily lives look like, what problems they encounter, what things that they find challenging in their lives and the reason that they do things. I think there are things that when we put on our marketer hat, we sort of take a, we disconnect ourselves a little bit from the choices that we make as a human being.

2:53And it's important to maintain both of those things. I can give you an example. When it comes to qualitative research, this was an example of a guest that I had on my podcast, The Curiosity Current. She was speaking about how she did a shop along with a mom and was trying to understand her purchases in the grocery store when she was shopping along with her. And she purchased organic strawberries, but not organic broccoli. And so that was a weird thing because you would think as a mom in this area, like she would potentially only choose organic, but why did she make that distinction? So when she was asked about that, she said, oh, well, the organic strawberries are for my kids and the broccoli that's not organic is for my husband.

3:37The guy doesn't need organic. He's already cooked. He's already the way that he is. But my kids, I have a bit more of a blank slate, so I care more about what's going into their body and he won't eat the strawberries and they won't eat the broccoli. And it was this moment of, okay, as a marketer, especially, that's something about their behavior that I should potentially know, but not something that I would have ever thought made rational sense until that human being tells you why that makes rational sense. So when I hear that marketers are customer obsessed, but they don't do any research or you get actually behind the curtain and they don't do a constant tracking or understanding or any sort of type of quantitative or qualitative research.

4:14It's sort of like a pet peeve to me, because if you're saying that your customer obsessed, you're not just obsessed with delivering them a cool experience. You're obsessed with meeting them where they are and having a full picture of them in order to actually deliver more of what they need in that moment. On that data point. So say they collected that data point because a lot of marketers, I think that I know, or I'm guilty of this too. Like we go deep in research we collect a lot of data um and that data just sits there we don't action on it um so how do you recommend someone set up a framework that what is the next step like how do i make sure that this gets action action so it actually influences like business decisions either it's like product or new marketing message or how we put our whatever how do we make sure it's action That's a great question because it's not something that only marketers are guilty of.

5:13Research companies are guilty of this also, especially as research organizations within companies are guilty of this also because they're, I'm going to butcher this. It was our CRO who said he used to work at a previous large B2C enterprise company. And he said, if only my company knew what they already knew. And what he meant by that was that there was so many disparate systems of research that was occurring. Marketing was doing a brand tracker. This product team was doing a package test. This other team was doing like a message test over here. And it was gathering all this data and information that didn't talk to each other.

5:51They were siloed together. So if only there was a way to put all of that data together for it to be more of a usable warehouse for everybody, this company would be unstoppable. And there's a lot I'd say you mentioned in the age of AI, and I'm sure we're going to get into this. But those are what those large language models are especially good at, is data warehousing, data packaging, and presenting those things and having those things queryable across an organization. And so there's market research tools that are coming out or will be coming out specifically aimed at marketers that help to warehouse all of that data and make it queryable for your specific thing that you're trying to ask, that you're trying to answer.

6:30So when you ask about how to potentially make those things actionable, I think that's the first thing is making sure that you have all those things available in that one place, potentially with an AI tool that you can reach at any point. The other thing is it actually starts with the research itself and the research question you're trying to ask and to be very clear and precise about what that research question is. So if you're saying my goal at the end of this is to decide blue package or red package, That's your question. It can be very, very easy to have project scope and bloat on this project.

7:06And suddenly what was a red or blue package is now, well, what's this font? Oh, but what's this? Oh, but what's the timeline? Oh, what's this? When that's the question that you needed answered and that can sometimes obscure your actionability. So you now have a bunch of teams that are trying to ask the same questions that are trying to ask different things. Let me get my questions in here. When in reality, you don't need those things. And as a marketer, it's a research skill that even researchers need to know because marketers are curious. Researchers are curious. They want to know every answer to every question, but you can't in that type of environment.

7:42You need to be very poignant about what you're trying to answer. Get that answer. Okay. Definitively in this package desk, we simulated a shelf. We had people go in and in, in a shopping, a virtual shopping environment. I set up Amazon. I set up a Walgreens. I set up all these different virtual shelves. The red package is the winner. The red package is the winner. Okay, we're going to go with the red package. We're going to take that one to market. Answer that question and don't try to answer anything else because then you have data that, to your point, is just sitting there because you just asked a bunch of questions because you were curious, but it wasn't actually what you were trying to answer.

8:18I feel like now that I'm going back in my head, a lot of reports you see out there are answering so many questions, especially the people who put out and um it's just even like foundationally in marketing when you're running a test or anything it should always be one one question you're trying to answer and there's a problem a lot of research where one research is not set up correctly so like the the results are actually biased or skewed or totally wrong so sometimes in that case like it's better and not even do research if you're not going to set up the research properly, which is crazy to think about.

9:01That could be hard to, especially for people who aren't trained in research. I myself did not get a degree. I don't have a fancy master's degree in psychology. I have a comm degree. And so what I, what I have learned from researchers is that there are things that you don't even think about that could buy us a result. So if you have, let's just say you have a question set and you're trying to do a multi-select of what things have you purchased at a grocery store in the last time you went shopping, please select the categories of all the things you purchased. Let's say for example, if you don't have those answer sets randomized, if you don't click that little randomized button, you will get an over-report of people that just select the first thing because there's a lot of quality metrics that sit in place of people who are just unattentive or people that, you know, are just clicking the top thing or when things are seemingly presented in a, um, in a more favorable light.

10:04So if you, if you ask a question, especially a yes or no question, or you ask the question where it seems like you want to get that specific answer. so um if you're like uh i'm trying to think of like an example of like do you enjoy barbecuing for example obviously the people are going to say oh well maybe they're asking questions about a barbecue maybe they're asking questions about a barbecue sauce i'm going to say yes and because that's what people just want to please the researcher they want to please that survey they want to they want to seem like a better version of themselves and these are all very like basic psychology things, but things that marketers might not have a fundamental understanding of.

10:45I think that's where also AI tools can come in because before, you know, you would just go in and say, you know, you'd go to Google forms or whatever and try to answer your questions. And to your point, you know, if you get a really biased result because it's what the marketer thinks is right, and it subconsciously ends up in that language, that's really hard to manage. and you could have skewed and biased results that lead you to the wrong place. But there are AI tools now that are backed by research knowledge and research language that you can go in in a more, for sure, not going to be biased environment and put in those questions and have it create a survey and a sampling plan for you.

11:28There was a really interesting thing I read recently. I'm totally blanking on the name. but it was this idea about Claude outputs and other types of language model outputs that said, you read a newspaper article about something you're deeply familiar with, something you've studied, something you've worked in, something that you're very familiar with. There's generalizations, there's oversimplifications, there's lack of concrete data, there's behind the times data, and you poke all these holes in it. On the flip side, you read a different article about something you don't have any concept or any sort of ground in at all, no understanding of.

12:07And all of a sudden, this article makes so much sense. You're learning a ton and you're going to take everything it says verbatim. They're both generated by AI, but one has your experience and one doesn't. So if this article has these crazy oversimplifications and things that you disagree with, the likelihood that that article also contains it is very, very high, except you won't be able to catch that. So a lot of times when marketers are going in and saying, Claude, create me a survey, ChatGPT, create me a survey, even those are still going to have those holes because it doesn't foundationally understand a lot of these research best practices.

12:49And also the prompts can be biased and what you're actually asking, how you're asking it. It's also the AI tools also trying to please you. We're still at that point as well in how these language models are learning. So it's very important for marketers to also consider that when thinking about going out and doing research and to potentially try and find a different tool that's backed by research organizations and researchers to help you get those more unbiased results. I think also one thing that you said a little bit earlier about the example of the person in the shop buying organic strawberries and then like non-organic broccoli.

13:29But if you would ask that person in a survey, like a different question, like a question about organic or not organic or something, her bias might come out and pick X versus Y. Like I always buy organic because I believe strongly in organic. Like she would probably never say I'm buying not organic for my husband because he has a ticking time bomb. Like that's something that like doesn't come out in surveys. I wanted to ask you the question because like people at family said that you had to like follow like the wallet versus follow the what people say. and how do you how and how do you set up or set up or think about doing research by following actual decisions people make versus just putting out a survey and people are just saying like I'm sustainable and then they buy something that's not sustainable or I like I care about the environment and then they buy something that's pretty much destroys the environment.

14:41So how do you follow that in a survey to make sure you are getting the right data back? There's a lot of things in there that we talk about with researchers. And the biggest thing is what's called the say-do gap, which we say a lot of in a lot of different industries, but essentially means the chasm between what you say that you do and then what you actually do. So an example would be, how many times do you go to the gym? That's a question of, okay, well, I'm a little insecure that I don't go to the gym that often. So I'm going to make myself feel better and say, I go to the gym four times a week, even though I haven't been to the gym in two months.

15:23So it's humans want to inherently look better, or that's something consciously, I know that I haven't been to the gym in two months. I'm intentionally trying to make myself look better on the survey. But then to your point about the organic, the shopper may not actually have any concept of what they're doing at all. They might have a psychology block or a bias of what I actually do. I'm not even consciously aware. So she thinks, yes, I buy organic. Of course, I buy organic all the time. And that buckets her into she's an organic shopper, she's a not organic shopper. When in reality, people are way more nuanced than that.

16:02And also to your point, I think a great example is I can use myself as an example. I'm getting my master's degree in environmental science. I really care about the environment. I care about my impact on the environment. But right now I have a young son who's 10 months old. I need my groceries delivered, which means plastic bags are being used, which means I don't care in which way you have to get this to me, just that you have to get this to me. You're going to use single-use produce bags, which I don't typically use if I go and do my own shopping. I'm going to be buying prepackaged foods that I wouldn't normally be buying, but it's because I don't have time to cook right now.

16:40So there's also seasons of life that affect how people function, where they are, what they're doing. Don't even get me started on work travel. I have my water bottle that I take with me everywhere. I don't use any single-use water bottles. But when I'm traveling, I absolutely do because this thing weighs a billion pounds. and I travel for work often. So there's a lot of things that make people very, very complicated. So when you ask, how do you actually get to that? It's almost an art as much as it's a science behind research. You have to ask things and utilize language in certain ways that get people to actually say the truth.

17:16So instead of saying, how many times have you gone to the gym this week and have it be something where people then have to say, oh, well, I went to the gym three times this week. You can say, ask the first question, how many times a week do you wish you went to the gym? I wish I went to the gym three times a week. Okay, great. How many times in reality can you get to the gym? Making it seem like maybe there's some external factors that are preventing you from going to the gym. It's not that you wouldn't choose to go to the gym. It's just that, you know, maybe you have to work late. Maybe you have to go pick up the kids.

17:51Maybe you're just like really tired at the end of the day because you just worked a full-time job. Maybe there's a barrier that's preventing you from doing this. So it's not necessarily it's your fault for being, you know, a lazy POS. You are actually, you're trying. How many times did you try to get to the gym? How many times you actually get to the gym? Breaking it down how people would feel more comfortable answering and giving them the opportunity to say, I wish that I did this more, But this is my current situation because then that can help you figure out in that say-do gap what people are saying they do versus what people wish that they had more of the ability to do.

18:28You know, what's also what's running through my head right now is how many, to say marketers in general, try to confine a buyer journey or a buyer process into stages or into something that's very logical. when right now you're buying decisions are way you said you have a kid and this like you have so many things going on here you travel there's so many different things in your life that can make your buying decisions different or make what you see during your buying decisions different like i have a kid and i do this i see that i see that i'm watching kid kids youtube here and i'm watching this here and this oh my my son has screwed up my spotify algorithm that i've worked real hard for.

19:22Exactly. I mean, I used to try it so funny because I, until Spotify, I like try to game it a little bit. Like I'm going to listen to this a little more just to make sure that I, and Spotify probably knows that now, but I think at first they probably didn't know that people are trying to game the behavior that they want to see at the end of. Right. But it's actually funny because at the end of the day, it always comes back to what you didn't actually listen to subconsciously, which is funny. I want to also go into, we talked about this up front, the AI conversation and research. Every marketer has the ability to have access to these AI tools, Claw, Gemini, ChatGPT.

20:06So do researchers have access to these AI tools. Do you think these tools are helping us get closer to our customer or is it also helping us or is it creating more noise in the research process? I think that it's both. I think it depends on how you use it and if you use it judiciously. I think it's very easy to panic and think, what is this going to do to humans? What is it going to do to the world? but there's, I sort of take comfort in the fact that it's happened before. People were having very similar conversations when the internet came to be. And we were like, what are these email things?

20:52Is this going to replace mail? Is this going to replace faxing? Like, Ooh, I'm going to get an email account. And it was a very similar conversation, but obviously the internet didn't take away our ability to critical think. It didn't take away our ability to do work and to be good marketers. Social media didn't change that. So, you know, so it's just going to mean that we have to adapt and we have to, you know, everybody's, everybody's using all of their tokens and figuring out what works and what doesn't for the, for their organizations, but it's going to just depend on how you use it. So it is very easy to get overwhelmed with the information.

21:32You can ask any AI, anything that you want, and it will give you an answer, whether it's a true answer or hallucination or, you know, that's debatable, but it will give you an answer. There's a, what's being talked about a lot in research right now is this idea around a real respondent, which is a human being sitting in front of their computer, taking a survey or a human being being watched going to the grocery store and going about their daily life. That is a real respondent versus what we're calling synthetic respondents. So synthetic respondents is if you went to chat GPT and you said, okay, I'm trying to determine red package or blue package.

22:10Here's my audience type. It's women under the age of 30 who don't have children, could be married or not married, who live in Southern California, who have purchased a beauty product from Sephora or an Ulta within the last six months. Tell me what you think, red or blue package. and it will give you an answer. I think that there's a really good way to use AI as sort of a BS layer detector. So let's just say you have 50 colors that you want to potentially test, every color of the rainbow. You could use AI in a pretty easy way to help narrow down that list. So it can tell you out of these 50 colors that it thinks orange, green, and blue are going to be your top three contenders for this particular product, then that is what you can ask real customers about, what they think between the three.

23:06So instead of asking customers what they think about the 50, you now have a somewhat tested, more narrow set of those three colors that you could potentially test among those real people. So I would say, yes, it can create a lot of noise, but so can a Google search, so can the internet, so can Reddit forums. There's things that have blurred the lines in what's made marketing challenge. We've been in data overload for many, many years. It's nothing new in that sense. It's more how you utilize it as a tool to get to your end goal. I like that answer because I think what you said at the top of the episode that AI has a great ability and probably we can ask about the research side of it too because it's trained on research.

23:52But I do think it has a great ability to take data points that would take a human weeks to sift through and connect them for you. find similarities, find things that stand out, find things that aren't standing out, validate something, maybe validate a question you have based on things that you, data you've collected, maybe through watching thousands of customers walk through a store and, or, because, or a thousand sales and sales reps talking to customers or. Yeah. It's exceptionally good at that. Yeah, but it's exceptionally bad at, not exceptionally bad, but it will give it, if you don't give it any of that context, it's going to find whatever context it needs to try, come up with an answer.

24:51But I want to ask you on the other side of it is, so it's really good at synthesizing data, maybe to come up with something. Exceptionally good. exceptionally good. If you have a ton of previous studies, and I will say, if you have clearance from your company to put your research into AI and you have that security clearance to do that, because I know that that's a whole other thing, it is exceptionally good at trend analysis. So I think the analysis of previous data is one of its strongest use cases, because to your point, for a human being to read all of that data, to comprehend all that data, to put it into Excel, to find all the charts, to do all the pivot tables.

25:28That was a whole role. That was a research analyst. That's a whole function. It's very, very good at doing that. I will add though, and I think this is where you're going with this, everything that's in AI is historic. It doesn't know about the future. It can make a prediction. It can say that based on the way that human beings behave in X scenario that I've seen before, people will react in Y scenario in the future, but it is not a perfect science. So when it comes to trends research or finding the next big thing or predicting the next big thing on social algorithms or what the next trend is going to be that as a marketer, you want to capitalize in that trend lifespan of like two weeks on TikTok.

26:14If that's your goal, AI is not going to be it for you because it doesn't know and it can't predict humans that well yet. I don't think there's ever going to be something that can predict humans that well. No, if there was... Not even humans. No, if there was, there would be tons of trillion dollar companies out there if you can predict what humans can do. And that's why there isn't because like you've been saying up top is humans are naturally illogical creatures. We're weird. And we decide on emotional, environmental, like the toilet paper example thing is you see a craze on TikTok and then you start having fear that you're not going to have toilet paper.

27:02And then you go to the store and then you want to you don't want to be the last person there. And then people are freaking out and it is just the spiral of inspire on a human's mind that I don't want to be the one left without toilet paper for months or years. So you just it's just a weird phenomenon that you have. But you can never predict that because you can never predict that tick tock happening. I think we've talked about this before we ever got on the podcast, but there's so many of these TikTok trends that you never guess that these ingredients would sell out because someone made a weird dish of ingredients.

27:40Tinned fish. Or dot cake. Or dot cake. Like, who would have guessed that dot cake would go viral right now? And everybody would start making dot cake things. It's like nobody would do that. Yeah. Yeah. The like, the thing that always weirds me out is that people were doing like tinned fish charcuterie boards, like fish shootery boards. I don't know. I'm probably butchering that. I'm like, I sound so old and out of touch, but it was like very confusing to me. Cause I was like, I thought sardines in a can were like disgusting, but now it's like very trendy. There's like, I would have never, I would have never thought things like that could catch on or things that were, that were popular in the past that were uncool are now like cool again.

28:20I don't know. Some things are very cyclical, like fashion can be very cyclical. So like pants that are like, I mean, we're on the Marketing Millennials podcast, right? So you can pry my skinny jeans out of my cold, dead hands. But like pants have come back. We're like back to the 90s type pants. So like sometimes trends can come back, but sometimes trends are just so off the wall. There's no way of knowing. Yeah, there's pendulum in all of society. um like yeah things swing back but there's also moment in time trends and those are hard to predict the moment in time versus uh what's actually a pendulum swinging trend that will come back like baggy shorts to skinny like baggy long shorts short shorts to baggy like i remember being in high school being young and be like why are people wearing short why do people wear short shorts and then every guy now is wearing like five like inch pants it's like and now it's going back to baggy again and like gen z's taking it back to baggy it's it's crazy uh i know i want to ask you also um besides the ai being really good at sensitizing data what do you think is ai is actually like surprisingly good at that people aren't thinking about um when it comes to research I think that AI is really good at simplifying processes of things that you do over and over and over again.