An outsider's case for why surveys will outlast every trend with Marin Mrsa

Listen

Watch on YouTube · Apple Podcasts · Spotify

Summary

Marin Mrsa, founder and CEO of Peekator, on why surveys are future-proof (they are a scalable communication tool, not a boring form of questions), why an all-in-one research stack is a dream the market keeps resisting, where AI should accelerate the research workflow and where it must never make the decision, and a Current 101 plea for more cross-competitor collaboration on panel fraud and for research tech to stay in its lane rather than be asked to be McKinsey.

Guest

Marin Mrsa — Founder and CEO, Peekator · Supplier-side

From this episode — top claims

• AI should be used to co-create — to suggest ideas, check conditions, and accelerate — but never to make the decision for you; asking AI to design the whole questionnaire and project is lazy, and a real researcher must stay in the loop to catch where AI's summaries and action points go wrong. — Ep. 59, Marin Mrsa
• Research tech should stay in its lane and do research — showing the insights about your clients — rather than being pushed by buyers to act like a strategy consultancy; clients increasingly want research tools to be McKinsey-style consultants while paying research-tool prices, and that is not a fair ask. — Ep. 59, Marin Mrsa
• Surveys are future-proof — not because the survey format is sacred, but because a survey is fundamentally a tool for speaking with your customers at scale, and businesses will always want to hear back from their clients directly rather than ask an LLM whether their client is happy. — Ep. 59, Marin Mrsa

aytm's take on this conversation

A survey isn't a format to defend; it's how you hear customers at scale, so keep reshaping it. Let AI accelerate the slow steps, never author the questionnaire or own the call, and stop demanding McKinsey strategy at research-tool prices.

aytm's perspective, voiced by Molly.

Full transcript

Auto-generated captions — speaker labels aren't always available and wording may be approximate.

0:00 It's a big trend these days to call everything is ending. Surveys are being around for 150 years and they survived wars, technologies, whatever it is, they survived. And it's not that I'm such a big fan of a survey. I'm more a fan of speaking with clients and I see a survey as a communication tool of speaking with clients in a scalable way. That's how I view them. I don't view them all as form of questions. I view them as if you have 100 clients, how are you going to speak with all of them? And I feel like in future, no matter what, we are still going to hear back from our clients and that's the reason why I think a survey is essentially going to be around.

0:49 Hello, fellow insight seekers. I'm your host, Molly, and welcome to the Curiosity Current. We're so glad to have you here. And I'm your host, Stephanie. We're here to dive into the fast-moving waters of market research, where curiosity isn't just encouraged, it's essential. Each episode will explore what's shaping the world of consumer behavior, from fresh trends and new tech to the stories behind the data. From bold innovations to the human quirks that move markets, we'll explore how curiosity fuels smarter research and sharper insights.

1:23 So, whether you're deep into the data or just here for the fun of discovery, grab your life vest and join us as we ride Curiosity Current. Today on the Curiosity Current, I am joined by Marin Mersa, founder and CEO of Peekator. Marin's path into the insights industry is one of those stories that immediately makes you lean in. He started Peekator in 2017 after selling his dad's old car and moving to Zagreb with no prior experience in market research and built it into a company now working with clients across the globe.

1:58 Since then, Marin has become a thoughtful and often provocative voice in the industry, writing about survey quality, research technology, data integrity, and why the role of researchers may become even more important as AI becomes more embedded in how decisions get made. So, today on the podcast, we're going to explore a tension that feels very alive in research right now. We have more tools, more data, and more automation than ever before, but that doesn't automatically translate into clarity.

2:28 So, in that context, we'll explore why do surveys matter? What should research technology actually be doing for us? And where does human judgment become even more valuable than it is now? Marin, welcome to the show. Thank you so much. It's great to be here, and I appreciate the introduction. It was really a nice one to hear, let's say. I love it. Well, I'm excited to chat with you today. Let's get right into it. I think something that's really struck me about your story is that you didn't come into this industry through the usual front door. You built your way in.

3:02 Um you went from hotel accounting to starting a research tech company from scratch, and that usually only happens when something really grabs you. So, I want to start there. What first made you look at the insights industry and think, "There is something worth building around here." Yeah, sure. So, the story like started I'm Croatian based, and we have a Croatian dream. Croatian dream is to own an apartment and own a restaurant or a bar and live off it, right? So, so that's Croatian dream, so live on the beach. Uh live on the beach, and that's the Croatian dream. So, my idea back then was I want to own a restaurant as well. But, my concern was how I'm going to maintain the quality of service and stuff like that. The staff, are they going to be good? Are they going to maintain the standards and stuff like that? And I was also back then working in hotel and I was big on and

4:02 I like I was always big on software tools. So, I wanted to combine both. I so, I wanted to create an app where you could book a person who would go to your bar or restaurant and check the quality of service. So, that was my first idea. Like, I was really hooked to that idea. I thought that it's the uh idea that is going to change the world. It's going to be amazing. I didn't knew that too much about research about what type of service I'm going to do and what that.

4:30 So, I was young. I wanted to start it. I wanted to start a startup and that's how I quit my job, sold the car, the only car that I had from my dad, of course, not not not like my own, and started it. Uh it turned out it wasn't the best idea. Uh it wasn't the idea that changed the world, but it changed me and I started with it and that was, let's say, the first like introduction to research world. That uh mystery shopping app that we created. So, that was the first initial initial uh intro to the world of insights.

5:09 Got you. I like the So, it was a mystery shop sort of uh entrance. Got you. Got you. Got you. I think that there could be something really clarifying about entering an industry from the outside. Sometimes you see patterns or blind spots that insiders just are innured to, right? From sitting and marinating in this space for a long time. So, when you first stepped into market research, I'm curious to know like what surprised you about how this industry works and what did you think it might have underestimated about itself? Yeah, so uh first of all, I I I really felt like it's a closed world. I don't know like I felt like it's a small closed world on inside. It was really tough to crack. It felt like everyone knew each other other and it and it felt really close. One more thing that I realized as well that so at that time I also attended an MBA.

6:07 It was really interesting for me that we get to learn about every function in business and that was from sales to finance. Operations, purchasing, IT, everything. So we went through all of it but there wasn't a word about insights, about research. So insights felt to me often overlooked and not and if you go and ask a person in a bigger like, you know, organization, where is your insights team? I'm sure they are they are not going to know who they are and what do they do. So that like struck me as well that most of the people most of the people who work in enterprises, wherever, they are not aware about insights teams and stuff like that. At least that was my experience back then. So that was quite interesting

7:08 thing for me to see at that at that age. That makes a lot of sense. And I think it kind of to continue in that thread, I think there's also the founder side of this which is always fascinating. You know, plenty of things begin as experiments but not many become real businesses. When did Picked Or stop feeling like an idea you were testing and start feeling like something that could really matter and make changes in the space?

7:36 Yeah, you know, so when I started it we had that whole first Uber for mystery shopping app ideas. So, that is the first idea. It wasn't the best idea as as as I said. We didn't uh, so we kind of evolved from it um, and we started uh, I bumped into So, yeah, the story was also like I was 6 months in like 6 months mm, into it. Couldn't find the best fit.

8:06 I like wasn't sure what to do and then I don't know. Uh, I lived in this flat and the owner of that flat worked at this really big bank. So, as an every, you know, entrepreneur I got I got uh, to sales mode. I asked him like, "Hey, could you introduce me to a person you know in your bank? If you know, if you have a research team to show them this app. Could this be uh, like an interesting thing for your bank?" And luckily he liked me uh, because I paid the rent, of course. But uh, he liked me and connected me to them.

8:43 So, I went to this bank like fully suit on and it was the first thing I had a suit and none of them had suits. So, that was like first thing like, "What am I doing inside a bank? I have a suit they don't have." But then in some way I convinced them to try it out. And that was like the first big thing for me because I was really young and especially back in Croatia there was a mindset you couldn't work like a big a bank like that if you don't have someone inside. If you don't have a connection there, you know, like So, that was more a confidence win for me uh, for the short time that we were doing that okay, so we could So, we could do this.

9:26 And that what a really good thing what we did because we were so young um, the next 4 years I literally find a some really good clients um, and went to this where I would meet them and just ask them for a specific problem they have and then let me go with my team try to solve it. And that's exactly what we did for 40 years with the with some really amazing brands. We just like uh try to solve a specific a problem one at each time and and but uh back to your to you like that first bank was the the the like the big win the first real win that felt like okay we are we are not now playing a business, you know, like we are working with really big bank. If they choose us, it it must be that that that we can do a thing or two. Yeah, that must be such an amazing experience and I I think that dynamic that you're talking

10:23 about I've seen it play out too here at AYTM this sort of like you know, you have this client or customer who becomes your co-creation partner essentially, right? And you and when that happens it can be absolutely magical because you are there's no chance that you're not you're building a solution for a problem that doesn't exist. Like you're so close to the outcome that it makes the the building of that so much more like valuable. I love that. I want to switch over to talk about surveys specifically in a little bit more you know, more directly. So we're in this moment where folks are talking about AI replacing methods, replacing teams, replacing entire workflows really. And yet you've been really clear that surveys are future proof. So I want to go right at that. If someone says surveys are old, they're slow, they're they're they're on their way out. What do you think that that person is missing? Yeah, you know,

11:21 like it's a big trend these days to call everything that to call everything is ending like oh this is all the the this is this and that, you know, like, and surveys are being around for 150 years and they they've survived, you know, like wars, technologies, whatever it is, they survived. And it's not that that I'm such a big fan of a survey, I'm more a a fan of speaking with clients, you know, like, and I see a survey as a communication tool of speaking with clients in a scalable way.

12:00 That's how I view them. I don't view them all as a boring form of questions or whatever it is. I view them as if you have 100 clients, how are you going to speak with all of them, you know, like and I feel like in future, no matter what, we are still going want to hear back from our clients, no matter what. And that's the reason why I think a survey is essentially going to be around. Of course, they're not going to look and feel how they are now. They're going to evolve as they as they evolved last 150 years. They of course they're going to evolve. They're going to be personalized. You you can do this and that. They're going to evolve, but I view them more as a communication a tool and not a boring form. That's that's what they, you know, mean to me. And I think a business, no matter what they are, they want to speak with their

12:57 consumer, with their client. They don't want to speak with an an LLM is my client okay with me or not. So, I think that's going to be an essential and that's the reason why I believe they are future-proof. Got you. So, it's really about the voice of the customer, customer communication at scale. Is that fair to say? Exactly. Yes. Yes. Yes. Makes a lot of sense to me. Let's get into this a little bit. So, research tech, I mean, we both know it's exploded. It's so prolific. There are more tools, more platforms, more dashboards, more promises than ever. But usually the best products come from one very specific frustration. So, take me back to day one with Picked Adore. What was the problem in survey research that you were most oriented towards solving?

13:45 And what did you want to build differently from other platforms? Yes. So, uh going back those first four years uh that I said earlier were amazing for us because we solved a lot of specific problems for our clients. And then at one point we said, "Okay, we have enough of experience. We want to create a single research tool that covers this all." So, uh what we felt in 2021 uh when we started uh the tool that we that it's the core product of of us now, we felt like uh the research process is broken in five or six tools.

14:25 As we saw back then, so you would so you would have a dedicated tool to script, then you would connect it to a panel or two, then you have a fraud tool, then you have an auto coding tool, then you have a reporting tool. So, we like looked at this and we are like, "Does this really need to be like this?" You know, like and then uh we decided, "Okay, we want to uh design a a powerful tool that can combine all of those in one." Of course, the the the idea behind it was you uh there were specific tools for each of the processes, right? And they are amazing, but you uh uh uh an average client will not need all of the functionalities that that specific tool has. So, our idea was, "Okay, let's get the most important ones, the 20% that consists the 80% of the output I get, let's get the most important ones and put them all in one tool that will

15:23 streamline the workflow, help them uh be more efficient and stuff like that. Because if you use like five, six tools, no way So, you you are you're just like uh increasing friction for you. There is a lot of like, you know, like going between the tools and stuff like that. So, that was the whole idea. We want to cut down the tools. We want to help them uh spend more time in insights, in outputs, activation, stuff like that. was the the the first problem that we wanted to solve.

15:53 I mean, tiny problem. Just kidding. It's a huge problem. Yeah. That's a huge a huge problem to start with. But you're I I totally get that. And I think that like that's a lot of what AYTM is trying to solve as well. So, I mean, I can certainly relate to that. I'm curious to to kind of keep that thread going, when a researcher or a company brings in, you know, an insights team, brings in a new tool, what should that tool make possible on day one? What has to be better like from the get, right, for it to be worth the switch from all the current systems that you're using? Yeah, you know, like it's a research stack is always evolving. I don't like I feel like there isn't there like isn't an end and that dream of us like, oh, you're just going to have one tool. That's just like That's just like a dream. All right, like one tool to run to to do it all.

16:50 So, if I would start now, I'm not sure that I would do it again the same to be fair, like all-in-one. It is It is a hard one and uh the change is significant one. Uh it often takes a long time to process it. Who wants to change all your stack, uh learn a new tool, stuff like that? So, going back to what you asked, I would say for for the first day, try to do a pilot, try to do like something easy, try to, you know, like an essential tool like like everyone ask me who how much the onboarding last and stuff like that. Uh if we need an onboarding if if we need like training of month then our tool is really bad, you know. If if you all need a all hands to show you how to use the tool, then the tool is probably not a tool, it's more I don't know like it's it's it's it's it's really not right? Because

17:47 tools should guide you by yourself and it should make sense uh uh for you. So, I would say that just like uh when when you start, you want to get an easy win, you want to get an easy quick win. So, uh a pilot, uh something small, whatever it is, just to feel it. Yeah, I mean, I think that's good advice in you know, in this world of research technology for people to understand what time to value looks like for any given tool. And I hear you on the I want to acknowledge your point about tools. I think that's so important. At the same time, I sort of I think as like a a researcher, I I think there's a tension between like robustness often and time to value, right? And we have to like allow that the more robust a tool is often the more time it might take to find that value in it. But that does lead me to

18:46 the next topic, which is that you know, PickedDoor's very AI forward. I keep hearing on this show, you know, because we talk to a lot of different people that AI is fantastic at removing friction, but that doesn't mean it understands what matters. So, from where you sit, what parts of the research process really do get better with AI? And where do you still look at the work and think, "No, we've still got to we've got to keep a human in the loop or it's going to fall apart. Yeah. So, we implemented AI like I like when the first wave come in, right? Like 2 3 years ago and we were quite specific about it. We didn't want to force AI everywhere just like that we can say oh we have an agent and you can do this and that. So, we wanted first to solve a specific processes that we that we saw that we could do and where like researchers would spend the most time. The first and like the first one that we saw that is

19:45 huge was auto to auto code the text that comes in, right? Like on open answers stuff like that. So, so that's like a huge one for me because I literally myself was doing that like 6 years ago. I literally was doing a spreadsheet and coding and I remember the days I spent on it. So, it was amazing when we implemented AI and we now you have it all inside the dashboard. It does it automatically for you. It codes it and stuff like that.

20:16 And it literally saves hours and days of work. So, that was like a first one that we made true and it was an amazing one. And then we had now we have like in probably everywhere all sorts of smaller things, but we did try to map again the process how it used to be now just like accelerated by AI. So, we made the the the the flow the same just accelerated by AI.

20:47 Where at So, it helps on for example auto coding, it helps with spinning presentations, some versions of it, of course not the final ones. Uh it helps of course all the well, like conversational AI inside the surveys. It makes the surveys much funnier much much more fun for the respondent to get like deeper insights stuff like that. I must say for me like where it doesn't help that much. I would say if you try to ask AI to to decide instead of you.

21:28 So to extend on that like if you just like ask AI okay design a questionnaire for this and this and that I want to run that. I think that's so lazy and and you could expect that from a non like researcher but if researcher does that like design the questionnaire and a project for me like what's what's the of course co-creation with AI is is something that is amazing, right? Give me ideas for it. Suggest this. Check the conditions. Check this. Check that but don't use it as decision instead of you.

22:07 Like that's that's really not what it's supposed to be. So yeah, I would say like things like that and of course you need to understand how the research works when the when you use it in summarization action points stuff like that. It gives good stuff most of the time, right? Until it doesn't. So So you are there. You are there to catch it.

22:35 And stuff like that but I would say it helps for a lot of things specifics like small ones but don't give it the decision. Don't give the decision to AI. You You still need You still need to decide. Yeah, I think don't give the decision to AI and don't give the question to AI to your point about showing up in kind of a a lazy orientation towards survey design. Like you have to understand your own business question, right? Exactly.

23:06 Yeah. Yeah, I love that. And I think like then to take us to our our next topic, there's something almost bigger underneath all of what we're talking about. If the world is moving towards automation, more generated outputs, more synthetic information, then maybe the real question isn't just what research survives, but what kind of expertise becomes the most valuable? What kind of human expertise? What kind of human judgment? So, I'm curious from your perspective, what kind of human judgment do you think this moment is making the most important? Like what should researchers be sort of crystallizing around? Yeah, you know, like it's a really tough one. I'm to be fair, I'm really not sure what to say, and I keep on like changing my thoughts around it. Yeah, I'm it's it's really tricky one. You know, like I think education is going to change as well with AI. Just the easiness of of these LLMs. Because, you know, like

24:07 Uh technology is used to they haven't So, for example, I was last month like on a trip uh to Croatia back, and I was spending some time with uh my mom and her friend. So, we are speaking about uh the people uh of 60-plus age. And they use AI all day. You know, like and so that was quite astonishing for me. You know, like uh uh a lady of 65 uses AI. It It's gave It's so It's So, it she gave it a name.

24:49 AI has a name for her. Uh it's her friend. You know, like So, a lot of things are going to change. Uh and I'm really not sure what type of skill to be fair, what type of skills we need to but I think first of all, the generic one we need to adapt. We are all going need to adapt how things are. I would say it will be the expertise in whatever we do. We still need to have because at the end of the day, how are we going to judge what AI is saying to us?

25:22 So, but then we we come to that for education. How education is going to work on really the natural if you rely on AI, you really don't know are the correct information or not, right? So, the skills I would say like entrepreneurs will do good because we do adapt. I think researchers will do good as well because our curious want to learn, want to explore, want to question. So, I would all I will always say question the black box because AI is really a black box. We don't know what's behind it.

26:01 The creators don't know as well. So, question the black box, stay curious, and adapt. I would think like you know, like we are all not quite sure what the future holds. I would say at this stage. Definitely, and I think I I love that you're calling out sort of curiosity and adaptability.

26:29 And I think another one that you're not using this exact phrase, but what it is bringing to mind for me is a quality I think of as discernment. Right? The ability to discern discern what's important and what's not. And I'm very curious personally to see how that plays out because I wonder if you can have discernment without the wisdom that comes with knowing all of the building blocks that make research possible. And when we start replacing those workflows with AI, how do you build discernment? And I think we're going to have to figure that out. Like I think that's an open question, which you said, right? Like it is an open question, and we're going to have to figure that out. Maren, we have a recurring segment that we do on the Curiosity Current called Current 101, where we ask all of our guests the the same questions. In your experience, what is one trend or practice in market research that you would like to see stop?

27:25 And what is one thing that you want to see more of? I will start with what I would like to see more more more of. And I would like to see more of collaboration with between the competition. Yeah, I feel like there are industry problems that we that we should collaborate and try to solve together. First of all, the fraud quality inside panels, stuff like that.

27:56 That needs a joint a joint action. I know there are already some actions on it, but I always feel it's a bit like more of a play for marketing than what it really is. Uh so yeah, I just like feel like more collaboration between players because we want to all grow the pie, right? We we want That's what we should aim for.

28:25 The second one what the thing which I would like to see less is and this could be a little bit controversial. I would like to that research tech stays for research, you know, like and let me expand for it. I feel like a lot of times uh companies ask us to be more than research, right? Like they want us to pinpoint exact things. They like almost want us to be consultants, want us solve the business cases, want us to show them this and that.

29:00 But they don't pay us as McKinsey, right? Like they pay us as a research tool. So, let us be the research staff, the tools, uh and let us do our job, and that is doing research about your clients and showing the insights. Uh and let the consultants be them, and you can ask McKinsey that they should do that, right? Like and that's a research tool that you pay like 10 grand per year. So, that's what I'm saying.

29:27 Uh I would say like we all have expertise, uh and we should do what we are supposed to do. Uh that's that's that's my take on it. I like that. So, it's a plea for people to kind of own their expertise and not sort of feel the the need to have to um orient outside of that and for everyone to have to be everything. Yeah, of of course if you can provide an advice, of course if you can go an extra mile. But I feel like there is almost this like research staff should do more of of this. You know, like of course that that that we could aim, of course that we could strive, but you know, like uh it's really not our role and in some cases uh to do that. And it's not fair of you ask of you asking, you know, like so um we all we we all need to know who who plays which role, right? Of course if we

30:24 could go beyond, that's amazing, but yeah, I don't think it's it's fair in in some points how we get addressed. Well, Marin, this has been an absolutely fascinating conversation, and I can't wait to see what happens for you and Picador sort of in the coming in the coming years. So, I'll be watching you. Thank you so much. It was a pleasure to be on. The Curiosity Current is brought to you by AYTM. To find out how AYTM helps brands connect with consumers and bring insights to life, visit aytm.com. And to make sure you never miss an episode, subscribe to The Curiosity Current on Apple, Spotify, YouTube, or wherever you get your podcasts. Thanks for joining us and we'll see you next time.

Produced by aytm

Curiosity Current is made by aytm, the consumer-insights company. Guests speak for themselves; the synthesis is ours. About aytm