Statistical Rigor in Social Insights with askpolly’s Erin Kelly

Social media conversations offer a wealth of consumer insights, but separating meaningful data from online noise requires more than traditional social listening.

In this episode of the Greenbook CEO Series, host Karen Lynch sits down with Erin Kelly, CEO of askpolly, to explore how the company is bringing statistical rigor to social media research. Erin explains how askpolly uses randomized sampling and AI to turn organic online conversations into measurable consumer insights, helping brands distinguish genuine sentiment from the loudest voices online.

They discuss the company’s integration with Microsoft Copilot, real-world examples of brand controversies, and how faster, more accurate measurement could reshape market research. Erin also shares her experience raising venture capital, navigating challenges as a female founder, and building a product that serves both researchers and marketers.

From overcoming startup hurdles to embracing new technologies, Erin offers valuable lessons for researchers and the next generation of industry leaders.

Key Discussion Points

• Beyond social listening: How askpolly uses randomized sampling to generate statistically valid consumer insights from organic social media conversations.
• AI-powered research: How askpolly’s integration with Microsoft Copilot makes complex data analysis faster and more accessible to researchers and marketers.
• Separating noise from reality: Why viral backlash and negative social media sentiment don’t always reflect how a brand’s actual customers feel.
• Building and funding a startup: Erin’s experience developing askpolly, securing venture capital, and navigating the challenges of entrepreneurship as a female founder.
• The future of market research: Why speed, statistical accuracy, and the ability to measure real-world behavior will become increasingly important for demonstrating marketing effectiveness.

Resources & Links

• ⁠askpolly

You can reach out to Erin Kelly on LinkedIn.

Many thanks to Erin Kelly for being our guest. Thanks also to our production team and our editor at Big Bad Audio.

Transcript

[Karen Lynch] Hello, everybody. Welcome to another episode of the CEO Series. I'm Karen Lynch. I'm so excited to be hosting today and talking to Erin Kelly, who's the CEO of askpolly. I came across Erin a few weeks ago, I think it was early September, when a news article caught my eye that I was going to talk to Lenny Murphy about on The Exchange, about how askpolly had raised three million in seed funding to accelerate their consumer insights engine for a company that was kind of doing something very different with social media measurements. So, I'm going to let Erin tell you all about askpolly. But first, I want to just take a pause and say Erin, welcome to the CEO Series.

[Erin Kelly] Thank you, Karen. It's great to be here.

[Karen Lynch] I'm so glad you are. And for our listeners, she's joining us from Canada, and we are talking to you from the East Coast. This episode will be airing in a couple of weeks, but there's been a lot of developments even before this seed funding was announced. So Erin, why don't you just level set everybody, tell us all about askpolly, and then we'll talk about how askpolly even came to be. But first, tell us a bit about the company.

[Erin Kelly] Sure, great. So askpolly, we're the first company in the world able to do statistically valid measurement from social media, organic social media conversations. So, we don't put surveys in front of people, we look at what they're naturally saying. So like, you know, anthropological research, if you will. But the key difference is it's not social media listening, it's measurement. We’re doing a randomized controlled sample of the population. So just like you would do with a properly done phone research, it's very, very, it's a very, very difficult technical feat to do a random controlled sample on social media  because as you probably know, networks, computer networks don't like to be random, just by nature, right. So, it was a five-year research project on how to develop that algorithm, conditional independence coupling, that allows us to match census data with any, well, we have countries all over the world that we're measuring now in Europe and North America and South America, and Asia, and all of that. We speak all multiple languages. So you put your question in, if you're a researcher, just like you would in a survey, and you tell it what you want to measure, and within minutes, it has a full statistically valid market research with a confidence interval, a margin of error, it's got all the demographic breakdowns, you can do any audience, and it's the first zero-shot, it was the first zero-shot AI in the world when it launched, which means it doesn't need to be pre-trained. So you don't have set things that you could measure, you free type any audience you want, any diverse audience, and any topic.

[Karen Lynch] So when you say social media, like, where's it pulling from? Where is this population?

[Erin Kelly] So right now, we're going across six different social media platforms, and we're adding more all the time. So at the moment, we're on Reddit, TikTok, X, Blue Sky, Threads, Mastodon, you know, a number of them, and we're adding in LinkedIn and YouTube, so we'll be able to get all of that data. And so, we’re going across all of the major ones, so we get all of those demographics, if you will.

[Karen Lynch] You have to tell me, what was the aha? Well, actually, before I even ask you that, tell me what you were doing before, because I want to kind of get to the aha that made you say, this is what we're going to work on, but what were you doing before that even led you to this point? Take me back.

[Erin Kelly] Yeah, so it's kind of interesting, my business partner and I both came from, like, measurement backgrounds, but different things. So I was in agency, and we were doing market research, and as you probably know, because you come from the same background, as I was doing it, I was realizing more and more, it's not precise enough, we're not able to… and our customers were pushing back on us, you know, I don't believe this research, you can't prove that it was your campaign that led to these sales, you know, that sort of age old thing that marketers go through. So I was starting to have… I actually started a company on my own to do social media research, but I made the same mistake that, well, I consider it a mistake now, that everybody else is making when they did social media listening, not doing that proper controlled sample. My business partner actually came from an academic background. He left academia to join the company, but he was working at the university, the local university, and the Department of Social Science came up to him, and they wanted to study suicide research. If you look at us, we do a lot of work in medical. Now, you can't do suicide research on the phone for a whole host of reasons. You can imagine how difficult that would be. But they wanted to know, can we get proper signals from social media, because that would be the place to go. And he started tackling that project. And that's what led to him realizing after he tried to do what everybody else is doing with listening, that unless it's statistically representative, you couldn't do anything with the data. You had to be able to use statistics, and therefore you needed proper sampling. And so that's how… and then he and I met, we were actually put together by people in the industry saying the two of you need to come together, we did that, and that's when the company started.

[Karen Lynch] Yeah. So how long has it been now? Like, how many years are you at this?

[Erin Kelly] We're four years post research, right, where we've made it something that customers can use. And, you know, and there's been a lot of, as you can probably imagine, making it easy to use was another challenge because our customers aren't necessarily statisticians. Some of them are, but some of them are also just marketers and things like that. And so we had to make a product that would be easy to use, easy to interpret. And our partnership with CoPilot, with Microsoft, has made that a lot easier.

[Karen Lynch] Yeah, talk to us about that. So again, it was early September that I saw about seed funding. And I was like, look this, and we'll get to that in a minute, because I was really excited that there was another female founder on my radar. Anybody who's been listening for a long time knows that is something that I am personally passionate about, because there are fewer of them in our industry than there are males. And I personally understand the challenges. I was never a female founder. I consider myself more of a female solopreneur. You know, I had my own company. I was a company of one. So I understand that world a little bit more. And until we reach, you know, full equality, I believe there are specific challenges that will always be on my mind. But then since then, when I was doing some research, I see, you know, two weeks ago, you have this CoPilot announcement. And I was like look at that. That does change the accessibility to your platform. So tell us a little bit about that and what it's enabling for your users, for the companies or the enterprises that are now working with you.

[Erin Kelly] Yeah. It's been actually a game changer because we always had the challenge with the product that you wanted to make it sophisticated enough because it's a very sophisticated product. Like with one question, you can get 10,000 permutations of data, right. And our research partners, like, you know, we have Harvard as a customer and all of these, they love having that rich data. But a lot of our customers just want, just give me the insights, right? I don't want to look at 50 graphs. I want you to tell me and make a nice report and put it in a PowerPoint and all of that. It's difficult for us to put things in PowerPoint. We don't have the licensing, whatever. So having the partnership with Microsoft where they, because it's through Microsoft and through CoPilot now, customers can take their data, put it into the Power Platform, Power BI, all of those things, but they can also take it into CoPilot. And you just ask in natural language, here's my study. I asked these questions. Please go through, and here's what I consider insightful. Tell me, is there a difference between men and women? Is there a difference between different regions and how people feel? Income levels, zip code areas, you can do any audience, like do vegetarians feel differently about this than do parents feel differently? So it will go through and crunch all the data for you, pull up, this is what, here's some really interesting things. Like it literally is like the consultant that you get when you go through an agency. And then you go, okay, I'm interested in that. I'm interested in that. Please give me more. And then when you're happy with it, it'll put it in report and put it in a PDF for you, if you want. It’ll make a PowerPoint for you with all the graphics. It will write a press release for you with the details and all of that. And this is all minutes, you know, the research is done in under five minutes. And then you go into CoPilot, you can have the whole thing done in half an hour, which is great. Like when I'm doing things for the news, for news pieces, they'll ask me, can you get me this, this week? I'm like, I can get you that in 20 minutes.

[Karen Lynch] Yeah. So that's really cool. An interesting use case to think about the news too. So, I have a question for you because as I mentioned, as I said, my career prior to Greenbook will come up during this conversation. As I mentioned, I was a qualitative researcher for some years before joining Greenbook and I had done some social media mining to sort of enhance some of the work that I did. So, if I was working in a category, this was very rudimentary, you know, back, I'm talking about maybe 15 years ago, you know, if I was working in a category, I would do some very basic searches to see what was being said about that category out there in the world, using some tools that were available at the time, nothing specific. And so when you talk about kind of asking a question versus what I might've done, you know, 10, 15 years ago, which is searching for a topic, my brain is kind of grappling with how can you question, how can you ask the question when what I'm used to is querying a topic? Can you give me some examples of questions that you can ask versus searching for topics? Does it work just like any other kind of search would be that answers are given to you, kind of forcing connections to the questions that you're asking. Just maybe give me some examples to bring it to life for everybody.

[Erin Kelly] Yeah, for sure. So, you can also use askpolly to design your research for you. Because that's often the hard part. Right. And I say to people, the kinds of questions that we used to ask were actually proxies for what we really wanted to know, right? I don't want to know how you feel about green or about almonds. I want to know if you're going to buy my product or what I have to say to you to have you buy my product, right, at the end of the day. But I have to ask like 20 questions. I have to beat around the bush because I can't just say, what will get you to buy my product? Now, you can ask that of askpolly. You put into askpolly, you say, my goal is, because sometimes it's not always to get people to buy your product. The goal might be take up a cause or believe a certain thing or what have you. Right? So, what is it going to take to have somebody to achieve this goal and what askpolly can do, we have message testing workflows. We have campaign impact workflows. So, I'll take campaign impact for a minute. You don't want to know, do you recall seeing this? And you know, people don't recall things very well. You want to know, did people see my ad? And if so, did they change their behavior because they saw my ad, you know, did they start recycling? Did they, you know, buy my snack? So askpolly can actually do that. And you just say, what effect did this campaign have on recall or behavior change? And what askpolly does, is she sets up all the research for you in the background. So, let's take an example. We did some work for the COP conference that was all about the environment. Remember a few years back? And all the governments were talking about greenhouse gases and all this. Well, it just so happened that the Canadian government put out a bunch of ads, poorly timed, right at that time. And then they wanted to know how effective were the ads? Well, everybody was talking about the environment at that time. Greta Thunberg was talking about it. Greenpeace was talking about it. Everybody was talking about it. So, how do you know what effect the government ads had versus everybody else talking? Well, what askpolly will do, and she does it all automatically. She tells you what she's doing so you can change the parameters if you want, but she will go out and find identical samples of people who were exposed to the ad and weren't exposed to the ad. How does she know if they were exposed? Because some people, not everybody says, I saw this ad.

[Karen Lynch] Yeah.

[Erin Kelly] Well, the whole world was exposed to the COP conference. Australia was, the United States, the United Kingdom, but only people in Canada saw those Canadian ads, so she looks for a signal of people who had the same amount of chatter about environmental things before the conference. And she got a sample in the United States, a sample in Australia, in England, just to be very safe, and then one in Canada. And then she looked to see, okay, the Canadian conversations went up the same as the others in Britain, but when the ads ran from the government of Canada, there's a little bit more. So she can say that that's with, again, you do, what are the chances that it wasn't those ads? We've got the exact dates, and we have the messaging of the ad. So she does those types of measurements. And then she can also say whether or not people took action because of those ads, because now she can isolate those people. So campaign, and same thing with the messaging, people will put in messages that they are thinking of putting out there. So you do the testing before you put it out. And she looks to see what other messages like that from other organizations the audience has been exposed to, and how they reacted to it. What snarky things did they say about it? Or were they like, yes, I believe that. And she measures it not just demographically, but in terms of the lift, like are they sharing it with their friends? And the persuasion, did it persuade them to say, oh, that's a good idea, or what have you. So she measures, and she also, when it's negative, she'll flag that too, and tell you why it would be received negatively, so that you can either correct it, explain it, or take it out.

[Karen Lynch] Yeah, one of the things that I think I read, you know, scouring your LinkedIn, kind of getting to know you a little bit better virtually before we actually had this conversation, is you're also sort of solving for the loudest person in the room bias, right? There's a lot of that on social media in particular where, you know, there's a lot of biases in qualitative research, right? And I've worked to overcome those throughout my career, and learning how to tune out the loudest person in the room in person. But in the world of social conversations, you know, that happens, and yet you really work very hard at this unbiased audience sampling. So let's talk a little bit about that, and kind of tell me how you recognized that, and knew that you had to come up with a solution for that.

[Erin Kelly] Yeah, this one's really fascinating, and I'll give you a concrete example. We had an example here where the government wanted to cap nurses’ salaries at one percent, and they wanted to know how the population felt about it. Now, if you're using social media listening, here's the problem. If you're just grabbing the first 10,000 posts, what would happen to the Nurses Association told all the nurses, go out and complain about this legislation. So, if you just grab the first 10,000, you are very over-represented in nurses, and guess what, they don't like the legislation. So you're getting like 98% hate the legislation. What we do is we start with people, just like you're supposed to do in market research, a randomized controlled sample. So we've got bankers, truck drivers, everything. Now, it was still unpopular. It was not long after COVID, and so people weren't happy, but you get a truer measurement of how people feel, not just how nurses feel. Now, here's where this is really interesting. It's interesting in that context to make sure you're getting an unbiased point of view, but it's also really fascinating, because we do capture the bias, and we put the bias in a bucket, and then we compare it to the unbiased. And here's what's really interesting. You really get to see companies, and companies where their marketing is in the biased portfolio, and then you see what the customers actually say about them. And it's really fun to see that pretty much as a rule, the companies that have great alignment between their marketing and what their customers say are more successful companies. And then you see companies that are in financial trouble, and you look at that bias and unbias, like you'll have a quick serve restaurant that’ll say, the kids love it, and the parents are saying, no, my kids don't like going there. They prefer to go over here. So figure out what your customers really like and promote that, as opposed to promoting something that they don't believe, and then they just ignore you. It's an interesting measure. And you see this all the time with the government. The government always misreads the population. It's crazy.

[Karen Lynch] Yeah. Well, and we probably have to table a lot of government misreading the population conversations for another conversation. But not that long ago, we were talking, I think, on a different podcast or on a different LinkedIn post about, recently, there was a very controversial ad that was viral socially, had a lot of negative backlash repercussions. But I'm not sure that it was negative for the intended audience. I think it was negative for an unintended audience. It was a sports betting app that used Sydney Sweeney in a very provocative way. I don't know if you know the ad I'm talking about. Female athletes were outraged because she was using her body in a way that wasn't really promoting female athleticism. It was really objectifying women. And yet, there were arguments all along saying that wasn't the intended audience. The intended audience was male sports betting. So, as you're talking, I'm thinking about that. And I'm thinking, does it matter to the brands if the chatter and the social story doesn't affect their sales in the end at all? Is that even of issue to them if what they're finding isn't relevant to them? Do you have situations like that where the social story that's going around that you're able to measure doesn't have an implication at all? Does that happen in your world? And I don't mean as controversy, but I mean that, yes, there's lots being said about it, but it doesn't have an ROI or it doesn't have a negative impact. It's just those two things are coexisting.

[Erin Kelly] Yeah. So if we take, for example, Cracker Barrel. Do you remember the Cracker Barrel thing?

[Karen Lynch] Yes. The logo. Was it the logo?

[Erin Kelly] Yeah. So we actually did that measurement. And here was the mistake that Cracker Barrel made, I would say, because it is important to understand whether your audience and your customers are complaining or other people are complaining who are never going to eat at Cracker Barrel anyway. So we created an audience, do you eat at Cracker Barrel. And we don't ask that… like we're not asking people that question. We're getting askpolly to go out and make a sample of people who have said, went to Cracker Barrel, took the family out to Cracker Barrel, so that we have good confidence that they have been to Cracker Barrel in the last year and to measure the frequency of the likelihood that are they regulars? So, she can actually group you as like, I love Cracker Barrel or I sometimes go there or what have you. So, she's measuring that. And then we compared that group to the group that was complaining about Cracker Barrel. And can you find evidence that these people have gone to Cracker Barrel? Now, some people go to Cracker Barrel and they don't. And we did it to the general population. And what we found was that the people who liked Cracker Barrel and went there either occasionally or often wanted them to change that. Wanted them to update their logo. They were in favor of a refresh. They're like, it's time. Right. And the people who are complaining were actually people that, you know, and it's not just because they don't say they go to Cracker Barrel. We look at, you know, the kind of, you can do with high statistical validity, the kind of dining people do, but we measured it against the general population. And it was really a minority of people that were against them updating their logo. So, and for various reasons, even the people like their people go, it's up to the company if they update their logo. Like I don't care one way or the other, but that's the company's business. But the customers actually were in favor of them changing it. And we did the Sydney Sweeney commercial and not the one you're referring to with the jeans.

[Karen Lynch] Remember? She had good jeans. Yes.

[Erin Kelly] Yeah. And it didn't have a negative effect on the brand.

[Karen Lynch] Yeah.

[Erin Kelly] So, in fact, it gave the brand a lift. But I can tell you a brand that really had a bad impact was Bud Light and they have not been able to recover. In fact, the more they try to recover, the more they're digging themselves in because now they're alienating everyone, right? They alienated the alpha males and then they look like they're groveling when they try to get them back. They sponsored the wrestling on the White House lawn and that caused people to say, wait a minute, they're not even American. Why are they celebrating, you know, American birthday? Right. And then that affected their Olympic sponsorships as well. So like, it's just every time they do something, they just dig themselves further and further. So that is a brand that really needs to take a breath and come up with a strategic plan.

[Karen Lynch] Maybe just go back, like, figure out who you are, right. Or go back to your roots or something like brand identity work, perhaps. I'm not sure.

[Erin Kelly] Do lots of market research. Because like they can't just keep trying things.

[Karen Lynch] Yeah. It's interesting though, because especially where we stand as kind of market researchers who are very tapped into marketers also, it's interesting to think about our own influences. You know, I find out about a lot of this stuff on LinkedIn where there are very opinionated marketers, right, in my circle who have a lot to say about these marketing decisions from a professional standpoint, like in the case of the Cracker Barrel logo redesign, there were a lot of really opinionated people bashing it, that decision from that professional standpoint. And so it's really interesting to think like, you know, yes, not only are you not necessarily the target, but you are not the customers. We in the professional space, weren't tapped into the consumer conversations, the mainstream user conversations happening in other places around the internet. So I think that that's also a great learning lesson for those of us who, our social network might be LinkedIn. Think about either way, we are a microcosm of the greater social conversation, so it's really interesting to think about where else you're pulling conversations from.

[Erin Kelly] One of the things I want to address too, is that if you're not properly sampling your data, the internet looks very negative. And that's because the negative people post more often. So we get the impression that everybody on social media is negative. It's not true. We've done that research and not just us, the university has done that research and published. When you correct for the number of posts, when you put the trolls down to one vote, if you will, there's lots of normal people and lots of positive people online, but they say their thing and they go away because they're not obsessed. But you get these obsessed people who tend to be more negative. And I think if you're a brand, you have to be very cognizant and careful of that because it looks like everybody's ganging up on you and on your brand, but they're not. If you do the analytics, it's like 10 people.

[Karen Lynch] Well, it's good that you're educating people as well in the work that you're doing. So, you mentioned the university and I want to go back to your founder story again, because you were working with someone in academia as well. Let's go back to the origins then. What were some of the, first of all, when you decided to go for this, when you decided to do this, what were some of the early hurdles you had to cross over? Some of the early things that were rough. And the reason I ask these questions of people when they come on this show is because there are other people that listen that either are facing those right now. I know that there are some startups that are paying attention that might be having those struggles right now and you've already overcome and gotten some funding. So, what are some of the things that you've overcome so that we can start to maybe inspire people? Maybe there are some people that are like, what do I have coming down the pike that I should be ready for? So, talk to me about some early challenges you've already addressed and let's offer some guidance to folks.

[Erin Kelly] Sure. So, the big challenge that we had was defining our audience, for one thing, because is your audience the analytics person? Is your audience the marketer? And they're very different audiences. Even if they're working for the same brand and at the end of the day, they're trying to achieve the same thing. One wants to work with numbers and graphs and the other one does not. And before we had the LLMs to bridge that divide, we went through so many different designs of our product. You wouldn't imagine. We were designing and redesigning. We didn't want to do the professional services. We didn't want to go down that route because we knew we wouldn't get funding. Your product has to stand on its own and even more so now, product-led sales and product-led growth is everything. And market research is traditionally a professional services environment, right? So, you're taking something where the customer is not used to doing their own. Now, there's precedents for that. I mean, it used to be you went to a travel agent to book your holidays. Now, a lot of people book their stuff online. They don't go through a travel agent. Some still do. So, it's a hybrid. But market research has always been considered very technical. It is. And that's the other problem is explaining to people why you want to go with our solution, which is doing that proper sampling versus there was a tendency to think, well, you're all the same. You're all doing AI on social media. It's not the same. Synthetic data is another layer now that we have to keep explaining. So, finding your blue ocean or explaining a very technical product to a non-technical audience because the buyer tends to be the marketer at the end of the day. The insights person will be the recommender, but they go to the marketing team to get the budget. And so, that's all very challenging. It's figuring out your user versus your buyer and having to explain to both of them why this is superior without getting too technical. Nobody wants to hear you drone on about sampling and probabilistic samples, right? So, marketing the product, taking it to market, that was difficult. And we had investors who it's like writing a book and sending in your manuscript. You get a hundred rejections before you get somebody who says yes. So, literally, I'd be pitching until I was hoarse. And then you get your first few ones. And once you get your first one, it becomes easier to recruit more, but it's getting that first lead who believes in you.

[Karen Lynch] Yeah. Yeah. Yeah. That's great. Let's go to the place where I said I wanted to go, which is doing this all while you were, we've talked, you have college age children, approaching young adulthood, not easy to do all this. I imagine as a female balancing demands of motherhood. I know that you have a partner, but there are unique challenges. It's a lot. And don't try to tell me it's not because I know it is. So, tell me your experience. Has it been more of a struggle? Do you think your experience is shared with other female founders out there? Do you think you would stand in that camp and say, yes, we are facing challenges that are unique, or do you think that you've come a long way? What are your thoughts on it?

[Erin Kelly] I think we have to stand up and say there are challenges that are unique because we're seeing backlash against women leaders now in the culture. Especially you look at places like the United States and things like that. And the way they work against women is to play on the guilt. And that's why when I'm sitting in meetings with women, when I see a young woman who says, I feel guilty, I worry, because the guilt is always, oh, if you put your child in daycare or somebody else watches them, they're going to be damaged. And I love to hold up my kids as an example. They're doing great. They're in university. They're independent. We have a great relationship. They're not unsure of who their mother is. And I like to say to young women, you know what? I spent so much time worrying about these things. And it was ridiculous. I have a great relationship with my kids. They're doing great. And I can tell you, I mean, it's anecdotal. I know lots of people who stayed home who don't have that great result. So, it's not because you worked or you didn't work. There's lots of confounding factors. But you can be confident. You can run a business. You can travel. You can raise venture capital. And you can still have children that not only are they not damaged, you've inspired them. And you've shown them you can stand on your own and you can do these things. And so it's actually, I would say, strengthened the relationship because my daughter is very ambitious as well. She's in engineering. And I think she likes to talk to me about the things she's doing. She knows I'll understand because I've been in a similar field.

[Karen Lynch] Yeah. Yeah. By the way, I'm just going to be smiling a long time for the fact that you're like, you can do this. You can do that. You can raise venture capital. Like it's just another thing to add to the list of strong working women. I really love that you've done that. It's great. It's great. Tell me about finding it with a male partner. Helped? Has that been an asset? Has that been just a no-brainer? Has it really factored into it at all?

[Erin Kelly] Well, it's really interesting because we kind of use the dynamic sometimes to see who are the misogynists in the group that we meet with. If I send stuff out, I'm the CEO of the company. I'm definitely the leader of the company. He's more on the back end of things. He's not really a public face. He's not that type of person. And you'll see people go, yeah, she's the CEO. We're going to talk to her about financing and all that. But then you get the people who just want to talk to him. We definitely saw that. And those are people we don't actually… it’s actually a good filter for us because we're like, yeah, you know what, this is kind of offside and probably indication of things to come if we're partnering or hiring or whatever these people.

[Karen Lynch] It's one of those things where it's good to kind of back to know who you are as a company, know your company values too. So it's nice to, I think, probably have that and have visibility into others for partnerships. So pretty neat. So what's your kind of thoughts on where the industry in general is headed before we wrap? You kind of mentioned being product led right now. So I know that that's one thing that's on your mind, but where do you see things headed right now? What's your glimpse into the future?

[Erin Kelly] I'm actually very excited about it because, you know, as I mentioned, like in the past, we've had to do research that was a proxy for what we really wanted to know.

[Karen Lynch] Yeah.

[Erin Kelly] Market research is actually going to become even more relevant now with the technology that's available. I know you read a lot in market research with response rates and stuff like that. Listen, go to where the conversations are happening. They're happening online. That is clearly, to me anyway, the future of it. We published a paper, not us, the Public Health Agency of Canada with a number of universities having used our tool to track foodborne illness in Canada. It was a first in the world. You know how in the past Google tried to do Google flu and things like that and it never really panned out.

[Karen Lynch] Yeah.

[Erin Kelly] But now with randomized controlled sampling applied to social media, it proved that not only was askpolly able to correctly identify a salmonella outbreak in a population, but the AI was able to do it faster than even wastewater monitoring, before wastewater and five weeks before anything else the government had. And you take that power and you apply it to market research. We can now tell you, was this campaign effective? A competitor has come into the marketplace with this price. How should you react and get a very precise measurement. Should I react by changing my price, by changing my messaging? And you can measure all of those things very quickly. My sponsorship, when I sponsored this and that, you can now measure very precisely that this sponsorship… and it doesn't have to be digital. This bus ad did more for my brand awareness and maybe buying than this thing that I sponsored on this website or TV ad. So, we can get very, very precise now. And I think market researchers that embrace that future where we're going, there's going to be exploratory research, which you can do with synthetic AI, but you can't do measurement. You can't say what effect did this scandal that got reported in the media have on our brand? How should we react? You have to be able to do that with measurement and measurement is getting really powerful now. So, I actually think we're going to be able to show the value of marketing, prove the value of marketing and of market research. And so, I actually think there's going to be, and there is, I'm already seeing it, high demand. Now, people want to do it themselves. They want it immediately. Like we talked about the news earlier. If something breaks on the news, you need to answer by six o'clock, right? Or as soon as possible. You can't let the thing drag on for a week before you respond. So, being able to figure out how to respond most effectively, that's what customers want. So, they want speed. They do want it to be accurate. They want to know they can trust it. I think that's a big theme right now. How do I know I can trust social media data? It's biased. So, you have to discuss about how you strip the bias out of it. The questions I get all the time, my aunt's not on social media. Yes, but she probably wasn't in the phone survey last week either. And that's why we sample, right? So, explaining those things. You don't have to explain it when you do phone research because it's so well known that people just go, I'm sure they've worked at the kinks. But with social media, you do have to explain it and it's more challenging.

[Karen Lynch] You know, it's so interesting. My brain is exploding with another thought too, which is sports fandom. No surprise to people. I'm a Mets fan. I have Mets stuff on my desk in front of me. I've interviewed somebody from the Mets organization on our other podcast, The Greenbook Podcast. And this year was very controversial for the Mets. There was some off-season decisions made that really had a negative impact on the fans. Social media chatter was quite negative, to say the least. I'm sure ticket sales were down at the beginning of the season. I myself was like, not going all season, not going all season. They have to get to 500 before I go. Come my birthday, my husband's like, what do you want to do? And I'm like, I want to go to a Mets game. So, so much for that say-do gap, right? I want to go to a Mets game. The next I'm putting pictures on social media of me and my children and my husband at the Mets game, having the time of my life in my Mets shirt. And I'm sitting there thinking, there's so much to be gleaned even from fandom, right? Where we might be bashing our team, furious with our team. And then the next thing you know, we're sharing a picture because we're there at the game, talking about what a good time we're having. And a platform like yours can probably suss all of that out as well. Right?And close that say-do gap on some level, or am I extrapolating something that is not accurate?

[Erin Kelly] You're absolutely right. And I'm going to give you a little stat, which I found very sad. Well, kind of hopeful as a Canadian, but sad for American girls. We talk about women and stuff. I was measuring, I'm a hockey fan. I mean, I come from Canada, so.

[Karen Lynch] Oh yeah, of course you have to be a hockey fan.

[Erin Kelly] Yeah. I feel like from the youngest age it’s all you think about. And we had our own scandal in Canada, hockey, right? Hockey Canada had a scandal. We won't go into it. But in the Olympics, as you probably remember the four nations, we won that one. But then in the Olympics, we came in second, both the men and women's team. Now I was tracking girls, I was doing a study on girls and their interest in hockey, and it's been going up and up. And especially as we came to the Olympics.

[Karen Lynch] Yeah.

[Erin Kelly] American girls and Canadian girls actually growing faster than boys. That's because all the boys in Canada play hockey already.

[Karen Lynch] Yeah.

[Erin Kelly] But here's what was really interesting. You see American girls and Canadian girls at almost the same rate as we go up toward the Olympics, being excited and saying, I want to play hockey because look at the women's hockey, you know, United States and Canada were in the finals, right? So it's very proud. Canada came in second, women's hockey, we came in second, the Americans won, but you remember what happened? Trump said, oh, I guess we'll have to invite the women. And he invited the men to the white house and the women went.

[Karen Lynch] Yeah.

[Erin Kelly] You just saw girls, even though Canadian girls, the women's team came in second, Canadian girls, proud, proud, proud joining hockey. And you saw this little, they just, the line graphs went in opposite directions where girls were deflated by that event, even though their team won gold. It's that controversy afterwards that brought down morale. Now it has since recovered, which is good, but it just gives you… And I actually gave a presentation on that, and I said to Canadian audience, this is actually our opportunity because hockey players are made 20 years in advance, right? If we want to win gold in 20 years, it's the girls today that we've got to get into hockey and Americans have just lost perhaps some stars from that event.

[Karen Lynch] Well, sidebar, my son is actually a big PWHL fan and we are New York Sirens fans. So, he is a season ticket holder. That was his college graduation, when he graduated last year, that was his graduation gift, was season tickets to the New York Sirens. Our whole family went to the big game that they had at Madison Square Garden last year. And it moved me to tears to see these young women talking about fulfilling their dreams of playing in Madison Square Garden. So really, really amazing. And just kind of watching all of these women in that league has been one of the greatest sports kind of accomplishments of my generation, which I'm sure for you as well.

[Erin Kelly] That's wonderful. And here in our team, the PWHL plays in the same stadium as the men.

[Karen Lynch] That is amazing.

[Erin Kelly] Starting this year.

[Karen Lynch] Yeah. And we have, you know, for us attendance is not as great as it is in Canada. So we, as fans, we always have that caveat, like, you know, most attended game here in the US.

[Erin Kelly] Well, thank you for supporting hockey.

[Karen Lynch] Absolutely, absolutely yes. You know, I may be a baseball fan, but hockey is our second favorite. And I have gone to more PWHL games in the last year than I ever imagined because of these season tickets though. They’re really quite fun.

[Erin Kelly] That’s awesome. So glad you're joining PWHL. We do the Ottawa Charge here. So yeah.

[Karen Lynch] Nice. Nice. Well, I will tell him that as well, probably the next time I see him, I will say, guess what? So I will say, I have a new friend.

[Erin Kelly] Yeah, absolutely. I want to come. I want to come and join you.

[Karen Lynch] Oh my goodness. Wouldn't that be great? Yeah. There's a few games where they said the arena is, you know, to be announced, so our fingers are crossed that it's going to be MSG again, because the only reason they wouldn't announce it yet is because they're working on those details because right now they don't play their full time, but it would really be so great. The girls deserve it. They deserve to play. They really do. So they play, anyway, it’s a total sidebar. They play where the New Jersey Devils play. Fine, but they're not as good a team as the New York Rangers. So anyway, enough about sports. Erin, is there anything that you wish I'd asked you on our time together before we wrap?

[Erin Kelly] No, I think we covered a lot.

[Karen Lynch] We really did. I'm so glad that you were here. What would be your final piece of advice that you would give anybody, you know, kind of, that would be a next generation leader, a next generation founder who's coming down the pike, final word of wisdom for them.

[Erin Kelly] Stick it out. I've been through a lot of troughs and if you can stick it out through the troughs… I think, a lot of founders give in too early. I think they give up too early. And I heard that advice once many years ago, and that got me through hard times and it really did work out. Now we're integrated with Microsoft or, you know, there were times when I didn't think that would ever happen. So it's stick it out. That's all I can say.

[Karen Lynch] That's great advice. So to all of you listening who are heading on this journey or heading down this path, stick it out. I'm with Erin on that one. Well, thank you so much for joining me, Erin. What a pleasure to speak with you.

[Erin Kelly] Thanks, Karen.

[Karen Lynch] And to everybody listening, thank you for joining us once again. To Emma, I appreciate everything you do to produce these episodes and to Big Bad Audio for cleaning us up. I appreciate you as well. We will see you all next time on another episode of the CEO Series. Bye-bye everybody. Take care.