BOUNCE - Charlie Butler (Co-CEO). Getting to the best decision at the right speed; Balancing data retrieval & fresh research; Why researchers still matter.

1 - > SPEAKER_00: No other industry is like it when it comes to market 2 - > research. 3 - > When it comes to stubbornness and resistance to change and 4 - > things being the way they are, you could spend all your life 5 - > complaining about why things are the way they are, or you can put 6 - > yourself in the shoes of the researcher and change, manage, 7 - > and articulate a better future. 8 - > And I think the reality on the ground is that humans make 9 - > decisions with so many biases that all you're looking for is 10 - > enough information to deliver enough confidence based on the 11 - > speed required.

12 - > The incentives for synthetic to or AI moderated call to work are 13 - > so high that I think it skews the human dimension. 14 - > So I think it skews humans' belief that it will be true. 15 - > When you want something to be true, you are more likely to 16 - > believe it will be true. 17 - > SPEAKER_01: Welcome to FutureView and a conversation 18 - > I've been looking forward to sharing for a while.

19 - > So Charlie Butler is the co-founder and co-CEO of Bout, a 20 - > business that seems to have achieved the rather unusual feat 21 - > of getting researchers, technologists, and AI 22 - > enthusiasts all excited at the same time. 23 - > Having started this company straight out of uni, Charlie and 24 - > his team have built one of the most innovative businesses in 25 - > the insight space, taking a very different view of how research 26 - > should work. 27 - > In effect, helping clients leverage existing knowledge and 28 - > then filling in what you might think of as the gaps with 29 - > automated research tools and partnerships with specialist 30 - > providers.

31 - > And given that Bounce works with organizations like DiAgio, 32 - > Coca-Cola, Pono Rico, and Tesco, it clearly seems to be working. 33 - > One of the other things I really liked about this conversation 34 - > though is that Charlie doesn't really come across as the 35 - > stereotypical technology founder. 36 - > Before Bounce, he founded a mental health charity after a 37 - > career-ending rugby injury. 38 - > And throughout the interview here, there's a real emphasis on 39 - > people, empathy, and helping researchers become influential 40 - > rather than replacing them.

41 - > As you'd expect, we do spend quite a lot of time talking 42 - > about AI, knowledge repositories, synthetic data in 43 - > the future of insights. 44 - > But we also ended up discussing entrepreneurship, raising 45 - > venture capital, leadership, optimism, fiction books, and why 46 - > most business books could probably be a lot shorter than 47 - > they are. 48 - > So, Charlie, firstly, thanks so much for joining today. 49 - > Really delighted to have you on the podcast.

50 - > Thanks for having me. 51 - > Not at all. 52 - > Now I wanted to get going with the traditional icebreaker. 53 - > So it's something that doesn't have to be deepest, darkest 54 - > secret, anything like that.

55 - > It's just something that people might not know about you unless 56 - > they know you very well. 57 - > SPEAKER_00: I suppose, well, people in the industry won't 58 - > know that the first entrepreneurial thing I did was 59 - > when I was in university before setting up Bounds was actually 60 - > set up a mental health charity, a men's mental health charity 61 - > when I was 18 or 19 and really struggling myself and with a few 62 - > friends who I opened up with, and we realized that was 2018 or 63 - > so, and very little public support systems and networks and 64 - > kind of peer-to-peer help, in particular in university.

65 - > And it kind of came off the back of kind of a career-ending 66 - > sports injury I had. 67 - > And I kind of had this loss of identity and not really knowing 68 - > what I wanted to do and struggled a lot, and then ended 69 - > up spinning up a mental health kind of community called Tribe, 70 - > which ended up raising 150,000 euro while I was in university 71 - > for kind of suicide prevention and youth mental health 72 - > services. 73 - > And that was kind of the first taste of entrepreneurship in 74 - > some way, but actually it was more around personal passion and 75 - > things that I found difficult.

76 - > Definitely something that I'd say most people don't know, but 77 - > that was kind of my uh probably my first passion, and it's 78 - > probably something that I will continue to spend a lot of time 79 - > thinking about in uh doing nonprofit work in for the rest 80 - > of my life. 81 - > SPEAKER_01: Yeah, I mean it's it's a fantastic story. 82 - > And and if you don't mind me sort of asking, digging a little 83 - > bit further into it. 84 - > So I I'd imagine, I don't know what sports it was, but I'd 85 - > imagine piecing it together, it's probably quite a macho type 86 - > of world, and you're not meant to admit any weakness and all 87 - > that type of thing.

88 - > And so how does it work? 89 - > Is it more like peer groups? 90 - > Is it's the likes of you starts to open up? 91 - > SPEAKER_00: Yeah, yeah, it was rugby that I played when I was a 92 - > teenager, so yeah, quite um macho and in in every sense of 93 - > the world, I suppose.

94 - > And I my whole identity was wrapped in that thing, and then 95 - > when I had to give it up, I had you know, my right leg is 96 - > effectively half metal, I had to have six surgeries from the age 97 - > of 14 to 18. 98 - > What ended up happening was I realized that like once I was 99 - > not in that team, I had this idea of like, well, what am I 100 - > then? 101 - > Like I have no friends now because my friends were the 102 - > rugby team, or my friends were can they only liked me because 103 - > of my sport or whatever it was.

104 - > So that was that's a hard thing to go through when you're a 105 - > teenager at the best of times. 106 - > And then I mean, like most cathartic vulnerability 107 - > experiences, my best friends uh, you know, I think we were in the 108 - > bar, a pub or something, and we ended up kind of going, Wow, 109 - > that period in school was really tough, and for a very different 110 - > sport, a different reason, we both had a went through a 111 - > similar struggle and ended up opening up to one another, and 112 - > then we kind of laughed and at how crazy it was that we were 113 - > walking beside each other in the corridors for five years, going 114 - > through a very similar bad experience and never spoke to 115 - > one another about it, even though we were friendly and 116 - > similar groups, and that kind of inspired a very simple message, 117 - > which was to kind of your first display of vulnerability having 118 - > such a positive response to it, encouraged us to try and like 119 - > get more people opening up and speaking about it in a very 120 - > public sense.

121 - > And funnily enough, I got invited on like a podcast before 122 - > podcasts were really a thing in university, and I remember after 123 - > doing that, there might have been four or five hundred 124 - > listeners, and I might have gotten four or five hundred 125 - > messages straight off the back of it. 126 - > And I think what that does is it gets you angry in some sense 127 - > because you're saying, How is this causing such a reaction for 128 - > such a normal thing?

129 - > You know, being in a bad place mentally is such a normal thing 130 - > uh for everyone. 131 - > So that kind of inspired us to do something about it, even on a 132 - > very small scale, and a very at a local community level. 133 - > And I think now it's you know just the world has come on leaps 134 - > and bounds around that sort of thing. 135 - > But it told me a lot about myself and you know what I was 136 - > capable of and what I could get through.

137 - > And I think also the type of business I wanted and life I 138 - > wanted to then have. 139 - > Yeah, definitely a kind of a I mean an unknown, but a very 140 - > central part of my and our story. 141 - > SPEAKER_01: Yes, as I said, we might digress a little bit from 142 - > the planned focus that I mean, I also think it's an incredibly 143 - > meritorious initiative, and even though things have moved on, I 144 - > mean I think men in particular are notoriously bad and doing 145 - > anything more than a kind of token check-in.

146 - > Of going, you know, are you alright, mate? 147 - > And they go, Yeah. 148 - > And then they don't go, are you really? 149 - > Well right.

150 - > And and as you say, I think that idea of actually sharing 151 - > something about yourself leads to. 152 - > SPEAKER_00: Yeah, and also like the reason why I think I was I 153 - > had the confidence to do so was because I was just surrounded by 154 - > exceptional women. 155 - > You know, my mum, my sister, I've been with my partner for 156 - > eight years, Ellen, like those people might have easily been 157 - > the key that unlocked the door, and also kind of without even 158 - > telling you subconsciously, like educated you on how to pull 159 - > things out of your male counterparts.

160 - > Like, even things like I remember my mum saying whenever 161 - > I was upset as a kid she would take us on a drive, because 162 - > whenever we were looking at each other, I was more likely to open 163 - > up. 164 - > You know, boys like looking in the eye is harder to open up. 165 - > And that's my theory for like why saunas have become really 166 - > popular in Ireland. 167 - > And I think most of it, because in saunas, if I'm sitting beside 168 - > my friend, we're usually looking out, not at one another.

169 - > And I think it's funny, like that learning is why. 170 - > Like, if I knew my friend I'd be struggling and I I wanted to try 171 - > and speak to him, sitting on a bench where you're looking out 172 - > or going on a walk where you don't have to look in each 173 - > other's eye, is like a learning I got from my mum on how to get 174 - > people to feel more comfortable opening up or displaying 175 - > vulnerability first, then allows people to then be vulnerable 176 - > after.

177 - > So, again, it's kind of the learnings from the exceptional 178 - > women in my life, and then that kind of is a probably a through 179 - > line through my whole life, but and how you carry that into your 180 - > male relationships. 181 - > SPEAKER_01: Yeah, yeah. 182 - > I mean, your mum's totally right about it. 183 - > It was actually a point that was raised.

184 - > I had a book I read about it's like bringing up boys, boys, and 185 - > actually the uh that was one of the points. 186 - > They said going for a drive, exactly that thing where you're 187 - > not slightly distracted, you're not looking at each other is one 188 - > of the tricks to do it. 189 - > So going back to Bouse, I mean, there's an interesting kind of 190 - > origin story there as well, which I'm sure you've told many 191 - > times before, but let's do it again.

192 - > Do you want to play the dragon's den type scenario? 193 - > SPEAKER_00: Yeah, I mean, so I went to Trinity College in 194 - > Dublin, which is like the number one university for 195 - > entrepreneurship in Europe, and it's why I wanted to go there. 196 - > My sister had gone, and she's a doctor, but I I'm the youngest 197 - > of four, about my parents entrepreneurial. 198 - > I kind of went in this pursuit of entrepreneurship and that 199 - > that university, and not many people from my school going 200 - > there, and that was like a really I just like enjoyed a lot 201 - > of those elements.

202 - > But very quickly, what I loved about college or university was 203 - > very quickly I was going to these talks from Michael 204 - > O'Leary, the CEO of Ryanair, and Isolt Ward, the founder of Food 205 - > Cloud, an amazing business in Ireland, and very quickly got 206 - > like the bug for wanting to set something up. 207 - > And luckily, there was lots of outlets for both finding 208 - > like-minded people, although that sounds kind of like 209 - > university jargon. 210 - > In reality, I was in the college bar and you know, drinking pints 211 - > with people who also had that ambition to set stuff up.

212 - > And I was incredibly fortunate to meet my now three 213 - > co-founders, Ronan, Josh, and Brandon, who were studying 214 - > computer science. 215 - > And there was this Dragon's Den competition coming up, or Shark 216 - > Tank for US listeners, kind of a very similar idea. 217 - > And we applied effectively with a business, or we're going to 218 - > apply with a business that was just really bad. 219 - > And in that Dragon's Den competition, we got a huge 220 - > amount of feedback, being like, You did no research, you spoke 221 - > to no customers, you validated nothing.

222 - > And we then kind of had the bug of, well, we love working on 223 - > this idea, but now that idea is really bad, we need to go do 224 - > research. 225 - > And very, very long story short, I ended up meeting the CEOs of a 226 - > lot of the big research agencies in Ireland as a student looking 227 - > for advice and help on how it works. 228 - > I learned a lot through that process. 229 - > I met with the likes of Qualtrics and Taluna and some of 230 - > these uh kind of software companies disrupting the space.

231 - > And then I met with you know the marketing directors of Diaggio 232 - > and Coca-Cola and Tesco in Ireland and said, hey, you know, 233 - > I'm trying to do research. 234 - > How I got meetings with all of them, I still don't know. 235 - > I was just maybe persistent and didn't take no for an answer, 236 - > and I was a student, so maybe they were just willing to help. 237 - > But what I learned through that kind of cycle and through 238 - > looking at all the other startups in this space was just 239 - > a lot of the core problems that still persist within research.

240 - > It felt like a the agencies had amazing people, but lacked core 241 - > competencies in technology, so we're never really gonna 242 - > innovate in this new machine learning world that was coming. 243 - > So far, companies were obsessed over the like tool they were 244 - > building and giving, you know, giving someone a fishing rod 245 - > instead of teaching them how to fish, that kind of idea. 246 - > So we felt that was going to hit a natural end because so much 247 - > research is custom and complex and starts at the end and moves 248 - > backwards.

249 - > So the more we became engulfed in this industry, we effectively 250 - > scrapped the other idea and started on this kind of five, 251 - > six, seven-year journey of building what we would foresee 252 - > as the future of how technology and agencies would evolve, all 253 - > rooted in the same objective, which for market research is 254 - > simply the understanding of the person you're trying to sell to 255 - > so you can build better products, sell more stuff to 256 - > them.

257 - > You know, that's all research is is the pursuit of a greater 258 - > understanding to make greater decisions. 259 - > And we felt there was a lot of flaws from a technology and a 260 - > people and a process point of view that with me and my 261 - > co-founders, we could try and iterate and build a solution 262 - > around. 263 - > And that that's effectively how we end up falling into the 264 - > research phase was trying to do research ourselves and realizing 265 - > there was no good.

266 - > If there was no good solution for us, then maybe there was no 267 - > good solution for everyone, and there goes a lot of room for 268 - > growth. 269 - > SPEAKER_01: And I I totally I totally see it. 270 - > It's fascinating coming at it from a sort of an external 271 - > perspective. 272 - > And I guess I won't get into the detail of the company.

273 - > One of the companies I work with is somebody I used to work 274 - > closely with who came in and became the chief data scientist 275 - > and wasn't out of the research background, out of the research 276 - > world conventionally. 277 - > And he was like, why on earth is it done like this? 278 - > And that's the other business he's he's set up is to is to 279 - > reconcile everything. 280 - > SPEAKER_00: So, yeah, really.

281 - > To a certain point, though, it's interesting you say that because 282 - > I think that's what enabled us to uh to do our like zero to 283 - > one, like the software we built and the way we pitched and the 284 - > way we understand the problem really gave us an advantage. 285 - > That kind of lateral thinking, you know, taking first 286 - > principles, putting it into an industry, you don't have the 287 - > like baggage of experience in in a kind of a funny way, but 288 - > definitely the kind of one to ten journey we've been on since 289 - > has been completely tied to the quality of researchers and 290 - > industry experts that can almost help explain why things are the 291 - > way they are.

292 - > You know, there's this amazing book, Same as Ever by Morgan 293 - > Housel, the financial author, which is like you should really 294 - > focus on the things that don't change in order to try and 295 - > predict the future instead of trying to predict the things 296 - > that will change, because I think it no other industry is 297 - > like it when it comes to market research, when it comes to 298 - > stubbornness and resistance to change and things being the way 299 - > they are, you could spend all your life complaining about why 300 - > things are the way they are, or you can put yourself in the 301 - > shoes of the researcher and change, manage, and articulate a 302 - > better future.

303 - > And I think when we combined our like technical superiority with 304 - > first principles thinking and you know, progressive industry 305 - > experts, it was a very potent mix that has definitely 306 - > accelerated our growth massively. 307 - > So you need a bit of everything. 308 - > SPEAKER_01: Yeah, I I and even referring back to the example I 309 - > was I I think Tom would probably say once he got involved in it, 310 - > started doing it, it was more complex than he thought, but 311 - > he's continued to roll up his sleeves and solve it.

312 - > So, Charlie, we should probably back up a little bit though and 313 - > let everybody know what bounce actually does. 314 - > SPEAKER_00: Yeah, for sure. 315 - > So what we try and do now is deliver insights that 316 - > researchers can confidently share and defend as quickly as 317 - > possible. 318 - > And that is what we are trying to do.

319 - > Back to like a very first principle idea. 320 - > Most research is rooted in I have a decision I need to make 321 - > or I have a question that I've received, and it is my goal to 322 - > get a good enough answer as quickly as possible. 323 - > So that is why we exist as a company. 324 - > The product that we've built to solve that is two has two clear 325 - > clear components.

326 - > One is a retrieval system, and in the simplest form, the goal 327 - > of that is to consolidate all past research in one place so 328 - > that you can have a conversation with your data and extract 329 - > insights that are fully sourced, sided, robust, backed up in 330 - > evidence. 331 - > It's the exact opposite of throwing something into Claude 332 - > or Chow GBT. 333 - > It's this idea of built by researchers, designed by 334 - > researchers, rooted in research integrity.

335 - > And the goal of that is to tell people what they already know, 336 - > very simply. 337 - > And that is the kind of evolution of knowledge 338 - > management in the in the in the market research world, which 339 - > people might be familiar with. 340 - > And the second component is what happens when gaps exist. 341 - > So let's say you're a tequila company, you say, Hey, I want to 342 - > know how to price this new tequila product in the on-trade 343 - > in Texas ahead of the World Cup.

344 - > What should we do? 345 - > Well, the first thing you should do is figure out all of the 346 - > pricing research you've ran in this category or what might be 347 - > publicly publicly available, consolidate all that 348 - > information, but you might not have anything 18 to 24-year-old 349 - > females, and they're a critical category for you. 350 - > So when we identify that gap via our retrieval system, the second 351 - > part of our offering is the ability to go out and run new 352 - > primary research to fill that gap.

353 - > So we have a system that takes that gap or takes any brief and 354 - > will then recommend a quantitative research study to 355 - > fill that gap. 356 - > So very classic survey methodology, audience, and 357 - > interlocking quotas to ensure it's robust. 358 - > We run that fieldwork using humans' respondents, and then we 359 - > will analyze that data with the original objective in mind. 360 - > So you can picture it like a flywheel design product.

361 - > They come in the top, they ask banks what do we know about this 362 - > topic, we identify gaps, and then we fill those gaps. 363 - > And it becomes like this repeatable decision-making 364 - > system within an organization. 365 - > And critical to that system is we have researchers pulling the 366 - > strings the whole way through. 367 - > So half our company are researchers, and half our 368 - > company are software engineers building in the AI space.

369 - > And what that allows us to do is effectively be kind of change 370 - > management powerful people within these orgs. 371 - > We work with Heads of Insight to reconstruct their organization 372 - > around this ideal flywheel. 373 - > And that's what we do. 374 - > So we're about 50 people now, offices in New York, London, and 375 - > Dublin.

376 - > Uh we have about 200 clients globally, mostly Fortune 500s, 377 - > but we work with startups and smaller brands as well. 378 - > And yeah, I suppose the why we exist is just to allow more 379 - > decisions to be insight-backed out of across the world. 380 - > So that's really the the raise on debt. 381 - > SPEAKER_01: And Charlie, what did you mean by making decisions 382 - > with information that's good enough?

383 - > That wasn't exactly the phrase you used, but it was something 384 - > along those lines. 385 - > I mean, is that suggesting that sometimes the uh level of 386 - > information can be over-engineered by the 387 - > traditional industry? 388 - > SPEAKER_00: I think without your uh uh leading me to water, I 389 - > think there's there's definitely a some decisions require 98-99% 390 - > confidence. 391 - > Decisions by the US government in relation to a war, they 392 - > should probably have a very high confidence in what they're about 393 - > to do.

394 - > Yeah, trying to decide whether you should price something at 1 395 - > euro 12 cent or 1 euro 5 cent might require 75% confidence. 396 - > And I think the reality on the ground is that humans make 397 - > decisions with so many biases that all you're looking for is 398 - > enough information to deliver enough confidence based on the 399 - > speed required. 400 - > Because my hypothesis is that a good enough decision made 401 - > quickly wins over time versus a perfect decision made slower.

402 - > And I think Barack Obama has spoken about this, the 51% 403 - > decision. 404 - > And if 51% confidence is good enough for the president of the 405 - > United States, it's probably good enough for the category 406 - > manager of a yogurt company in the UK. 407 - > And that is not to say that's always the case. 408 - > I'm not, you know, poo-pooing the idea of you know robustness, 409 - > confidence, but it's rooted in what is the decision you're 410 - > trying to make, what is the level of confidence you'd be 411 - > satisfied with, and what speed do you need to make the decision 412 - > by.

413 - > Because slow research is really, really important. 414 - > Slow research uncovers human insights and understanding that 415 - > just requires time and that's necessary. 416 - > But sometimes speed is the most important trigger, and it's how 417 - > it's either I do nothing and make a decision, or I get as 418 - > good as I can and then make a decision. 419 - > Our role at Bounce is to accelerate people up that curve 420 - > to get the highest level of confidence based on the speed.

421 - > So I either need something in an hour or I need something in a 422 - > week. 423 - > That is going to increase the level of confidence or 424 - > robustness we can deliver to that person, but it's all rooted 425 - > in what decision you're trying to make, and it's all 426 - > context-specific. 427 - > Yeah, I think it's very well put. 428 - > Like the biggest problem in research is like it's probably 429 - > inefficiency and attribution.

430 - > Inefficiency being how slow and expensive it is to actually 431 - > deliver good enough insight. 432 - > And attribution is when you make a decision back by insight, how 433 - > did it go and how do we capture value of whether it was a good 434 - > or a bad decision? 435 - > I think those two things are the things that we're trying to 436 - > solve. 437 - > It's really hard to, but given the setup of structures and 438 - > teams and businesses and how agencies and tools interact, 439 - > it's quite messy and it's it's hard to then make change happen.

440 - > SPEAKER_01: Yeah, yeah, definitely. 441 - > And do you have a particular area of specialization? 442 - > I mean, and the reason why I ask that is for instance, like 443 - > around sector, because I'd imagine different data sets, if 444 - > we're looking at that kind of knowledge management piece of 445 - > it, probably have different challenges. 446 - > You know, even something that we've talked about, if it might 447 - > be TV ratings as opposed to, I don't know, beverage retail 448 - > data, probably have different considerations around them, I 449 - > would imagine.

450 - > SPEAKER_00: Yeah, for sure. 451 - > So yes and no. 452 - > So on the we work with all sectors in that space because 453 - > our our system, again, you have to remember that the data that 454 - > goes into the retrieval system is all client-owned data. 455 - > It's a closed system.

456 - > So whether you're a tobacco company or a betting company or 457 - > a yogurt company, the data, the differentiation is in the 458 - > extraction of the inside and the citation and the identification 459 - > of like data conflicts and that sort of thing. 460 - > It's not in the um the data itself. 461 - > Like our value is not in the data we provide. 462 - > It's the same on when we run new research.

463 - > We are we are integrated with panel partners to deliver the 464 - > sample. 465 - > So that's why that's what allows us to be industry agnostic. 466 - > In house, our researchers are specialists in different areas, 467 - > whether that be methodological or industry based. 468 - > We're big enough now that we can kind of have enough people that 469 - > are go to different areas.

470 - > You know, if something comes in on, I don't know, creative 471 - > testing, we'll know what researcher to put on it, or if a 472 - > certain sector will say, Hey, oh, I know the person who's 473 - > worked in that sector a lot. 474 - > But the technology itself is built to be more industry and 475 - > methodology agnostic. 476 - > And what that means is like a lot of our clients, we will 477 - > recommend they use certain specialist agencies or tools in 478 - > certain scenarios.

479 - > So let's use a scenario where we retrieve an answer for them and 480 - > we say, hey, you probably need to go and run some qualitative 481 - > research in the pharmaceutical space. 482 - > We will say, hey, here's who we recommend you go to. 483 - > It's almost the same with panels. 484 - > There's a great transparency and trust that's built when if we 485 - > are the orchestrator or the conductor of their kind of 486 - > research execution, we can tell them who to use based on our 487 - > vetting and recommendation.

488 - > SPEAKER_01: Got it. 489 - > And obviously, I haven't seen the system actually in action, 490 - > but in terms of the user experience, I mean, is it a sort 491 - > of LLM type of approach where you type in and go, I don't 492 - > know, what can you tell me about females 18 to 24 and their 493 - > consumption? 494 - > SPEAKER_00: I have uh I have strong opinions what the future 495 - > like UX is going to look like for a lot of these companies. 496 - > So currently, I don't care whether any of our clients ever 497 - > log into the Banks platform at any point.

498 - > And the reason why I don't care about that is because I think 499 - > one of the big things that have struggled, like have caused a 500 - > lot of research tools to struggle getting in, struggle 501 - > getting embedded in an organization is integration, 502 - > workflow, alignment, like making it easy for you to be the thing 503 - > they go to. 504 - > So in a perfect world, the engagement with bands would look 505 - > like one of two things. 506 - > I'll get more specific in a second.

507 - > Either they email us their problem, and our researcher 508 - > takes that, uses the system, tells them what they already 509 - > know, identifies the gaps, recommends the solution. 510 - > And effectively, we are just a hyper-efficient full service 511 - > agency as it feels to the customer. 512 - > Now, our gross margin about 80%. 513 - > So I don't care whether people have a SaaS platform or not.

514 - > You know what I mean? 515 - > So that is a perfect world as it is now. 516 - > Researcher or a person comes to us with problem, we solve 517 - > problem. 518 - > We use that through a researcher trade on our system.

519 - > There are self-service elements of our system if they want to 520 - > engage in it. 521 - > Like you said, they could type into the system themselves. 522 - > What do we know about tequila drinkers in Germany? 523 - > Or they could, when the results come back of a quantitative 524 - > research study, they can pull out the insights, export it to 525 - > an editable PowerPoint.

526 - > But we don't charge for consultancy. 527 - > So we allow the client to either use self-service or our people, 528 - > completely dependent on their need, because we want to reduce 529 - > friction. 530 - > How I see that going is instead of emailing us, they we could 531 - > just be integrated into their Microsoft Teams, their Slack, 532 - > their Claude, whatever it is. 533 - > So when they pose the question to Bounce, it could just be a 534 - > webhook, which is Slack Bounce, and it sends the question or 535 - > sends the brief.

536 - > And that means that there is zero change in their day to have 537 - > it in Imperial brands or in Coca-Cola to be able to extract 538 - > insights, identify gaps, run the research. 539 - > So right now it's researcher self-service or full service, 540 - > depending on how you need. 541 - > It's the exact same offering. 542 - > It's dependent on the client experience.

543 - > In future, I imagine it to be way more amalgamated within 544 - > existing technologies, infrastructures, systems, and 545 - > it's our job to it's right with them. 546 - > So that's that's a massive change I see happening over the 547 - > coming months and years. 548 - > SPEAKER_01: Yeah, that makes a lot of sense to me. 549 - > And just so I've got a full understanding on the kind of 550 - > what I think of as the kind of the repository bit of it, the 551 - > knowledge management side.

552 - > So what exactly do you integrate? 553 - > I mean, do you take historic survey data? 554 - > SPEAKER_00: Might you take desk research, focus groups, there 555 - > are two forms of data in the market research worlds. 556 - > There is unstructured data, which is PowerPoints, docs, 557 - > Google Slides, industry reports, Mintel, all this sort of thing.

558 - > Like basically reports with graphs and text and X and Y 559 - > axes, and that is like output level data. 560 - > Um, and we've built a model to ingest them, auto-tag them, 561 - > understand them. 562 - > So it requires no manual effort when we get that data to have a 563 - > full extraction of that data and to understand it, and not 564 - > conflicts, et cetera. 565 - > We've had to build our own custom model then for ingesting 566 - > structured data, which is things like Circana, Nielsen, spins, 567 - > more complex Excel-level data, which might require additional 568 - > context in order to be ingested and understood.

569 - > We call that input data. 570 - > And the reason why both of those are important is depending on 571 - > the question you are posing, you might not want the summarized 572 - > output answer. 573 - > You might want the specific crosstop on the survey. 574 - > So we've ingested survey data from all of the different DIY 575 - > quant tools that are out there.

576 - > We've ingested qualitative transcripts, we've ingested 577 - > syndicated data sources, we've ingested market research report 578 - > data, we've ingested what you name it, and it becomes, we can 579 - > do that and host it in our system, or we can effectively 580 - > through an MCP or an API integrate into where that data 581 - > already lives. 582 - > So if someone's already done the work to build up a repository in 583 - > SharePoint, we should just be able to integrate through there 584 - > and you don't need any effort.

585 - > If you haven't built up a repository, we will build one 586 - > with you and make it as easy as possible to do that. 587 - > But the technical complexity to do that is exceptional. 588 - > I see a lot of like people launching like repositories as 589 - > part of their offering over the last few months. 590 - > I'm excited to see when they realize the technical complexity 591 - > of adjusting the volume of data sources and to be able to 592 - > guarantee the citation to quality of the answers and the 593 - > sourcing.

594 - > It's taken us a long, long time to do it right. 595 - > It's a very obvious use case of AI because it's like, oh well, 596 - > throw in reports and summarize the answer. 597 - > It's not that simple because if people are making decisions on 598 - > the answers we are giving, we have to be 100% confident in 599 - > them. 600 - > And when we when gaps exist, we need to be able to fill those 601 - > gaps in a very comfortable research first way.

602 - > So that system has been years in the making, um, and we are right 603 - > at the kind of frontier of how we're leveraging AI, how we're 604 - > leveraging our own data, and how we're kind of building it with 605 - > our clients to ensure that it is doing the thing it says it it 606 - > does, because there's a lot of AI disillusionment at the 607 - > moment. 608 - > And once the way we will pilot it is we'll say, hey, send us 609 - > questions that you already know the answer to through manual 610 - > effort, then run those same questions in cloud or co-pilot, 611 - > and then run those same questions in bounds and just 612 - > compare the quality of answer with the speed with the cost.

613 - > It's a meritocratic system, and that's where I think that's 614 - > what's driving a lot of our growth at the moment is you can 615 - > actually prove you're better, which is something in the 616 - > research phase that is very difficult to do. 617 - > SPEAKER_01: Yeah, very much so. 618 - > And and and everything you were just talking about there, 619 - > Charlie, I guess becomes part of my concern around the 620 - > repositories. 621 - > So I imagine myself, let's say I've sent it to you, or if I put 622 - > it into the system.

623 - > And and then is there's this question around what's the 624 - > source of the data, and then related to what we talked about, 625 - > what's the confidence level around this relative to the 626 - > level of decision that I'm looking to make? 627 - > SPEAKER_00: There's a few ways that you try and solve it in the 628 - > repository space. 629 - > One is how do you help clients pose the right question with the 630 - > right context? 631 - > So we have like training on the structures upon which you pose a 632 - > question in order to extract the right answer.

633 - > Another thing is when we are pulling an answer, we are only 634 - > pulling it from a smaller number of sources than it could pull an 635 - > answer from. 636 - > So let's say you've thousands of documents in there. 637 - > Before we create an answer, we will show you the data or the 638 - > sources we are going to pull an answer from. 639 - > And you can manually select or deselect or add it anything else 640 - > you manually want to do back to the researcher has their hands 641 - > on the wheel.

642 - > It's not just this free-flowing agent. 643 - > Then when the answer is delivered, we will call out 644 - > where there are conflicts in data. 645 - > So let's say one of your reports says X and the other report says 646 - > Y. 647 - > We allow for manual overlay of which one are you going to take 648 - > as truth, and then that trains it further.

649 - > Then when the answer is given, every single statement has a 650 - > citation, like a PhD style. 651 - > So it's a natural language like you would see in a Claude or 652 - > ChatGBT from a user experience point of view. 653 - > But every statement you can click a citation and it's going 654 - > to take you to the like Henry verified points that you would 655 - > put in. 656 - > So the quality of data that goes in is one of the really critical 657 - > steps.

658 - > And then there's all of these steps to ensure that it is 659 - > effectively shareable and defensible. 660 - > So the use case we say is that when you get that answer, you 661 - > can go do a meeting with the stakeholder who was posed that 662 - > question. 663 - > And when they say, Well, why do you believe that? 664 - > That every single statement is defensible, backed up in data, 665 - > and makes the person look smart, feel smart, and be able to 666 - > convince that person to act, which is another big challenge 667 - > of the research space.

668 - > SPEAKER_01: Yeah, it is directly related to the project that we 669 - > were touching on kind of before the call, where this question of 670 - > conflicting data sets and going through the decision flow and 671 - > going, what am I going to place the most credence on? 672 - > Which is why this is why the human is so important, Henry. 673 - > SPEAKER_00: Like the the conductor of this orchestra is 674 - > the most important person, choosing the pace of the music 675 - > flow, choosing who goes when, choosing who you go to next.

676 - > That is why we're building a read a system for researchers to 677 - > elevate the quality of researchers in organizations. 678 - > We're not one of these companies that is like, there will be no 679 - > research function in the future, research doesn't matter. 680 - > Now we finally democratize insights. 681 - > We're like the reason why we're building it the way we are with 682 - > research and insights professionals is because we know 683 - > how to do it, we know how to do it well, and we know the 684 - > importance of the conductor through this whole process.

685 - > And when things are wrong or where things are challenged, the 686 - > ability to then dive in and play one of the instruments, or you 687 - > know, to use that analogy is going to be so, so integral to 688 - > make sure that the insights functions are the ones that are 689 - > building these intelligence networks in the companies and 690 - > not people who don't understand research integrity and guidance 691 - > and consumer psychology and the emotion behind what decisions 692 - > actually drive us decisions.

693 - > It's really important. 694 - > Yeah. 695 - > SPEAKER_01: Uh Charlie, a bit left field, but where are you on 696 - > the whole synthetic data digital twins side of things and all the 697 - > rest of it? 698 - > Is that part of the offering that Banks?

699 - > SPEAKER_00: No, it's not. 700 - > I'm very glad I'm not in that world because it's uh it's a 701 - > very fun debate to sit from the sidelines and eat popcorn and 702 - > watch. 703 - > Um my opinion on it is I've a few thoughts on it, but uh like 704 - > I don't take advice from people who aren't experts in a space. 705 - > I don't think anyone should take my advice on this.

706 - > But here's my from people who know comp uh I know founders who 707 - > are building in that space, and I obviously know look, we've had 708 - > to speak to them because in the future, instead of providing 709 - > human respondents, some of our clients might want a synthetic 710 - > response. 711 - > Some of my thoughts are on it are this. 712 - > One, we are already doing forms of synthetic modeling. 713 - > I sin I think it's one of the worst marketing positionings 714 - > I've ever seen.

715 - > Totally agree. 716 - > Yeah, calling it synthetic is just a terrible idea. 717 - > It makes it feel fake when it's not fake in a lot of these 718 - > cases. 719 - > So I think they have a marketing problem.

720 - > We are already doing forms of synthetic modeling. 721 - > If you've ever done MacStiff or any of those methodologies, you 722 - > were doing forms of mathematical modeling to give confidence. 723 - > So this is not new in that case. 724 - > Third, and this is more on the concern side, we haven't seen 725 - > it, but we aren't as close to the frontier of this, be good 726 - > enough to really roll out to clients yet.

727 - > But as soon as it is, we will. 728 - > We are very open-minded and a clock of a client are very 729 - > proactive, they want it to be true. 730 - > But my concern is this, and it's the same in the AI moderated 731 - > qual space, and again, some amazing businesses in that 732 - > space. 733 - > When I see a lot of venture capital money flooding a 734 - > category or a space, because the incentives are great if they're 735 - > right.

736 - > And what I mean by that is because the the amount of money 737 - > that goes into human sample and incentivizing and the amount of 738 - > money that gets spent on qualitative research right now 739 - > is so high and it's so slow, those two processes, the 740 - > incentives for synthetic to or AI moderated call to work are so 741 - > high that I think it skews the human dimension. 742 - > So I think it skews humans' belief that it will be true. 743 - > When you want something to be true, you are more likely to 744 - > believe it will be true.

745 - > So I have skepticism purely because the amount of money 746 - > flooding into the space, and because the incentives for 747 - > wanting it to be true are high. 748 - > And naturally, as a founder, my spidey senses tingle a bit. 749 - > And I say, I don't think that there is a high risk for that 750 - > not being as far along as it could be, because people are 751 - > flooding capital and want it to be true. 752 - > And if it does work, it will be transformative to the org.

753 - > If qual does replace quant, huge ramifications for the industry. 754 - > If synthetic does replace sampling, huge ramifications for 755 - > the industry. 756 - > But back to the Morgan Housel thing, things that don't change. 757 - > I am still a bit skeptical.

758 - > I wish all the founders in that space the best. 759 - > And if they solve it, it will make my business better. 760 - > So I actually don't even force in the race. 761 - > It it will make me it will help Bounce deliver its value 762 - > proposition faster, better, cheaper, which is always good.

763 - > So they're my thoughts. 764 - > SPEAKER_01: So what comes next for the business? 765 - > Where where do you hope to be in I don't know, three to five 766 - > years? 767 - > SPEAKER_00: Honestly, if I have a year like the last year for 768 - > the next three to five years, I will be very happy.

769 - > Back to the point that we spoke about at the start around you 770 - > know, the juice is the squeeze, or you know, Adam Brand's 771 - > phrase, you know, instead of the pursuit of happiness, it's the 772 - > happiness of pursuit. 773 - > They sound cliche and cheesy, but it's how I live my life. 774 - > Right now, we have doubled the business in revenue every year 775 - > for the last three years. 776 - > We will double again next year.

777 - > We are growing at a really crazy but very manageable speed 778 - > because the productivity of the business with less people is 779 - > amazing. 780 - > I love what I do. 781 - > I get to hire incredible people who I love working with, and we 782 - > are in control of our destiny. 783 - > So we raised a small amount of capital, like$7 million over two 784 - > rounds.

785 - > We are now a cash flow break even again. 786 - > So we are using our revenue to continue doubling year on year. 787 - > So everything feels within our destiny, which is really 788 - > important when the macro environment is the way it is. 789 - > Like, you know, with what the in advancements in AI, with wars 790 - > going on across the world, with all of the stuff going on, there 791 - > are so many uncontrollables that could kill bounce that if I 792 - > thought about those, I would be sleep at night.

793 - > So all I think about is very controllable items, which is 794 - > what we call a bounce, like a great company. 795 - > And a great company is where people are like growing in their 796 - > careers, well parried, enjoying their work and feeling 797 - > challenged. 798 - > Like it's quite a simple set of ingredients that we try home. 799 - > And if we can keep going the way we are with our vision from a 800 - > product point of view, there are so much technical challenges 801 - > that are really rewarding to solve.

802 - > There are so much business challenges and clients to work 803 - > with and try and win that are really exciting. 804 - > And I think the industry is in such a fascinating place that 805 - > I'm honestly just like I'm 28 years old, so I almost forget. 806 - > I've obviously been doing this since university, and like I 807 - > have no ambition to retire. 808 - > We are do building an exceptionally interesting 809 - > business and an exceptionally interesting time with 810 - > exceptionally interesting people, and long may that 811 - > continue.

812 - > That's kind of my uh my current feeling on the whole matter. 813 - > SPEAKER_01: One of the intriguing things for me around 814 - > it, Charlie. 815 - > I mean, I should be careful about this because I don't want 816 - > to slag off the whole industry, is that the consumer insight 817 - > sector's been so stodgy and it's been really difficult to change, 818 - > and that researchers, by their nature, always see the problems 819 - > rather than the opportunities. 820 - > But it seems like you found a client base that doesn't see it 821 - > like that.

822 - > Yeah. 823 - > SPEAKER_00: Yeah, but you also have to you have to help them 824 - > gain their voice. 825 - > Like, I've a huge part of our job is giving people confidence 826 - > to elevate themselves within the organization and not be 827 - > defensive but be offensive and show them how a really clear, 828 - > easy use case of AI that they can bring to their boss and look 829 - > really good. 830 - > So I think we're trying to take an opposite tack to a lot of the 831 - > software and research companies that have entered this face over 832 - > the last 10 or 15 years.

833 - > And a lot of that is research first, it's quite empathetic to 834 - > the problems. 835 - > And I do listen to, you know, insights wanting a seat at the 836 - > table. 837 - > Instead of moaning as to why they don't have a seat at the 838 - > table or why, you know, you know, teams are getting cut. 839 - > Our job is to try and help elevate them and help them 840 - > become the most important person in the team, to help them 841 - > transform the way insights are delivered.

842 - > Because back to why research exists as a function, it's 843 - > ultimately just to help companies make better decisions 844 - > that are rooted in human understanding. 845 - > Like that is it. 846 - > So it's a really important role these people play. 847 - > And if they lack confidence or if they lack articulation or 848 - > they lack entrepreneurial skill sets, it's our job to give it to 849 - > them and help work with them.

850 - > And yeah, we've got exceptional clients. 851 - > Like we've clients who gave us a chance when I was a student with 852 - > no understanding of research, barely a platform, and they're 853 - > still working with us today, like Prinoricar, Diagio, 854 - > Coca-Cola, Tesco, some of these really early customers. 855 - > And now we're working with you know 10, 15 markets and teams 856 - > within a lot of those organizations because we've 857 - > co-created, we've listened, we've built with them, and we 858 - > have a very compassionate understanding of what the future 859 - > looks like.

860 - > And I think that's something that a lot of tech companies 861 - > have lacked and still lack when you hear them speaking on stage 862 - > around replacing research orgs and we're gonna make your job 863 - > faster. 864 - > It's like people don't want to work faster, they want their 865 - > life to be easier. 866 - > So I just think we try and think about things a bit differently 867 - > and work with them. 868 - > And maybe that's what has allowed us the success we've had 869 - > to date and hopefully allows us to continue on this very fun, 870 - > hard journey.

871 - > SPEAKER_01: Have you got any kind of key elements of advice 872 - > you would give to other founders if they're looking to raise 873 - > money, particularly in this type of environment? 874 - > SPEAKER_00: I could do hour rays on this topic alone, but I'll 875 - > try and keep it brief. 876 - > Um raising money is raising debt, is the first thing. 877 - > Only raise money if you absolutely have to, if you are a 878 - > venture business.

879 - > And what I mean by venture business is you are have the 880 - > ability to actually deliver venture returns. 881 - > A massive mistake raising money is when you don't have venture 882 - > returns, you give away a portion of your business, and then you 883 - > are effectively destined for failure in both stakeholders' 884 - > sense. 885 - > So, a lot of the reasons why a lot of software companies in our 886 - > space are failing is because they raise too much money at too 887 - > high a valuation that they were never going to achieve, and 888 - > they've had to do mass sets of layoffs, restructures, re-orgs, 889 - > and now they're trying to turn their software business into an 890 - > AI business.

891 - > And that is challenging and not really the fault of a lot of 892 - > companies, but raising equity is a serious, serious thing, and 893 - > you should treat it like debt, is kind of the first thing on 894 - > that. 895 - > The second thing when going into fundraise is you need to treat 896 - > it like a sales cycle, like a funnel. 897 - > If you can assume a win rate of 0.5% on getting a term sheet, 898 - > you need to make sure you're speaking to enough people.

899 - > You need to make sure you are going to the right venture 900 - > capital firms. 901 - > You need to do your homework, like you're trying to do pitch 902 - > to a customer. 903 - > Too many founders, I see they're like, Oh, there's no way we can 904 - > raise funding. 905 - > The environment's so hard, there's not enough money, it's 906 - > so unfair.

907 - > How many companies, how many bench firms do you speak to? 908 - > And they go, Oh, like 20? 909 - > And you're like, What? 910 - > Speak to 200 and then maybe come back to me and like learn from 911 - > every single meeting you've had.

912 - > So I think there's a naivety around the process. 913 - > I think there's a naivety around when you should and why you 914 - > should raise, and what are you actually trying to achieve as a 915 - > company? 916 - > So I think a lot of that zero-to-one homework is my are 917 - > the two most critical things you can do. 918 - > And then the third is be aware of their incentives.

919 - > Like a venture capital firm exists. 920 - > And yes, some of them might be lovely and some of them might be 921 - > mean, and some of them might be able to introduce you to the 922 - > on-person or help you in a specific challenge. 923 - > Most of them is capital to deliver a very important 924 - > function to scale and then exit. 925 - > So again, being aware of their incentives, be aware of what you 926 - > are trying to achieve.

927 - > And then if you are going to raise, make sure you treat it 928 - > like a numbers game. 929 - > These it's a very ruthless, hard slog, and you need to make sure 930 - > you're putting enough volume at the top of the funnel if you 931 - > expect something at the bottom. 932 - > So that are they're kind of the top of mind things that I would 933 - > say. 934 - > Well, I'd love to disagree.

935 - > SPEAKER_01: I don't. 936 - > Now, on a slightly lighter note, what would your partner say are 937 - > your best and your worst trout? 938 - > SPEAKER_00: Oh gosh. 939 - > I don't know if you've ever met an Irish person with 940 - > self-deprecation or compliments, it's not something we naturally 941 - > come to.

942 - > I think the UK are the same. 943 - > I'm quite optimistic, blissfully optimistic, and I I live on 944 - > energy and enthusiasm for anything. 945 - > You know, going for uh we're going for a nice dinner 946 - > tomorrow, and I can't wait for it. 947 - > Like the energy I bring to small things in life and big things in 948 - > life, I think hopefully is reverberates to other people 949 - > around me, like the people I hang out with.

950 - > I like to be hopefully very enjoyable and optimistic to be 951 - > around and a good problem solver because of that. 952 - > Um and I like to think I'm considerate and doing things for 953 - > other people and putting people first, whether it's in work or 954 - > Ellen, my partner, or anything like that. 955 - > I think there are two things that I at least aspire to be 956 - > like very real. 957 - > We would call it in our in Ireland being sound, which has 958 - > two meanings being kind and being reliable.

959 - > And I think I would love to be described as sound and 960 - > optimistic if I was to be described as anything. 961 - > There are things that matter a lot to me. 962 - > And then, yeah, flaws. 963 - > Um, I do think there's a selfishness that comes with 964 - > building a company that is I need to check myself with, like, 965 - > particularly if I'm lucky enough to have kids in the coming 966 - > years, there's a selfishness that comes with that.

967 - > And I luckily I have exceptional friends that uh would call me 968 - > out if it ever became too much. 969 - > But I imagine that's a trait you need to watch when you're 970 - > building a company and you're so obsessed about trying to make 971 - > that company a success. 972 - > Yeah, I think that's probably something to be to be very aware 973 - > of. 974 - > SPEAKER_01: Final question: what are some of your favorite sort 975 - > of recent pieces of media?

976 - > By which I mean it could be books, music, whatever, TV, 977 - > film, all that type of Yeah. 978 - > SPEAKER_00: So I actually, for the for about I haven't read a 979 - > non-fiction book in maybe four or five years now. 980 - > I stopped. 981 - > I've I love fiction books.

982 - > My favorite books are kind of general historical stories, like 983 - > fictional stories. 984 - > So the classic, like Kite Runner style books, pachinko, these 985 - > types of amazing books. 986 - > I just love those styles of content. 987 - > It takes me out of the world of bounce and it takes me into the 988 - > real world in a way that's really refreshing.

989 - > So you'll never catch me reading nonfiction books. 990 - > In college, school, I would have tried to read them and I never 991 - > became a reader. 992 - > I never got into it. 993 - > And it felt like work, it felt like I was switching off bounce 994 - > and then switch on like learning.

995 - > And most non-fiction books can be an essay, or they could be a 996 - > long read, or they could be a New York article. 997 - > So I'm generally pro-fiction books for a lot of reasons. 998 - > Big podcast person, but again, because I think so much about 999 - > work all the time. 1000 - > Most of my podcasts are non-startup related.

1001 - > Although, I mean, the best podcast probably like the 1002 - > founders podcast is unbelievable. 1003 - > It's an amazing the deep dives they do on companies and 1004 - > founders or acquired. 1005 - > Sorry, founders and acquired are the two, the company and the 1006 - > founders ones that I love. 1007 - > And Tim Ferris is someone who I listen to since I was like 14, 1008 - > who had a huge influence on my life.

1009 - > Yeah, they're probably the ones from the top top of mind. 1010 - > As he said, I think you'd be you'd laugh. 1011 - > My Spotify is probably sports podcasts, Olivia Dean, Jamie XX. 1012 - > It's like it's nothing that you would put as the like you know 1013 - > founder that is all he thinks about his work because I put so 1014 - > much of my time into Bance that I think when I get to turning on 1015 - > my brain in another way, I think I get a lot of clarity of 1016 - > thought and outside thinking from fictional books, music, 1017 - > interesting podcasts.

1018 - > I don't know. 1019 - > Not everything has to have utility towards growing Bance, I 1020 - > suppose, maybe is the answer. 1021 - > And a lot of that is switching off well, enjoying other forms 1022 - > of content, being connected to the real world, and not just 1023 - > being like an AI obsessed founder who doesn't think about 1024 - > anyone else because I don't think that's very healthy 1025 - > either. 1026 - > SPEAKER_01: Charlie, thank you so much.

1027 - > It's been great, really, really fascinating, and really enjoyed 1028 - > talking to you. 1029 - > Thanks so much, Henry. 1030 - > Well, I hope you enjoyed that episode as much as I did. 1031 - > I can genuinely say that Charlie is one of the more impressive 1032 - > founders I've had on the podcast.

1033 - > Obviously, what he's built with bounce is impressive in itself. 1034 - > But what really struck me is how thoughtful he is about the role 1035 - > of researchers within all this change. 1036 - > There are plenty of people talking about how AI is going to 1037 - > replace insight teams. 1038 - > Charlie's perspective is almost the opposite, that the best 1039 - > researchers have become even more valuable when they're 1040 - > equipped with the right technology.

1041 - > Thanks as always to Insight Platforms for their support, and 1042 - > of course, to you for listening. 1043 - > See you next time.