Why BoldVoice charged from day 1 and ignored the Duolingo freemium playbook — Anada Lakra

0:00I think the worst thing that someone can do is like take some playbooks and some elements of a playbook from one app, combine it with a playbook of a different app and create this mishmash strategy. Delingo strategy has always been huge premium base, right, that they monetize a portion of. And because they have this like all the languages, all the levels, they're teaching the basics of everything. They now have all these new skills. Give this product away for free. Even if you monetize a percentage of it, you're fine. Our philosophy from the beginning is we're not going after this like huge base of learners who are

0:32at basic level. We are actually intentionally ignoring them. Hello, I'm your host David Bernard. My guest today is Anata Lacra, co-founder and CEO of Bold Voice, the AI speech coaching app for non-native English speakers. On the podcast, I talk with Anata about scaling to 10 million in ARR with a team of 10, having every employee do two user interviews a week, and why their hard payw wall win backfired. Hey, Ana, thanks so much for joining me on

1:02the podcast today. Thank you for having me. I'm pumped. Yeah, I'm really looking forward to this conversation. So, you like so many builders built something for yourself. uh [snorts] you uh came to the US and you know what let me I I'm I'm about to tell your story. Let me just have you tell the story because I think it's a fantastic story and a fantastic example of how when you face real world problems that you need solutions to sometimes software is a great solution and that often is the best way to build

1:32something. Yes. So let me go all the way back. I was born and raised in Albania. I lived there for the first 18 years of my life and I moved to the US for college. When I moved here, I had not I had never actually been to the US before. It was my first time visiting. I didn't have any family or friends here. So, it was a pretty big move. And I felt pretty prepared for it at the time because, uh, you know, I'd applied to colleges,

1:56figured out the whole process on my own. I'd gotten into Yale, which was, you know, especially at the time, a huge accomplishment and felt, you know, very confident when I first moved here. So imagine my surprise when I started to speak and people were asking me to repeat myself and were clearly not understanding me and it really again surprised me because by that point I'd learned English for 10 years and uh I was basically ape so passed all the tests right got A's and everything so why were people not understanding me and from that experience I realized that you can learn an language on paper but it's

2:32actually very different to communicate it confidently in real life your pronunciation your accent, your fluency, your conversational skills, all of those things can get in the way of being clearly understood. And I again I realized there was this big gap and all that learning had not prepared me for that moment that I would actually communicate in a native English speaking environment. And uh it definitely played a big role at the time both in classroom settings where I'd raise my hand and say something and people would ask me to repeat myself or later when it came to job interviews uh where again I would have to think about I'm not actually

3:07sure how to say this word. Let me think of a synonym on the spot because I don't want to butcher this. So I realized it's a pretty big problem and uh you know many many years later decided to start Bold Voice to help people again at this last mile of fluency. They've done all this learning on paper, but they still maybe struggle to communicate as clearly and be as confident in English as they are in their native language and the job

3:30of bold voices to fill that gap. Awesome. And then your co-founder kind of has similar story. How did how did you meet him and what's what's his story? His story is that he actually moved to the US when his parent immigrated when he was four. So he's essentially a native speaker but his parents when they moved they were in their 30s and even though you know well educated software engineers they still re face the ceiling of having this accent and feeling like well I can't speak as well and as beautifully as my American colleagues and therefore I don't feel comfortable managing people and they were never like

4:07being offered those managerial opportunities and those promotion opportunities to a great extent due to their communication confidence. So he's seen this from the perspective of his parents which was clearly again very aligned with our mission and uh we actually met in both of our first jobs out of college in consulting and I think we very clearly and quickly both realized that financial services consulting was not what we wanted to build our careers around and we both wanted to actually build something do something more entrepreneurial and so we actually both together left that uh consulting company to build a startup in a very different space, but we're both

4:46now with Bold Voice second time found co-founders together. Uh, which I says a lot about the co-founding relationship and that trust that you need to have with your co-founder. That's awesome. I I didn't look deep enough in your LinkedIn. Was it Was it Mackenzie and you're just not wanting to say it or was it I cannot share cannot confirm more than I [laughter] there. You know, it's funny because I I feel like there are so many stories though. some of those consulting company, you know, they they just have a really high bar to hire and like you can make fun of them and maybe they overcharge for some of their services

5:18and maybe they give obvious, you know, answers to some folks, but boy, I I've known a lot of really sharp people who go on to be very successful founders who got their career started at McKenzie or Accenture or some of these consulting companies. So, like I know there's a little bit of a stigma around it, but like honestly like I just know so many

5:35great people. No, that's very true. But actually, let me tell you this one funny story. I applied to Mckenzie in college and I did not I mean I got the interview but I did not get an offer for Mckenzie and that was one of the cases of those interviews where I felt like how am I going to especially when a big part of being a consultant right is that clarity and the communication confidence. I struggled to actually communicate my ideas that confidently and I actually got an offer from a Swiss consulting firm where my interviewers were non-native speakers so we could get way faster into the core of

6:12the ideas as opposed to like the accent and the communication. And then when I was in business school about you know five six however many years later I got my offer from McKenzie because by that point I'd worked on my communication and on my accent. So, it's a kind of a really interesting again another another validation point of the the problem we're going after in a little bit of a

6:34funny full circle moment. That is so fun. I'm glad I asked because that is such a fun story and and a perfect example of of why you built this company. Uh so then take me through you were co-founders with Ilia is his name, right? Yeah. Um so you and Ilia founded a separate company. maybe just speedrun how that went, but then like you know the start of Bold Voice, did you raise a bunch of money? Like how'd you get that first product into market? What did that look

6:59like? That first startup that we did was a recruiting technology company and we were bringing more data analytics primarily through a third party references to employers so they could make better hiring decisions from the start instead of suffering the the the cost of hiring a person was not the right fit. We were 23 24 at the time. Uh it was our first startup. We had barely had any any experience but it was an incredible learning experience of how to build the product, how to talk to users and in hindsight there were a lot of again the typical mistakes that first- time founders made. We made the same

7:33things. We started obsessing with the solution before really understanding the problem and talking to users. So I learned a lot through that experience that you can just build a product in a vacuum and then bring it to customers. You actually have to involve them from the very beginning of that process. And we we had some traction with it. We had over a dozen companies using it, thousands of candidates going through it, but we just didn't have quite the momentum to raise a big round at the time and be able to continue with it. So eventually we sunset that company and I through the that experience though found

8:05my passion in product and in building the product experiences talking to users, understanding and converting that into like beautiful functional UX and and again product features that solve problems. So with that I went into a different uh series B at the time startup in New York. Then from there I went to both product at Pelaton which was really fun and another kind of consumerfacing product that at the time was you know exploding and it was a really fun

8:33experience to to build product there. And then we started BoldVoice when I was in business school and I was again thinking about you know having done a startup before you realize how difficult it is. you realize that it sounds sexy from the outside but actually the inner workings are just grinding nonstop and the more you accomplish in a way the more the intensity ramps up. So I was very picky about exploring ideas because I knew that if I was going to do something again it needed to be something where I could see myself spending the next decade of my life working on that problem. And the

9:09boldface idea came to me because again in a way it had been in my head for for a decade since I first moved here. And it was also inspired by what I was starting to read again this was 2021. This was before chat GPT and some of these latest things that we're seeing now with AI but already I was starting to read more about audio AI becoming more powerful. So the idea there is, can we build something with AI that can hear you as well as a coach would? Because for when it comes to speech coaching, accent coaching, you can't really just give people thumbs up, thumbs down. You

9:43really need super precise feedback for it to work. So at the time, it seemed like with how audio was progressing, we could build those ears on the machine to actually build a product that could function like a speech coach in your pocket without needing that person in the room. And so that was a thesis and you know now we're we're here 5 years later and we're still building around

10:05that same idea. Yeah. So so then did you raise money? Did was this a side project where a business school? Did I quit his job? Like what was the actual founding story? Yeah. So I was 6 months shy of graduating from Harvard Business School. I decided to say no to my Mackenzie return offer so I could focus full-time on this. And I give myself a deadline. I was graduating in 6 months. I told myself within those 6 months, I need a term sheet. And if I don't have a term sheet, then I can start applying to jobs, but if I do, of course, I'll continue this full-time. And in the

10:40meantime, I'm doing this 110%. I'm not entertaining any any other projects. So that was a great that that land was a great forcing mechanism to have that 100% again focus that I think is needed otherwise a startup won't happen if if your attention is split in a in a lot of different ways. So got to work very quickly. The first thing that I realized is I'm the user but I'm not the expert here. I'm not the speech coach. So first thing I did was to reach out to a bunch of different speech coaches who are experts in the field. I had this idea that if we got Hollywood speech coaches

11:15to work with us that it would be especially differentiated and very cool because these are people who work with celebrities and actors and they often also have an acting background themselves. So they would be really great people to have involved in this project. So that was my first thing that I did reached out explored conversations with them. I also once we had our first coach Ron who's incredible and still works with us now 5 years later put together a pilot with him where he recorded a bunch of videos I found a few people about a dozen people who wanted to who were willing to do a beta run of

11:49this and we were sending them manually these videos uh and they were giving us their speech recordings back and emails and the coach was giving them feedback asynchronously. So we're just doing the dry run of super manually without any lines of code being written just to see how this would actually go and work when we actually had a real product out there. So with all that being said, then I started to pitch investors and through that process, we landed our first term sheet of around 500K. And then once we had that term sheet, we also applied around the same time to YC. Because we had that term sheet, we had them

12:25accelerate our interview process. So, we had the term shoot in hand. We also got into YC. So, at that point, Ilia quit his job and we're both in it full-time. Awesome. And so, since you went through YC, you both moved to the valley for what, 3 months to go through the program? We would have had it not been that this is summer 2021 and co. So, we actually built it out of New York. I have some slight regrets that that was the case and we didn't get to do that. Although we do go there every every few months or so to see our adviserss there and to

12:59like a lot of our investors are based there. So it's good to go back but yeah we didn't have that inerson experience. So in a way we we yeah we're we didn't have any distractions. We're just in a tiny office in New York just working 14our days to make as much progress as possible.

13:16When did the product actually launch? During YC? Yeah, we launched it uh in the first couple of weeks of YC. How did that go? I know you were charging from the very beginning, right? Which was maybe a little counterintuitive compared to what you know a lot of venturebacked companies would do, especially at the time in 2021 when money was free and there was still a little bit of a just like build it, grow

13:38it fast kind of mentality. Mhm. Yeah. The biggest thing here in the way I think about charging is that especially early in the early days, you really want to figure out who is your customer. Who are you building for? Because the tool that you build, although it has a value proposition, right? That could be pretty broad. There will be an ICP that will be your initial target audience that you need to hyperfocus on and make the product better for them before you can think about multiple ICPs. And so charging for the product is a way to already figure out who that ICP is. And this way you

14:12don't get feedback or worry and get distracted about feedback or usage patterns or issues with people who are just never going to be your ISTP. But you can really figure out who has the willingness to pay. If they have the willingness to pay that generally correlates with, okay, this problem clearly resonates with this person. Let's then narrow down and focus on making that person's experiences as uh

14:33great as possible. It makes a ton of sense. Not everybody does that, but it it really does make a ton of sense. How did you structure that? Was it just a standard 7-day free trial to a hard payw wall? Did you have any kind of like free tier or any kind of free stuff that people could continue doing in the app or was it just pay or

14:52leave the app? At the time we had a very simple, I believe $10 a month that was just the standard payw wall. We always had a oneweek free trial because we believe that people do want to need to experience a product before they can commit. we were not trying to, you know, stick them with a product and a payment that they regretted, but having again that free trial was still a much stronger signal than just totally ungated. Um, and then yeah, we were not, especially in the YC days, we're not trying to revolutionize, you know, the the business model. Let's just do the triedand-true just like one week trial

15:31simple $10 a month so that we can focus our energy on actually building a good product experience building a good onboarding building the core product instead of just having to do like a ton of experimentation on the payw wall. Obviously that came later. We've done a lot of variance on the payw wall and the pricing packaging. And today we actually lead with an annual subscription of around $150 a year. And then we also offer a monthly as a as a fallback option for people who prefer that. And we also have a higher tier that comes with a few additional features like AI chat which comes at $200 a year. So we

16:07have certainly evolved it but I still think my philosophy is do optimize the payw wall. It's important, but don't overoptimize the pay well compared to the actual product value that you're delivering to users, which is the onboarding is that storytelling and that prepping the user for a successful experience. And then obviously the core

16:27day one plus product experience. Yeah, I I love that perspective and and the way you put it in the notes was optimize product value, not just the payw wall and onboarding. I kind of harp on it a lot in this podcast. So, I get to get up on my soap box again because you put it so eloquently, [laughter] but if you're if you're spending so much time optimizing the onboarding and the payw wall and you're getting a ton of trials, but you have a massive trial cancellation rate, if you think that another two pages in onboarding, another hack on your payw wall is what's going to get those trial conversions, that

17:00that's not how it works. It's because people actually get into the product and they're not finding value. like that's a massive optimization lever on free trial conversions that not a lot of people are really spending enough time thinking about after the pay wall. So, I love that you brought that up and that's how

17:17you think about it. Yeah. I mean, look, all these hacks of like, oh, just one copy change and it drives 20% more people. Like, they're so tempting because you're like, oh, I'm just one iteration away from unlocking this huge product value. And of course, again, I'm not saying do no testing at all. It is an important product surface area, but I think people have taken it maybe to an extreme where they're putting so much effort in that as opposed to again working on the core product experience. Another really interesting thing that we found in some of those experiments is that the initial

17:49metrics actually look really promising. For example, the start trial rate does increase with a few changes that you might make. And I can give specific examples if that's helpful. But at the end of the day when you actually look who converted uh did those payments go through or did they fail? Did people ask for refunds the delta after all that work with the bottom line is actually we found for a lot of these experiments not that big if not sometimes even negative even though the headline number of trial start rate is higher. So, I've kind of gotten a little yeah disillusioned with these like again almost get rich quick

18:25quick schemes that feels like paywall optimization is becoming and it's like at the end of the day just make a better product for people and that's how you drive more revenue. Yeah. And to be clear, I mean some of those hacks work and you're experimenting with them. You try stuff. But I love and I'm just going to read it from my notes because I love the way you frame this is that your the ultimate judge, the ultimate metric you should be measuring against is not trial starts and some of those early funnel, not even like initial conversions and stuff, but it's net revenue after refunds per

18:54exposed user. And then here's the kicker on a matured cohort. How did you decide on that specifically as your kind of gold standard metric to measure any experiment against? It came from the experience of running a lot of these and implementing some changes that looked really good and in hindsight again at that time maybe we just didn't know that we had to wait and see all these other metrics mature and also we didn't have instrumentation. It was really easy day one to see start trial rate, right? And

19:25that's where you get the most data. That's where you know you're you're most likely to get a a p value that is statistically significant because you get so many people at that level. But what we learned and started to implement soon after we we started to really iterate on the payw wall is that's not what actually matters. you have to actually look at the metrics of actual subscription conversion which again that's that's like a pretty well-known thing. I don't think that's new. But the more important thing there is that some of those subscriptions even though initially they convert the payment doesn't actually go through and a lot of

20:01the the refund rate can be a huge delta. So what I found is that anything that kind of forces or tricks people into starting a trial you'll pay it when it comes to refunds. And that is a worse experience because those people who ask for refunds are generally pissed. They're not happy that they have to go through that process. You're also creating this negative user sentiment

20:24around the product that you don't want. And then of course the matured cohort piece is important because people won't ask for the refund necessarily on day one. They'll ask for the refund after they've actually seen that it the payment hit their bank account. So you have to think about okay trial start on this date give it an extra week for them to convert give it at least two extra weeks for the dust to settle from refunds. So actually again it's very tempting to make a quick decision on trials but you have to give it more time to see the entire story play out. Yeah, it makes so much sense, but it's hard to

20:58do and it's hard to find the discipline and then you like you were saying the tooling I mean revenue cap we just you know in the last year or two finally expanded the experimentation window where you can actually go back and look at 3 months and six months and I think we now it's like over a year that you can look back uh and and you know we've tried to make that really easy because it's one of those things like in if you're running your own analytics Like it's just hard to do that and then have the discipline to go back and check it and all that kind of stuff. It's all

21:31these things that make sense, but then you actually have to execute. You actually have to do it and then have the discipline to keep looking back and maybe reversing some of those decisions that you were really excited about. Like I mean because I would imagine you do sometimes pick a winner on some of those early metrics, then you don't see till 3 months later that it's that it's something you actually wanted to reverse. Does that happen frequently or

21:55how does that work for you? Yeah, I I don't think we've implemented a decision that we've had to reverse, but yeah, there have been many things where we've been really excited about the initial result. And the hardest thing is not necessarily the discipline and checking back in the data. It's the fact that it takes just way longer. And you have data you're excited about, you want to be able to action it. And the idea of just sitting there on something that could be really promising for an extra month or depending on x number of months depending on how much uh again users you need per to to validate and

22:27and get to statistical significance. That part is hard of like just sitting on that and and being patient and it also means you can't iterate as fast right you can implement those learnings as fast. So I get it and I was that impatient person. What we found is for example, I'll give you uh two examples actually. When it comes to the the payw wall, we tested a hard payw wall where we basically removed the X dismiss button. So people had to start a trial in order to proceed. And then we also tested a version where there was still a dismiss button on the main payw wall,

23:01but then if you dismissed it, we gave people an offer of X% off that they could not X out of. So they would have to start a trial if they hadn't started at the baseline price. They would have to start it at the lower price. Again, it's a trial. It's not a payment. So they still can't cancel it. But both of those look so promising on the trial

23:21numbers. And it makes sense, right? People have already put in all that time in the onboarding flow. They actually want to see what this app is all about. And we're like, great, this is great. Why are we sitting on this? Why why why shouldn't just we put this in front of everyone? And the answer came down to the refunds. Then the refund spiked by

23:38way more than the gain in star trials. And again, it took us a while to get to to that data, but once that data came in, I think it was very clear that that was not the right thing to do. So again, I I feel like I've been a broken record already about this. But I I think my takeaway from this is that just the mechanical clever things don't actually work. And you'll pay the price for for those things. And the only thing that actually works is once we have something that generally makes sense when it comes to the pay walls, just work on the

24:09product. So, how does it work now? You said you tested a fully hard payw wall and then you tested the X to a fully hard payw wall with a discount. What happens now if they X out of the payw wall because you said that you don't really have a premium experience, right? We have a very limited premium experience. So, if you ex out of the payw wall, you're taken to the speech

24:30assessment. You'll be able to speak. you'll still see the assessment results. We'll break down all the things that we heard. It's already a pretty valuable experience. Like some people work with coaches to just get that assessment. So, we give that for free and then you can also access the rest of the day one experience and do a practice set and I believe also coach video for free and then only on day two you can start a trial at that point if you wish to continue. So we've kind of yeah found the middle ground and let's give some of the product experience but we also have

25:01an experience where we have daily plans. So it's easy to say okay the first daily plan for free on us after that again you don't have to pay right away but you have to start a 7-day trial to continue. It's essentially a reverse trial at that point because you are like giving value but only for that first set of lessons and the speech assessment. These are the kind of things I think more people should be experimenting with is is there something is there a great experience I can give to folks that helps them see the value of the product but that there is some kind of limit to it and a great

25:38example I just brought this up on a live stream yesterday I've been looking at trading card games apps a lot in the past few weeks because we're going to do a live app valuation on subclub of one of these trading card game apps and it's like a card scanner to see the value of your Pokemon cards and Magic cards and things like that. And one of the really cool things they did, and I hadn't seen this exact user experience before, but what they do is they give essentially the full value of the product, but then they limit it to a certain number of scans, and then they make it really

26:08clear. So, you've got your whole scanner UI where you're scanning your Pokemon card and it says like, you know, 35 scans left and then you scan one and then 34 scans left and then you scan one, 33 scans left. And so it's really prepping you for that moment that the payw wall comes, but it's giving you the full value of the product and a really fantastic experience generally. I mean, some of them are really bad at scanning [laughter] the cards as I've [gasps] learned. My kids are into both Magic and Pokemon and they've been pulling me in recently. So, I've been actually using them with their card collections and

26:42I've collected a few now. So, some of them actually are a bad experience. And and this is kind of your point though is like those apps that are a fantastic experience. It's way better for me to have that experience and then use the product and then I'm going to be a product user for life. I'm going to be, you know, actually subscribe versus if I had done a 7-day free trial. It looks really great. I started the trial but then I immediately canceled because the product's not delivering. So if you can actually deliver value and people actually care, that kind of finding a way to do a reverse trial, finding a way

27:14to do some kind of limited experience, trying to just find ways to like give value really preps the customer for that next payw wall that they're going to be way more likely to convert. And the cool thing is that at least for us, you don't need to do that at the expense of that payw wall and onboarding. There will be some people who through the onboarding experience have seen enough about the product that they're prepped to start a free 7-day trial. There will be some people who are more skeptical and maybe they have been burned in the past by free trials or other apps that have shady product

27:49experiences and they even just putting their credit card down, they don't want to do it even if it's for a free experience. So for those people, you can offer that. I think offering for everyone and just removing the free trial payw wall from the experience. I think that could be tricky because you can also maybe risk that people think, oh, this is just a free product and then get kind of upset when you show them a payw wall. So, I think we've kind of found an interesting nice middle ground where we do show them that pay wall, they can X out of it and they can get a

28:17little bit of a free taster there before they they commit to putting their credit card down for that free experience. Do you think you'll ever have a more kind of premium tier where people can get some value and continue using the app or like what's your philosophy on that? Because I know when we were prepping for the episode, you mentioned that that you really enjoyed the uh Jim Canu, the uh chief product officer at Dolingo, his episode like their big thesis at Duolingo is that that fremium base is

28:45the growth engine that they lever. What's your philosophy on that? Do you ever see Fremium or do you think that this form of kind of reverse trial and giving away value is enough to keep propelling the business forward? Jim is great. I'm a big fan of him and the work he's done at Dualingo. I will say though that it depends on the it depends on the product and it depends on the the business model. And I think the worst thing that someone can do is like take some playbooks and some elements of a playbook from one app, combine it with a playbook of a different app and create

29:15this mishmash strategy. Delingo strategy has always been a huge premium base, right, that they monetize a portion of. And because they have this like all the languages, all the levels, they're teaching the basics of everything. They now have all these new skills. Give this product away for free. Even if you monetize a percentage of it, you're fine. Our philosophy from the beginning is we're not going after this like huge

29:39base of learners who are at basic level. We are actually intentionally ignoring them. We're not teaching the basics of English. We're not teaching words. We're focusing on a very different audience that is actually somewhat ignored or dep prioritized by a lot of these apps which is people who already have a pretty good command of language and want to take it to the next level. So people are intermediate to advanced level and for whom English is important enough that

30:02they want to put more effort behind it. And those people are generally career professionals. They care more about English. They're willing to pay. The alternative to what we offer is a way more expensive one-on-one speech coaching experience with a English teacher or an accent coach. So, it's a totally different audience where the intent is much higher and the ROI is also way higher than as a hobby learning the the basics of Spanish or Japanese or something like that. So, totally different ICP, different value proposition, different willingness to pay. And so our entire business has been optimized around that. And it makes sense that again for for Dualingo if

30:43they do way too heavy monetization at the sacrifice of their free base that's the free base is what got them here. So that doesn't make sense. Whereas for us that was never the case. We're always hyperfocused on this person who has a willingness to pay if they get a product that solves that need. And if obviously the the cost of that product is a significant reduction to working one-on-one with a coach, which is the case for us. As you say that though, it does make me think that at some point, you know, as you grow and have, you know, more time and resources to experiment with, then maybe there is a

31:18cheaper tier that is aimed at the the more casual but still willing to pay. And again, I think this is something if you're building an app and you start to see because I imagine this wasn't like from the get-go you realize, oh, we're going to sell this to professionals who are willing to pay hundreds of dollars for this. So, if you start seeing yourself pulled in that direction, like embrace it because it can like, you know, $150 a year and then a premium tier above that that's 200. And honestly, like you're probably just scratching the surface of what you could do. And then you mentioned you even have

31:51one-on-one coaching that you know it's not a huge part of the business but is like $400 $500 an hour. So there is like massive willingness to pay when it's on that side of things. And so by not chasing that consumer and lower willingness to pay market, you're building a really fantastic business on that higher ROI, higher intent, more disposable income and and all those kind

32:10of things. Yeah. And that served us well till this point. And I I'm not going to say we'll never deviate from that strategy. I think for sure there will be experimentations with the packaging especially as we go right now we're still primarily serving people in the US who are immigrants and who also have that higher ability to pay whereas outside of the US spoken English fluency is sometimes as important right because you can get a higher paid job you can get better opportunities you can go to study at a different international university and so forth it unlocks even in some cases a bigger delta uh between

32:46the opportunities you could get without that clear English and those you can get with clear communication. But of course the the base ability to pay in general is lower and I think as we do more of the international strategy we certainly can revisit and should revisit like the localization and things like that. So, I'm not saying this is a strategy necessarily that we will rigidly keep as we continue to scale, but certainly the principle that we're serving a different learner level where they are higher intent and willing to pay for a product that solves their need instead of just like a hobby learner. I think that will

33:23most likely continue to be the case. How did you figure out that that was your market? Well, in the early days, we created this product. I went back to the basics and I said, well, I was an international student when I first experienced this problem. And it's like such a almost a cliche playbook, right? Let's just put this in colleges because college students are great adopters. And again, this is what happens going back to my point about like you take a playbook from a totally different product like dating app or something where they all starting college campuses or Facebook famously or whatever and you apply it to

33:54product where the ICP and the value prop is super different and of course it's not going to work. So what we found in that process is that it's not like international students were a terrible audience for us. I mean we still do get a lot of students in the app, but we clearly realized that a they are a bit younger and they have more they still have some plasticity and they're still surrounded by native speakers in very social environments. So their rate of improvement naturally will be higher without needing this additional resource if they just focus on it themselves. The second thing is the willingness to pay

34:26and the ability to pay for college students is of course lower. And also just the the the fact that they maybe have not seen the price of their communication as much because they're not yet in a meeting where people will overlook them and listen to the native speaker over them. or they're they haven't yet been in a job interview where they realize that the reason that that didn't work out was their

34:50communication and all of those things. So both the willingness to pay and the realization of how much that impacts your income and your career advancement has not happened yet. So we found that actually the audience that resonates the best with this were people who were in the workplace already and who were already seeing the cost and therefore the ROI of investing their career and of course they were making having actually income that they could put towards this and then is there has there been any kind of more direct B2B play with companies with and and because I did see on Hacker News the founder of Expo said

35:28that they talked about h having an employee expense this as a solution to help them improve their job performance. Yeah, we've seen that quite a bit with just inbound companies will reach to us to make this happen for their employees or we're also seeing that people will you know buy this and then have the company pay for it which is great. We haven't pursued it as much as a intentional go to market purely because of resourcing. We're still a very small team and we've gotten to this point by being relentlessly focused on the consumer experience and I think the muscle of building a B2B company or B2B

36:08go to market is pretty different and I think it'll make sense for us at a future stage potentially but for now the consumer market is so big that we've just let it be our main intentional focus and then we get this nice additional effect of inbound from companies because they know that the consumer product is so loved. Yeah. One of the things you touched on there and we've kind of been touching on throughout the entire conversation is user obsession and focus. I love your approach to understanding users. So, why don't you take me through kind of the layers of how you speak to your users, how you interact with user, and how you

36:48let that impact the direction of the product. So we've put a bunch of things in place to make the user obsession not just something we talk about but like a muscle that we keep fresh because I think the user base is always also going to change and their expectations are going to change. You're going to bring

37:05in naturally newer people as you expand. So it's really important to not just have this do a bunch of user interviews at the beginning and then be like okay that's our user and then just like not evolve that understanding. So the things that we've implemented is the for example we send out automated invites to users and we pick a variety of different people. Sometimes it's subscribers who have been with us for two or three years. Sometimes people who just started trial a week ago and converted or didn't convert and we ask them to jump on a call with our team for 15 minutes and share their feedback and that invite

37:38automatically invites the people on the team who have availability in their calendar. So what that creates is that basically every single person on the team, myself, my co-founder, our engineers, our growth people, everyone is in at least two user interviews a week. So we're actively talking to people for at least 15 minutes like twice a week. And that's just one example. Another example is we've created this super user group and we hold weekly or so calls with them. And that's a great opportunity for this is the people who already engage with the app a lot to figure out how do we make these experiences better for them. And

38:15where that's been really impactful is to actually show to have our engineers demo new features that we're working on and get feedback from those users on those features so that they're a part of the product evolution and decision- making before the the feature is even out to users. So that's been really powerful to figure out what is making sense, what's not making sense. One thing about our product is that it's deceptively very complicated when it comes to pronunciation, speech. It's not like teaching grammar where there are correct and incorrect rules. There's just so many elements and details and ex uh exceptions and nuances and it can get really complicated to

38:57build which is what we want like a really user friendly product experience while also not making it over complicated and intimidating to use and understand. So it's really important to get our real users eyes on features to validate is this actually making sense before we ship it. So those are just two quick examples but sometimes we also bring people in office and we watch how they actually use the product and we watch what they tap like what they're spending most time on. Of course we also have analytics around all of this uh that we also watch but being able to actually verbally talk to users and

39:33actually see them interact with the the the product I think is an underrated skill. So building these again automated ways for us, semi-automated ways of course then we actually are in these calls, but semi-automated ways to to have more user facetime has been super helpful. I love that level of customer obsession and it speaks to being such a small team and we'll get to this later. You're only 10 people and at $10 million in ARR, which is incredible. A million dollars per employee as as a small scrappy company is incredible. But that user obsession, you know, you're meeting with the users two a week, everybody at the

40:08company. So it's 20 calls a week at least. You're have a super user group. And then you bring people into the office. There's like even layers to it. But that that qualitative feedback and seeing people actually struggle is so different than the quantitative like just look at the analytics. And then when you see the analytics and you're like, "Oh, it looks like something's going on here." then you actually can like verify that with the super user group or with these inoff testing and those kind of things. So I I love that like kind of multi-layer approach and then you're also like not overindexing on the super users because

40:43you're having meetings with the random folks and then you're actually having people in office to get like the human touch. So, I I absolutely love how you've built that and structured it to kind of touch all those different points and not kind of because, you know, it is a bit of a trap sometimes where if you only did the super user group, you'd probably be pushing the product in directions that work for them, but not some of the more casual users. And then if you only did the kind of casual user sampling, you may miss some of the

41:12things that the super users care about. And so by layering it all like that, you're really getting the the kind of breath of qualitative feedback that you just never would get in analytics and never would get if you didn't do it in this multi-layer strategy. Did that just come about organically or how did you

41:28decide on on that approach? I mean it's evolved over the over the years for sure and more recently again we have been figuring out ways to build more and more automation. All of us here in the company use cloud a lot. that we're figuring out how do we do these things and so that we don't have to if because if you have to have a person email manually, you know, a bunch of users and that's great one-off, but it's a one-off. And so these automations help us again make sure that we maintain that muscle instead of just making it a one-off effort. But I will say though to your

42:02point about the diversity of perspectives, one really important flag is that you're always going to get more feedback and you're going to get more people to jump on calls who are happy users because they're like, "Oh my gosh, my favorite product reached out to me for my feedback. I want to book this call." And people who did not have a good experience are going to be less likely to jump on those calls. So it's really important to keep that in mind and to be aware of the bias in the feedback in the features that people are also asking for. So a lot of the time people who did not have a good

42:32experience or did not feel like the product met what they were looking for. Those people might not want to jump on a call, but if we give them just like a three question mini survey in an email where they don't even have to click on anything, they can just like respond in an email, they'll be more likely to answer that. Or when someone goes to cancel their subscription or their trial, having like a mini flow there

42:53that asks them what actually went wrong. So yeah, I think just being aware of that bias and making sure that you're not overly indexing on like making those happy people happier, which is important, but that could also lead you toward like making this like super niche, super technical product. And you also want to create a product that you know has as especially as it grows wider

43:13and wider appeal. Such great advice. Gosh, there's so many things I want to talk about. We don't have a ton of time left, but maybe we can kind of speedrun some of the verality and kind of growth things y'all have done because you you've gotten some really cool growth wins and some viral moments and things like that. What would you say were the kind of maybe top three most impactful lessons on the growth side or or top three kind of viral

43:37moments? Well, for sure the top one was the X and Oracle that we built. So yeah, our team had this idea that if we could build a tool that could predict where a person was from just based on 10 seconds of them speaking, it could have this kind of viral potential. And so we built and trained a model in house to to do exactly that. So I think this also speaks to again a really important part of our success is that we have an audioML team inhouse building these and training these really differentiated models from scratch with our own data and we also have the expert speech

44:14coaches who are a huge part of how we label the transcribe and just train these models to be better. So that's a really big differentiation in our product and the Axent Oracle was built based on one of those in-house created models and um it worked. Uh it was the most magical thing I think we've done and the best part about it is that the way that it it spread is that every month I write the investor updates. I've been doing for 5 years. That month that that the Axen Oracle was built, I sent it out I think that even day after or so to our investors. One of them was an

44:49angel investor. It was Korean and he was like, "Oh, this is cool. I'll put it on threads and he posted it on threads." And then I believe the next day we had 100,000 people in Korea try this thing. So again, wild and unlike anything we've seen before. So of course when that happened, we're like, let's make this better. Let's let's invest a lot of time in making sure that making the model more accurate, making sure it's not doesn't crash and all these things. And it was a really wonderful way to get people familiar with the product. And we we made it like there's no there was no

45:20gate. There was no free trial. It was like pretty magical experience. You don't even have to download the app. By the way, my co-founder actually made this decision to just put this in a separate web app where you just go to aoracle.com. You record your voice instantly. It spits out your feedback, your results. So, it was optimized for this frictionless viral moment that then exposes people to Bold Voice as a brand and then invites them into the app to continue discovering more things about their speech and then actually improving and working on their pronunciation and

45:51speech as well. Has that been fairly evergreen or have you taken lessons from that in your ads and other viral content? For sure. Yeah, it's actually surprised us how much it actually has lasting appeal. There are people who use it, who keep coming back, it will randomly go viral. Again, with viral things, what makes them beautiful is also what makes them tricky. You can't overengineer them. Like with ads where you know exactly, I'm spending this much, I'm getting this much out of it. But one thing that has proven true is that when we invest into it. So, for example, we recently added this really cool feature

46:26where it doesn't just tell you the country where your accent is from, but it also goes down to the city level. And uh once we implemented that, it had another viral boost because again had this new unique edge to it. So yeah, we we can constantly see it be have this evergreen value with like kind of random unpredictable spikes, but we have a few levers where we can boost it by adding new features or another thing that we did is that we showed people a snippet of the sample of the the full results that they would get in the app. So they're not just just seeing this result

47:02and then they have to make the leap of faith to go onto the app. We're telling them, by the way, here are some of the other things you'll see on the app. And that invites them to get to the app or it boosts the conversion rate to get them to the app. Yeah, that is so cool. What's the number two top growth hack lesson, viral

47:18moment? Probably when it comes to viral things, the second thing that we did, which is uh again kind of funny and unpredictable, it's just like social media and Tik Tok. There was this trend that was going around with people crying on Tik Tok when the app gave them low scores. And in particular, there was

47:35this one creator who god I I don't know. She got like 300 million views in like a few months or something pretty crazy like that. Again, completely not boosted by ads. So again, another really cool way of getting the product in a more light-hearted, accessible way into the kind of people's feeds. So we're not just doing like this some more serious marketing, but we also have these other lighter ways to get in front of people and make them aware of bold voice. Of course, when you do humorous content, you're going to get huge viral spikes, but they're not going to have the same level of intent that you can get from

48:15like, you know, paid ads, for example, or word of mouth, obviously, where you can really target the messaging, but it's still a really nice, I think, machine to have working alongside the paid marketing efforts. Well, for number three, then I want to ask, what is your kind of breadandbut growth tactic? You do you rely on paid ads? What's the kind of the bulk of what

48:35you put your growth effort into? I think it's a mix for sure of productled, viral or sharable features plus uh paid marketing. On the paid marketing side, I know it's very vilified and a lot of people are against it. And I think that's taken to an extreme and realistically even those people who are saying ads are bad are actually running ads. My experience with it is that there there's a good way to run ads and a bad way to run ads. The good way to do it, I think, is to just

49:01be fully aware of your unit economics. So you're not spending money on again low value customers when it comes to again LTB over CAC. So there's quite a bit of discipline there that needs to happen to make sure that the unit economics are strong. And one really powerful thing that I have found in that process is using payback period to your advantage and making that as short as possible because it gives you upfront a lot of clarity on a user's LTV. And what I mean by that is when I already referred to this before with having most users by default being presented with the annual plan as opposed to the

49:40monthly plan. What that means is that with an annual plan, we get the cash up front from users after their 7-day free trial. So, we know right away from a particular channel, it drove this many subscribers. This is how much we spent. And then we know exactly how effective that marketing was. As opposed to monthly where you get the initial monthly subscriber, but you don't actually know will they retain one month or will they retain 24 months. And different channels actually do produce

50:08very different durations by customers. Especially Tik Tok I found to be a flimsier channel in terms of just intent and and loyalty or attention span, whatever you want to call it. So it's really helpful to have annual as a way to add that discipline where up front you know what your LTB over CAC is at least for the first year. And also by getting instant CAC payback that gives you the safety that you can then reinvest that money directly into

50:36acquiring the next customer. Yeah, it makes a lot of sense. Gosh, so much I want to talk about. Well, let's dive into being a 10erson team and hitting 10 million an air or as a 10erson team. How [laughter] how do you do that? It's a good question. I think it's a mix of things. Probably the biggest thing I'd give credit to is just the people we've hired and how much just everyone is a doer and an owner. I think everyone here drives a ton of value for the business. And part of how we do that is by hiring people who are kind of this founder types who want to have that huge

51:15impact and who want to be very ambitious with what they build and who don't require also a lot of you know direction and coordination. They can you know are the types of people who can just get things done largely independently. Another thing that we do as a company is minimizing distractions to building. So we have one teamwide company meeting on Mondays where we just all have an opportunity to sync and everyone is

51:40aware of what everyone else is doing. But other than that, it's really heads down time and again talk to users. So being very also focused on who the user is super impactful and having everyone work toward solve that user's problem instead of just building and writing code. So I think a big part of it is is thanks to that and now of course also with the latest AI tooling that we have now right with Claude everyone in the company is using Claude for the first time in my life I'm actually shipping PRs myself which is crazy I'm not a technical person and I'm not here to

52:14claim that now I'm the highest like producer of code in the company I'm definitely the lowest but just like the amount of like ownership and freedom and and just agency that everyone gets and that's just amplified with the fact that this tooling allows us to build even faster. Another example is our designer. He has done an incredible job not just building things on Figma but translating them to prototypes and even getting them to almost 80% there when it comes to then the engineers taking it over and coding it. So we've also optimized the workflows so that we can get things iterated on and shipped even faster. I

52:54realized as soon as I said that some people are probably going 10 million in ARR that's not that big and 10 people that's a lot of people but it's a very different business and that's why it's so impressive. This is not you spent 15 million on paid ads to get to 10 million in ARR. This is not you had one viral moment and blew up to 10 million ARR and you're just going to kind of you know tank after that viral moment. It's like you've built a repeatable machine, you have machine learning and AI experts on the team. Like you've built a real company and have done so profitably. So

53:24yeah, speak to that as well. Like a lot of people who get to 10 million, it's a very different business at the 10 million than what you've built. I feel like again, we've been around here for 5 years now. We're not like an overnight success story. We're also not like an AI rapper where we found like there's this arbitrage short-term arbitrage opportunity of just like putting something out there in front of a user building around a specific model and that means that I mean hopefully again right like our value proposition is more durable our business model is more durable the investment we've also done on the AI and ML capabilities has

54:02also proven to be the correct investment because still to this day you cannot talk to chat GPT or claude or whatever have and get this kind of feedback that BoldVoice gives you. And we made this bet that when it comes to non-native English speakers speaking English, that is a too maybe niche of an opportunity

54:20for the more general models to go after. And yet, it's a huge opportunity when you just look at one company going after it and how big the market of non-native English speakers is. So yeah, we've invested a lot in again building a product that doesn't go away in the next iteration of Chad GPT or claude. But of

54:40course the landscape is always changing. So we're also spending a lot of effort and on the technical side on ML side across the board to continue to innovate and make the product better and better and always a step or hopefully more ahead of where the more general generalized models are. Well, the last thing I did want to touch on before we wrap up is the kind of shift away from just focusing on accent to kind of a broader mission with the company. So, I wanted to give you a chance to kind of talk through that. For sure. So, from day one, we've actually called ourselves an accent training app. And the reason

55:16for that is that the vision has always been we need to help people speak more clearly and confidently. But we also know that the biggest pain point that we're hearing from a lot of people at this last smile of fluency is their accent. And it seems like such a differentiated and spiky pain point that we can create a product and scale a product initially that focuses on that wedge. And now we've reached the point in the market, in our brand, in our revenue as well where we and also we raised the series A that allows us to build towards that vision of getting to the next level and using this wedge to

55:50expand into this larger market of spoken English fluency. Again, we're now going after early level learners. just the focus is still largely the same learner. We're just going maybe a step earlier instead of last mile going a step in the penultimate mile of fluency and then expanding from just accent coaching into also things like confidence of of speech

56:12ability to convey your ideas clearly. Did you use the right phrase? Was it idiomatic? Did you use the right sentence structure? Did a maybe a grammar mistake slip in there from your native language? Is your point coming across? and all these things that create confident communication beyond just particularly pronunciation and accent. And so this is just to me a natural evolution of the product uh and the next concentric ring around the core that we've built around accent training that feeds into our goal ultimately of there are a billion people out there and English is for better or worse in the global language of business and the one

56:47language we universally have decided to adopt and speaking that language at the most confident and fluent level is going to create a lot of opportunities for a lot of people. So, let's build every piece that it takes to get someone to that complete fluency and confidence. I'm curious personally if this expanded TAM also includes things like coaching against uh verbal ticks and and I mean as you were saying all that, I'm like, I've been podcasting for six years now and I still say um and like and like I I don't feel like I'm even the most confident communicator as a native English speaker. Is that ever part of

57:26who you might serve with this product or or do you think that that's just a whole different product? It's a good question. We've talked internally about speech coaching for native speakers as well and just more general things. My impression of that market though is that it feels more like a nice to have and a one-off problem as opposed to this repeatable higher stakes problem which is what English coaching for non-native speakers is. And because we still have such a big market and we've just scratched the surface of that market, our intention for the foreseeable future is to stay focused on that person. At the same time, there are

58:01a few things that our model is getting better at hearing that could be applicable both for native speakers and non-native speakers. Things like vocal fry or filler words and other things like that that we're also again in the process of of adding to the product that could benefit everyone. But for now, yes, the focus is clear on non-native English speakers because we think that that's the both the bigger market and the sharpest pain that people are

58:28actually willing to to pay to solve. Yeah, I I am so glad I asked that question because it's just a perfect example of as a product builder, you can go chase all these like, oh, maybe this is a bigger TAM and maybe maybe we could serve those people. But your answer was we're focused on the most critical painoint with the highest willingness to pay. It's just such a fantastic answer and such a clear great almost summary of this whole conversation is that you're very focused on a very specific pain point that has a high willingness to pay that matters in somebody's life that matters in their career that matters in

59:06how they communicate even day-to-day life. And so it's just such a cool kind of summation of your approach to all of this. episode is a great great place to end on but I do want to ask the three questions I asked uh every guest now at the end is what's the most impactful experiment or change your biggest win over the last year maybe what I already shared around the adoption of these automations and coding tools to make everyone even more of a builder than they were and make people be twice if not more productive and increasing that velocity it's almost a trope now in 2026 but

59:42golly Even just in I thought I was a heavy AI user, but just in the past few weeks, I've like completely changed how I work with computer use, it's just getting good enough where it can just like do stuff. So, I just send it out on task like it's an assistant updating spreadsheets and fixing this and like it's pretty incredible. So if you think you're super AI native and fluent, keep trying because it's every I feel like every two or three weeks there's like you just get deeper and deeper and deeper and it's it's it's getting crazy good. Anyways, what's the worst experiment or change that you've

1:00:16implemented in the last year? Your biggest fail? I already spoke about this, but the payw wall stuff. Just stop overoptimizing every single word and every button and everything. like optimize them for a little bit and then move on to the actual product. Yeah. Uh great answer. Great answer. You won't hear me argue. [laughter] Um and then the last thing, fill in the blank. Growth would be easier if if for us if accents and uh speech was not so taboo to talk about. We get so many people in user interviews who tell us they love this product. It's improved their lives so much. But on a day-to-day, are they actually, you know,

1:00:55posting about this on LinkedIn as much as they're telling us or are they bringing this to their colleague? There's this kind of shyness about talking about something like this because you don't want to offend people. And at the same time, I feel like this does them more of a disservice because at the end of the day, again, it's not about accents, it's about communication confidence. And that's a big part, by the way, of why we're also doing this expansion. First of all, it is to give people a better product that better solves their need and captures more of the things that makes them struggle with their English beyond their accent. But

1:01:28second, it is also to kind of escape a little bit this almost antiviral word of accent. And uh that's why also the accent oracle was so fun because we kind of turned it on its head and we made a fun viral experience out of a word that is not as people are tend to shy away from. It's always a a thing that it's on my mind of like as we grow, it's really important to figure out how we combat and and how we boost the word of mouth and kind of combat this idea of I don't want to share this product with others

1:02:01even though it's very helpful for me. And I think there are other industries that went through this and have seen an evolution like mental health was something where people did not talk about it and then it almost feels like overnight but of course not overnight with a lot of hard work from a lot of people involved it then became something that people are talking about more publicly and it got distigmatized and I hope that we're part of that conversation of doing the same thing for

1:02:23speech. Yeah. And I love [clears throat] your framing too because it's not and I love even kind of the pivot away from overly focusing on the accent because it's not about like completely removing an accent unless that's what you want because I mean some people do want that. They want to just you know be so clearly a native English speaker that people would never even think they were from another country. But like you know if you're from Spain and have a little bit of a Spanish accent like I mean I I think accents are cool. Most Americans think, [laughter] you know, accents are cool, but but then it just does come down to

1:02:56communication is that it is sometimes harder, especially for native native English speakers, and I've mentioned this to colleagues, you know, my European colleagues, I feel like uh have way less difficulty with certain languages and understanding people with those accents because they themselves aren't like you brought that up. that like when you're not a native speaker, it's easier, but when you are, uh, I have found it more difficult. Um, but I love that that it's not about removing the accent. It's just about helping be a more clear, confident communicator. Such a great mission and such a great kind of

1:03:31expansion of the market. Absolutely. And again, that's the the goal is to have everyone be as confident in English as they are in their native language. Accents are a very small part of it. At the end of the day, the goal is confidently communicate and get your point across and always be understood without having to repeat yourself or be asked, "Hey, where are you from?" from the first moment of of opening your mouth, which is something that we

1:03:56actually hear a lot in user interviews. Um, so we just want to remove those roadblocks while also allowing people to be as as confident as themselves as they can possibly be. Yeah. Amazing. And fun place to wrap up. So, thank you so much for joining. This is such a fun conversation. Likewise. Thanks so much. [music] Thanks so much for listening. If you have a minute, please leave a review in your favorite podcast player. You can also stop by chat.subclub.com

1:04:24to join our private community. [music]