How ElevenLabs Builds, Prices, and Grows AI Consumer Apps

Welcome to the Sub Club Podcast, a show dedicated to the best practices for building and growing app businesses. We sit down with the entrepreneurs, investors, and builders behind the most successful apps in the world to learn from their successes and failures. Sub Club is brought to you by RevenueCat. Thousands of the world's best apps trust RevenueCat to power in-app purchases, manage customers, and grow revenue across iOS, Android, and the web. You can learn more at revenuecat.com. Let's get into the show.

Hello, I'm your host, David Bernard, and with me today, RevenueCat CEO, Jacob Eiting`. Our guests today are Jack McDermott and Tanmay Jain, both working on mobile growth at ElevenLabs, the leading AI voice lab. On the podcast, we talk with Tanmay and Jack about how earned media can drive paid performance, building features that make for good Tweets and why stripping out your onboarding quiz might beat optimizing it.

All right. We've got a special one today with two guests and Jacob with me as well. So let's go one by one so people familiarize themselves with your voice. Jack, thanks so much for joining us on the podcast today.

Jack McDermott:

Hey, it's great to be here. Thanks for having us.

David Barnard:

And Tanmay, nice to have you.

Tanmay Jain:

Yeah, super psyched to be here as well.

David Barnard:

And Jacob, always great to have you on the podcast.

Jacob Eiting:

I'm excited to be a fly on the wall today with our big crew, but I'm super excited. I'm a fan of ElevenLabs and all this stuff. You guys have been super ambitious in a bunch of areas, so I'm excited to talk about AI, what's been going on.

David Barnard:

And that's exactly where I wanted to get started. ElevenLabs is super ambitious, and one of the cool things y'all are doing is not just being a foundation company and an API company servicing all sorts of apps and enterprise solutions, but you're building consumer apps. So I wanted to start with why build consumer apps as a model company.

And I mean, in some ways people listening might see it as competition. They're a developer, they want to experiment with voice. Why are you competing with your customers on the API side? But I think it's a really cool way to build a company. So yeah, I'd love to hear thoughts behind that.

Jack McDermott:

So I think an important part about a ElevenLabs is we really think of ourselves as a researching product company. That's kind of how we got our start that it all goes back to the research at the beginning, and then at the end of it is being able to deploy that out as individual products. And you may be a developer and that consumers that are built on that through an API. You may be a creative person or creator who builds on top of our web or on top of our mobile interface, and ultimately we're beginning to move towards consumers as well.

So we really think one of the underlying parts of ElevenLabs do is kind of the blurred line between creator and consumer. We think that that line is beginning to blur a little bit. So at the end of the day, we want to have a direct relationship with users, with consumers, with creators, and ultimately with developers at some of the biggest enterprises. So yeah, we kind of start from the first point of being able to build on top of great research and then to deploy it in a variety of ways and products.

David Barnard:

It's so powerful to be the first best customer of your API as well, and having those consumer products can be a superpower for the API. It's a better API when internally you can be testing against it and know what the experience is of building against that API. Anything to add there, Tanmay?

Tanmay Jain:

No, I think Jack nailed it. That's exactly it. We're an AI research lab, but we're actually not, we're a product and research company. And we're very deep on the product, very, very deep in the craft.

Jacob Eiting:

There's a few, OpenAI being maybe the biggest example, and just kind of a surprise, violated what I expected was going to happen, which I thought all these companies were going to be behind, back of the scenes, enterprise, API companies. Traditionally, that's been where the focus is and where the scale is. So have you found it hard to be both? How does the company ... Because I'm mostly an API backend company, I'm struggling here. It's a lot. So how do you both balance the needs? Because it's very different to build ... Obviously there's some overlap there.

Jack McDermott:

One thing that I'll say that I think is really unique about ElevenLabs is people might think of ElevenLabs as one enterprise or one scaling startup, but inside of ElevenLabs, if you were to take a look at it, we really look like and feel like maybe 10 or 12 individual startups that are all building, shipping, and [inaudible 00:04:37] all at once. And so we keep our teams intentionally really small. Our mobile team is about 12 people right now and growing. But each team that we kind of set up as a pod or a speedboat, as we like to think of it, has their own kind of self-contained development, design, growth, product, and can kind of freely operate within that.

And that's true across all things voice agent to a creative platform to research team, which against some of the bigger labs is surprisingly small. And so we kind of like to find and play on that specific balance of small teams with big goals and big opportunity. And that's been one of the key things that I think from the outside looks really unique about ElevenLabs.

Tanmay Jain:

Yeah. And it's definitely something I noticed as well. I run the ElevenLabs flagship app on mobile. There's a lot of times where we have to think about how does this behave in web? How does this behave in mobile? How do we manage that crossover? And the guidance that we're currently gotten is just go ship, go ship fast. We trust you'll make the right decision. And if there's any gaps along the line, we'll figure that out and we'll ship our way out of them as well.

David Barnard:

One of the interesting things y'all are doing as well is creating multiple apps. And we do see this with OpenAI creating Sora, the health app was a spin out and ChatGPT, of course, being the kind of main app. But how do you think about being a multi-consumer app product company?

Jack McDermott:

One thing I think that's kind of part of ElevenLabs is from the beginning we've built a very horizontal platform. You can create speech, you can create sound effects, you can create music, you can create speech to read aloud an audiobook or to be able to play into an ad. So I think we have a very horizontal set of use cases that you can build on top of. And one of the things that we try to think of on our team when we build apps is really what's the form factor and what's the ideal use case that kind of meets an individual where they are.

And so if you're a short form video creator on the go, you may need or be able to use a subset of those horizontal capabilities, whereas if you're a consumer that's trying to convert a PDF into lifelike audio to be able to improve reading and comprehension on the go, that's its own separate use case as well. So I think we really try to start from when we're spread that far as far as what we can do as a platform, you kind of have to force yourself to think about what does the ideal user persona need? What does that look like as a form factor? Is an app the right approach to be able to take? And at least twice now, we found that's just the right way to go about it and has allowed us to continue to move really quickly and be super, super focused on specific users.

Tanmay Jain:

Sorry. And to share an example, I previously was at Canva and that was a story where we actually had a super app and they tried to spin out a separate app and actually ended up coming back to bring it into one core app. And the reason behind that was, compared to ElevenLabs, which is horizontal API focused, while Canva serves to be the design tool for everyone and semi horizontal, it is one product. And so they tried to spin out a version called Canva Stories focused on Instagram stories. And the product was amazing. It won Apple Awards. It was beautifully designed, very, very unique use case, which was very prevalent at the time, 2019, 2020, but it just could never crack those acquisition problems. It could never crack the retention, it could never crack brand awareness. People came to an accepted, "Well, we want Canva, this is only a unique ... This is a subset." And so because of that, it just never really took off.

And so it takes a very unique and a very precise company structure to make this succeed. And the benefit as well is we now have two different testing grounds for a bunch of different things for two very different use cases. And so, so often I will go and test something or Jack will go and test something and we'll swap notes. For example, we're testing paywalls right now and we're going through an exercise where I'm testing a couple of hypotheses, Jack's doing the same, and we're curious to see what... We almost double or quadruple our surface area for learning.

David Barnard:

And so just to be clear, Tanmay, you run the flagship app, the Eleven... What's it called? Just ElevenLabs?

Tanmay Jain:

It's just called ElevenLabs, yeah.

David Barnard:

And then Jack, you run ElevenReader, which is the even more kind of consumer focus where you can upload a PDF and generate audio. A really cool app, by the way. My son and I have been playing with a text to voice and he was tinkering with wanting to build an app and we went right to ElevenReader. I was like, "Oh man, this is so good." So how do you all think about that kind of cross-functional, not stepping on each other's toes, but incorporating those learnings across both apps?

Jack McDermott:

Well, I think one thing that's been interesting for us is we've actually had to intentionally create a bit of space between the two apps because when you go into the App Store, the Play Store and you type in ElevenLabs, you might be expecting to get a specific result when you're doing a branded search like that. And if you're trying to find an app that allows you to turn anything into an audiobook, you might not actually be top of mind with ElevenLabs.

So we've tried to really intentionally create our own brands and our own kind of look and feel for each app, but we share a common design system, a common set of technologies. A lot of our developers have worked on both apps by now and have a lot of experience between the two, and then have a few others that are kind of in the test phase or in a small bet right now that we're incubating. So I think it's just being very clear about the role of each app and the role of each audience that you're trying to serve and being smart about to what extent do you have to dis-intermediate or pull the two apps apart and run them both. And I'd expect that we'll build and launch a lot more as we go forward too.

David Barnard:

And how do you think about growth with the two separate apps? I'd imagine you're doing some ads and growth focused things very specifically for the Reader app, which does have that very different kind of use case and product promise. And then with the main ElevenLabs app, I imagine you're doing very separate things, but then you can kind of like as a brand also kind of be the text to voice company. So yeah, how do you think about growth across both apps and the brand?

Jack McDermott:

There's a couple of shared parts that I think underpin both of the apps' growth. One is, from a fundamental building block level, we share a lot of the same technologies. So when it comes to having to stand up on a database, an MMP, testing tools, these are a lot of the shared parts of the apps that we can lend and move together. So when it comes to growth, we can also lend a lot of things between the two apps. So creating great UGC content to be able to boost on paid or to be able to post on organic, a lot of times those creators or agencies are doing a lot of the same things across a variety of brands and we can kind of pick and choose what seems to be working well on the other.

Or I think a good example is in the MMP world, like being able to map and define specific events between apps that show the greatest amount of signal, even though the underlying use case of the app is going to be different. So that's where I think it starts to change is really when you think of a retention and engagement and monetization, that's where we diverge quite a bit is that you have to build specifically to that use case and to that audience.

Tanmay Jain:

One of the things that this affords, and this came up in the super app conversation versus separate apps is we were really able to tailor paid plans or monetization directly to the value that that ICP or that user seeks. So when we think about ElevenReader, they don't think in terms of credits and tokens when it comes to a listening or a consumption focused app. So Jack was able to really align the costs to an hour's experience and a consumption focused experience versus in ElevenLabs, the flagship app, it's very, very creative focused and users to our benefit have been trained and primed to learn about credits and characters and all these sort of different digital economies. And so by separating out, we were able to really basically get licensed to go and try out a variety of different pricing and packaging in a really, really unique and useful way.

David Barnard:

Since we're on the topic of user acquisition, I am curious about how y'all think about running paid UA and kind of building the engine of growth. And I know you use a lot of really polished videos. I heard another podcast that one of your colleagues was on that you even hired a motion designer specifically to get that earned media that is these big launches, but then you kind of trying to build a paid UA flywheel that runs continuously while also doing these kind of big launches. So tell me about how you think about those things and then how you actually execute.

Tanmay Jain:

We always think about our launches as two buckets. There's moments we go out and seek earned media, and then we have those moments where you need to go on forecast sustainable month-to-month growth. Both things have to be true at once, and that's something as a growth and product person you constantly have to do. And even within our earned media, we think about what level we kind of targets at. Is this something that goes on all our channels? Is this something that go on email? Maybe people don't care about on our LinkedIn, the ElevenLabs flagship app because it is short form creative focus. So understanding who fits where and why is really, really important.

And then on the paid side, we think about, okay, well, what is the channel? What is the product? What is the specific message that we're trying to do? And how do we set the right sort of models so that we can figure out how to do this in a really sustainable way? Because we look at that as a sustainable growth channel and not as a let's go and launch continuously as well. So I think the gaps between launches aren't like a traditional company where it is once a year, once every six months, we would probably do more than once a quarter, to be honest, or I would say monthly is probably the average cadence for a team in ElevenLabs. So we get a lot of opportunity to explore these sort of viral upside moments while also having these continuous growth engines.

David Barnard:

What are the primary channels that you do use in paid?

Jack McDermott:

One, we operate across a lot of the paid ad channels that you'd expect. So we're operating across Meta, TikTok, Apple Search ads, Google Play, Google Ads, and really taking a fine-tuned approach to be able to looking at the results and the data in an incremental way each week or each month. I think one thing that's kind of unique about ElevenLabs, and again, going back to that idea of a horizontal platform that we have is we're in a nice position where we don't need to always optimize for say day six conversion or day eight version or even day zero conversion. We're in a position where we do have and we're trying to build a brand and we're trying to build an audience and we're trying to drive engagement that we ultimately want to capture a lot of value from.

But one thing that's kind of nice about the position that we've been able to build with a great deal of brand work at ElevenLabs through all of these flagship launches where we announce new models and gain a good amount of earned media is it gives us a little bit more flexibility on paid to be able to test new creatives or to really not have to nail that week one payback, but to really allow ourselves a bit more flexibility in order to determine and test into the right payback or LTV to CAC. So that's kind of the way that we thought about it.

And of course at the end of the day, it's a lot of work because each new channel is going to operate in its own way and you really get to understand what type of creative is going to perform, what's the right approach across each one. So that I think is an important part, that we don't view it as just paid in the aggregate, but a really fine-grained approach to each individual channel in the way that they layer up together toward growth.

Tanmay Jain:

And I think earned media is something that a lot of companies really underrate. I think earned media is really a compounding effect. Every launch we do earns more eyeballs, which means that the next launch earns more eyeballs and it is truly compounding. And then it compounds, not that we've ever run the numbers on this, but I can say with confidence, it's compounding down in our paid media as well.

The first place we start is ElevenLabs branded search. It's the lowest hanging fruit. We get a lot of it because we do a lot of earned media and we're able to go and scoop that up. And I'm pretty sure when people get hit with the ads, they've heard about us, they like us, they know that we provide the most expressive text to speech model. And so when we hit them with an ad about that in a really unique way, they're more likely to convert as well. And I think that results in better sort of CPA or CPI, whatever we're looking at.

Jacob Eiting:

Yeah. It's hard to tell what the efficiencies are. It's hard to measure, obviously.

Tanmay Jain:

Exactly.

Jacob Eiting:

The brand is always really hard. But yeah, do you all ... This is an old trick back in the day, but launches and coordinating with the platforms was a huge advantage at the time. It used to be launches were bigger on the App Store, for example. Do you think about cooperating or working with the platforms on what they want to see, like your App Store reps or your Play Store reps? And are you thinking about that very much or is it more like, no, we're not bigger than that, but that's not really our main lever?

Jack McDermott:

We do have great relationships with our reps and with the teams at both Apple and Google. But I think the way that we've approached things is like you have to start, especially nowadays, with the power of your own story of what the launch is and how you want to own it and tell it. And we've been able to really invest. We hired our own internal creator who's on staff to be able to film and create short form content.

I think for a long time, the pendulum has been swinging from who has the story to tell? Do you have to rely on a platform to be able to get on the front page of the App Store or can you really create an organic or earned or paid approach to being able to get a really great story out there?

So we spend a lot of time upfront, probably one to two weeks of kind of refinement of what's the story that we want to go out and lead with. And there's so much noise out there, especially in our space, where you're building and launching AI models that seem to be a new-

Jacob Eiting:

Yeah, every week.

Jack McDermott:

... state-of-the-art benchmark every single week, right?

Jacob Eiting:

Change a name, put a number on the end. Cool. Beat some benchmark I've never heard of before, great. It sounds awesome.

Jack McDermott:

Yeah. Right. And then so I think in that specific environment, if you don't start with a very clear storyline yourself, then you're not going to be able to rely on anybody else.

David Barnard:

Yeah. Tanmay, I'd love for you to dig deeper into that. You were telling me at your time at Amazon, the whole idea of write the press release first and how that's taken on a different form at ElevenLabs. How do you plan for that?

Tanmay Jain:

It was really interesting. I think six weeks in, had to go off to my first launch, we don't have long onboarding processes here at ElevenLabs, and straightaway got handed a doc and Jack was super helpful with this where it was like, okay, write out your core value promise, your secondary value promises very succinctly, very tightly. But then the first thing you're going to do is turn that into a Twitter thread and we're going to think about what is the main headline? What's the interesting hook to get you to the second thing that gets you to the next sub-Tweet and the next sub-Tweet? And how can you encapsulate that story in terms of a number of small Twitter threads that really capture the audience and drive that earned media?

And then the next activity is, well, then how do you turn that into a video and how do you express that through video? Because video is so important in all of these channels where it's driving so much more performance and organic impressions. And that looks like sometimes we go and make a motion graphic video where we have our in-house team who are forever churning out the highest quality videos I've ever seen, and I don't know how they do that, all the way to more people in camera focused content.

And it's something that as a company, I feel like you get trained on because you spend a lot of time watching all the content that the team puts out across these various different sort of tiger teams shipping all this really great content. And then you also just get refined. You put a script out there, the creative team will give you some feedback. You write a Twitter thread, Jack or Luca, our head of growth will write some really meaningful feedback. So you get a lot of opportunity to get these little reps out as well.

David Barnard:

I think it's such an underrated kind of product and growth lever. If you're working on a feature and it doesn't make for a good Twitter thread, you better have a really good reason to be building [inaudible 00:21:17].

Jacob Eiting:

I get a lot of crap sometimes at internally at RevenueCat because it's like, "Oh, we manage by Twitter interactions." And I'm like, "Well, it's not the worst guiding force." You know what I mean? At least you know somebody's mad or wants something, right? And it's true, if we build something and I post it and nobody cares, it's like most likely the feature ... There's a correlation at least that stuff that slaps on Twitter tends to be really well received.

And to replace X, Twitter with wherever your customers are, obviously it's just a proxy for being where the users are and staying really in ... It's the classic YC advice, just talk to your customers. If what you're building doesn't resonate with them, why are you building it? Who's it for exactly? So I'm all for vibes based, like product roadmap setting and stuff.

David Barnard:

Any lessons from your time at Canva, Tanmay? What was that a way you approach things at Canva as well?

Tanmay Jain:

Canva's a bit different. So my time was a bit earlier in the process, it was 2019, 2020, 2021 to 2022. So very different time in terms of the market. This is very much pre-AI before GPT 3.5 became a big thing. And so there it was much more kind of old-school. I mean, we still had around the media, but when we launched the video product, we had a 12-month run-up, we spent three months figuring out our messaging.

And yes, we had a great video on the backend. We had a lot of the same things that we do now, but that time period and that time horizon was extended massively. And it was all about, we're launching on this day and we've got the news articles going on exactly that day. We've lined all these big ticket items up. And I can't speak to what's happening internally now, but definitely in ElevenLabs, it's launch fast, launch often, really nail it and understand what resonates.

David Barnard:

Yeah. That's got to be a lot of fun moving to more at that startup. I love the way Jack described it early on. It's just such a cool way to operate a company, especially a multi-product company, is that you're almost like little startups inside the company. And so I love that kind of flexibility and focus for teams to be able to focus on the individual products they're working on and move fast. So that's really cool.

And then speaking of that, one of the unique things I was reading about the team structure at ElevenLabs is that engineers that you hire are filtered against having a solid product instinct. And you almost think of the engineers more as the PM and then the more PM type folks are more growth leads. So how does that work?

Jack McDermott:

Yeah. One thing that we've seen is we have a belief that we want to either be building or growing, right? And we kind of view those two sides as the key to cogs in the wheel to kind of drive our ultimate growth. And so what does that mean? On one hand, I think that means that what we see at ElevenLabs, and this is true for myself and for Tanmay as well, we have a lot of former entrepreneurs, a lot of former startup founders on the team, probably the highest amount that I've been at, at least in the company by far. And then on the engineering side, we look and we screen for those that build side projects on their own, that can identify a user pain point and really build and be obsessed with that.

A lot of engineers that join our team just love the ability to kind of like the freshing feel of being able to engage end to end and own a product cycle from idea stage all the way through [inaudible 00:24:46] execution. So it's been a bit successful as well, like self-selecting the types of engineers that want to build and own in that style of culture. And I think it's self-selected on the growth side as well.

David Barnard:

Where do you slot in there, Tanmay? Are you a growth product manager? Are you an engineering product manager? What's your role specifically?

Tanmay Jain:

Yeah, me and Jack have had multiple conversations about this and it is whatever it takes to make the product succeed.

Jacob Eiting:

That's what I always tell people. People are like, "What's my job?" I'm like, "I don't know. Help developers make more money. Figure it out."

Tanmay Jain:

Quite literally. I will merge sometimes very questionable PRs that my engineering team loves. I will review Figma designs, I'll share payroll experiments myself, I'll be reviewing creative content that our agencies are putting out. It's everything that needs to happen and you just kind of figure it out. And I think that's why it is like catnip to founders or former founders.

Jacob Eiting:

Yeah. Well, I was going to ask and follow up with that is like, what have you done or what does the company do to kind of, I guess broad scope there is one thing, but is there something they , obviously being just at ElevenLabs is exciting, it's good to be at a company that's growing and doing something unique, but is there anything they do to get that ...? Because I find it's hard, people that build stuff, they want to build stuff, and often their own stuff, right? To get them to jump on a team is tough. So what does ElevenLabs do to get folks to come into the board? It's a good environment probably, which I guess, Tanmay, that's kind of what you were hinting at.

Tanmay Jain:

Yeah, I can say what the sell to me was, I've only been here for about four months now, but when I came in, it was super fresh and the sell was very much be a founder. I've been at big companies, I've been at slightly smaller companies. My wife works at a PM at a large FAANG company as well, I see her day to day. The last thing I'd want to do is spend my life in PR and legal reviews as a former founder who you would ship something and get feedback the next day.

And then you come to a place like an ElevenLabs, it's like, "Hey, we are a successful company. We have amazing quality talent who take extreme ownership on everything that they do." And you get to go and join this A class set of people and you get given full autonomy to go own a space and move as quickly as you humanely can and want to. It's like catnip for a founder. The only rate limiter really is your own capability.

Jacob Eiting:

Right. It's tough to set up. I was going to ask, do you have an overarching, for example, marketing function that's like setting, okay, how do we talk...You're doing a growth campaign and they look at it or something, because it's something I think we sometimes, I don't want to say struggle, but I do like setting teams to just do whatever, but then your challenge sometimes is consistency and the voice. How do you do that in that setting?

Jack McDermott:

Yeah. One thing that's been nice too, so we have all of these small pods that build and grow individual products or apps. And then of course there's a set of shared or org-wide roles that help to bring everything together. So think of everything from operations to creative and brand to SEO and our web development side. So at the end of the day, there are some things that are important enough and that are wide enough in their scope that we want a cohesive brand or we want a fully operational organization.

So I think it's trying to find the balance between where does a traditional hierarchical or matrix kind of work? Where do you need that shared support? So every little team is not having to rebrand themselves, but there's kind of that shared set of guidelines. And there's also just a lot of work that goes into ElevenLabs to make sure that there's standards and guidelines around brand and comms and ops, like the ability to do that now on a really small team too, because you have access to a ton of AI tools to help you expand that impact. So there are a ton of shared roles that can drop on and you don't have to reinvent the wheel each time.

Jacob Eiting:

So you can just like grab off the shelf like the, "Hey, here's our standard how we do this." And then maybe you mess it up and it has to be corrected.

Jack McDermott:

Yeah. Or like, "Hey, what does great look like in SEO?" Oh, well, there's a few folks here that have decades of experience in that specific area that might lend well to work that we're trying to do on mobile web to app or the way that we're trying to solve a specific problem. So being able to be very collaborative, like at the end of the day, I think we all view ourselves as part of the ElevenLabs team and have a really tight-knit group. It helps to do a week-long offsite that we've done the past couple of years, we...

Jacob Eiting:

The whole team, like everybody, the whole company.

Jack McDermott:

Yeah, the whole company. So we do...

Jacob Eiting:

It's in the hundreds probably, right? If I'm guessing... Yeah.

Jack McDermott:

Yeah. Now we're going up to 400 right now and we just got everyone together in Italy this fall for a week. The year before that, we all met up in Croatia. So to build both the human connection and then to let great people operate, I think has been the key.

David Barnard:

Yeah. As Jacob was alluding to, it does seem like it's a really hard thing to build those kind of teams and get that kind of founder mindset. I'll say personally, the engineers I've had the most fun working with on my own products are those really product-minded engineers? How do you filter for that? Is there a specific way you frame the job rec, the interview process?

Jack McDermott:

We take a good, hard look at public GitHub profiles to see who's engaged in what. And we look at specific companies too that we know have really strong product building cultures and try to attract folks that want to join a company like it.

Tanmay Jain:

I was saying as well, it becomes a referral flywheel, right?

Jack McDermott:

Yeah.

Tanmay Jain:

The people that come in, then they also know people like themselves.

Jacob Eiting:

People they want to work with, right? Yeah.

Tanmay Jain:

Exactly.

Jacob Eiting:

It's a crazy thing. There was a period people go like, "Oh, you're just going to be drawing from a limited pool of folks." But the risk reward is just so high to do that. And you have to be careful because sometimes people will refer, "Oh, I knew this person from college. I never really worked with them." You got to dig a little bit and be like, "Okay, how do you actually know them?" And maybe bonus referrals maybe creates that issue.

But yeah, it's often really hard to tell until you've worked with somebody. Even what somebody says they've done, right? Obviously we know the internet, everything's a lie, nobody knows I'm a dog, but it's very imperfect. But yeah, it does just unlock a ton. And I think about engineers and product folks too, kind of just everybody, they get so conditioned at other places, Tanmay, as you were talking about, just waiting for the PR review or the legal review and become so conditioned to not take risks and things like that.

And it sounds like you all operate in this sense, which is like... Especially in a period where the growth is so rapid, if ElevenLabs has a hiccup or doesn't iterate for a quarter, you're done. There's somebody coming way fast. There's so much to be gained right now that there's somebody right behind you ready to take that space. And so it's not that you don't have time, I mean, it's not that legal and PR review and all that's not important. It's just that it's maybe not existential or it's less existential than moving.

And so it's like, okay, how do we balance...? And that takes, I don't know, I mean, assume founders and leaders who are at the company who are tolerant to mistakes and risk is probably one thing. And then making sure that folks know you can just do things, which is not typical corporate culture I think in the world.

Jack McDermott:

The one thing I was just going to say on that note is like, we hear this a lot, that speed is becoming the moat. And I think that we kind of live by that and have tried to really embrace just the speed and size of iteration and the amount that we ship. And so to some extent, it's like, to your point, Jacob, understanding what bricks might fall. And our support team does an amazing job because there's often blips in the model that might come up or specific very important things, but things that you often need to make a trade-off between, are we going to get this model 105% fine-tuned or are we good to move quicker and to continue to push the edges?

And so I think when that becomes the company's guiding principle and ethos, then it's pretty clear that moving quicker is more important than ensuring that there's messy coverage of every little thing. And so I think that that's why these things, the small team sizes, the founder-led background, the speed, they all need to fit together.

Jacob Eiting:

Yeah. You need people who take responsibility too, right?

Jack McDermott:

The advice is not to go out and just find a bunch of former founders and bring them into a very hierarchical culture. It's not going to work. You're going to have a lot of clashes and they're going to leave. And there's going to be lots of ideas that compete. But when you create systems that fit together, small teams with driven, product-minded engineers and founders that fit with this go fast and ship move, then I think that's where you really start to get harmony and you can kind of deliver bigger results.

David Barnard:

Well, one of the challenges with moving fast and then, even as much as I praised it, building toward the Twitter thread or the viral TikTok is that you can sometimes end up building kind of flash in the pan, gimmicky things. How do you think about balancing that kind of moving fast and creating those big launches, but then also kind of building toward this more retentive use cases, a product that is going to be used for years and meaningful for years versus just kind of moving quick and slapping out a bunch of features that don't end up getting used.

Jack McDermott:

It's a really good question. It's one that I can already reflect on over the last year and a half, a couple of the features or growth campaigns that we've built and launched that might in hindsight look like a flash in the pan. One thing that our CEO and co-founder, Mati told me a couple of weeks ago, he encouraged a quote from Brian Armstrong at Coinbase where "Action gets information."

And so I think at the end of the day, if you're willing to take action, you're going to ultimately get more information. You're going to get more information about your market, about your users, about your pricing model than just not pursuing that. So I do think that the way to think about it is that in hindsight, it might be a flash in the pan, but in forward sight, it might actually help guide your future decisions. So I think that that's been one thing that I've felt to be true here.

Tanmay Jain:

When you think about these launches, it's not we're launching for launching's sake. It's actually about what is the most impactful thing that we can build that is going to provide a better user experience? How can we build it as furiously as quickly as possible and then express the value of it in the best possible way? We deeply think about what are the next things we launch. Sure, some of them might hit, some of them might fail, but we're always coming from the lens of like, how does this improve the user experience? Not, wouldn't it be cool if we ship this thing?

Jack McDermott:

And I tend to find, I think you brought this up, Jacob, I tend to find the X for Y to lead you down the wrong path. When I think of the times where we've developed or built or grown tools that haven't worked, it tends to be applying ideas or concepts from somewhere else and being like, "Wouldn't it be cool if...?" Or, "Did you see that? Could we do that?" And that direct one for one application versus starting with what's our unique take or what's our unique story on this or what are the capabilities that we have at ElevenLabs that no one else has? And I think when you get back to that space, you start to unlock a lot more creative ideas that ultimately move beyond the flash in the pan stage and actually allow you to build compound growth over time.

David Barnard:

Yeah. On that note, Tanmay, you had a really great post on LinkedIn about don't just build another Spotify Wrapped. I mean, it's a perfect example of what Jack was saying about how everybody thinks, "Well, how can I build this Spotify Wrapped of my product?" But your post was great about, that's not a great way to think. Tell me more about that.

Tanmay Jain:

Yeah, I saw this pattern and I feel like it becomes truer and truer every year. I thought every year people would learn, but it only becomes truer and truer, is that every year Spotify Wrapped will drop like clockwork and it's amazing. And a random product out there, not to name names, will do their version of a Wrapped.

Jacob Eiting:

Mine this year is Phil's Coffee. I got a Phil's Coffee Wrapped in the mail this year, which that's how I knew it had jumped the shark.

Tanmay Jain:

Yeah. That's how you know when Wrapped has gone beyond mainstream and post cool is probably [inaudible 00:37:55].

Jacob Eiting:

I love Phil's coffee, don't get me wrong, but maybe we didn't need it Wrapped. It's fine.

Tanmay Jain:

Yeah. Maybe we skip this year. So I think the reasons, people forget why Spotify Wrapped really took off, right? It took this passive consumption user behavior and then it turned it into this natural shareable experience, right? People generally resonate and identify themselves around the music they listen to and you present it in this beautiful format and you share it and it goes viral and it's all around a compressed time period. And for a growth PM that's catnip, you've taken a single user experience and you've added a viral factor around it. So certainly you've got a K-factor you can measure and it's super fantastic.

But what people miss is you have to think about what your natural shareable moment in your product is, and I think I put in the post some of the products that I think do this really well. The GitHub Skyline, it's probably something that didn't take that long for someone to implement, but it's consistently sat there. And occasionally every so often you'll see people put a screenshot of their Skyline out there. Or I think even the RevenueCat team, you'll put the billboard up on Times Square and then people will go and stand in front of it and take a photo and then post it. I think the

Jacob Eiting:

I think the [inaudible 00:39:07] on that one is gone, by the way. I think everybody's figured that trick out.

Tanmay Jain:

Yeah,.

Jacob Eiting:

But it's all right. It's how it goes, right? Everything becomes Wrapped eventually and you have to move on to something else.

Tanmay Jain:

Exactly. Everything becomes Wrapped and there's a short half-life, but you have to find what resonates. So even in the ElevenLabs team, we handed out to people who are building on the agents platform, this is how many hours of agents interaction you've had, right? Very similar to the YouTube, you've hit 100K or Stripe, you've hit this sort of milestone. These are natural shareable moments that align with the value that the user gets from the platform that they are personally very proud of and that then they are willing to share. And actually going and figuring that out is really important for you.

Jacob Eiting:

It was really interesting, we've built various versions of share this screenshot or share this chart in our dashboard over the years. Never, almost never, saw them in the wild. The biggest shareable feature we ever built was just building a mobile app because then people could easily screenshot. And I guess, this was not intentional, but making the home screen of that fairly aesthetic and fairly easy to screenshot and share. And 100 X, the number of times I saw our product on X overnight. It wasn't even a sharing feature, right? You know what I mean? I mean, maybe we should have thought, I mean, obviously in hindsight we could say, yeah, clearly we knew that virality would go up with the mobile app. And also just we had a miss there on not having a mobile app for as long as we didn't.

And I kind of thought it as like this toy, just kind of like, "Oh yeah, we're a business B2B. Why do we need a mobile app?" And I may be taking it from Stripe, they have a mobile app, but it's like, how big and important is that for me? It just never was that important be on a certain stage, but I totally missed this. And then in hindsight now, it's obvious. It has almost the same usage on a daily basis as our dashboard, you know what I mean?

And people were yelling at me for years to do it. I just was like, David included, but I just ignored it. But it's exactly your point, right? Is that I was being like, "You guys are wrong, it's what other people do," da, da, da. Rather than just being like, "Why [inaudible 00:41:02] today? Okay, I'll build that." That would have been the better algorithm and would have got me to this sort of unlock much sooner, right? But yeah, it's just like surprising sometimes the second order effects you get. I'm sure the Spotify people never knew Wrapped would become the thing it did, right?

David Barnard:

Yeah. It's hard to manufacture those shareable moments, often they end up coming more naturally than you anticipate. And when you try and build for it, you screw it up. Even one that I think has worked really well is Flighty's. Flighty kind of does a wrap, like the year-end flights and that has gone fairly viral. A lot of people post and that kind of thing.

Jacob Eiting:

But it's a little more compelling than how many coffees did I order this year?

David Barnard:

Absolutely. But Ryan even talked about, it was either on his App Growth Annual ... Yeah, I think it was his App Growth Annual interview with Charlie that was released in the sister podcast of this launch where he talked about they struggled with that. They went through multiple iterations before they landed on something that was actually truly shareable. And I mean, I think that was kind of your point, Tanmay with that post and kind of what you're saying is, you got to iterate towards that real shareable moment and then try stuff. And if people aren't sharing, then it's not because you didn't surface it well enough or whatever, it's like that's just not a shareable moment. And then you keep iterating and you find that shareable moment.

Jacob Eiting:

It's very analogous to like April Fools as a marketing gimmick, right? That was such a big deal for a number of years. And then I don't remember when it was, but at some point it just jumped the shark and it became cringe to do it. You know what I mean? It still happens, but I feel like it peaked and then ... But I guess that's more of these things having half-lives. And it's anything in the attention economy, right? It's like you find something unique, it's able to get attention, then you're going to attract a million fast follows to that space and then it's gone in two years.

But I mean, that's what keeps it interesting, right? That's why it's growth hacking. It's not formulaic or whatever. You have to constantly be thinking of cool stuff to do. We're always like, yeah, we borrow stuff obviously from other brands and platforms, but then we're just always coming up with the most unhinged, as unhinged and weird as we possibly can dream. A lot of it never makes the light of day, but one out of 10 stupid things somebody says we're like, "Actually, you know what? Maybe we can do this." And then you get to multiply that across multiple products too for you guys. It's like, you can kind of go in a lot of directions, which is cool.

David Barnard:

Yeah. Have y'all found any key unlocks in the consumer apps at ElevenLabs? Are there shareable moments or are you still iterating toward finding that kind of loop?

Jack McDermott:

Yeah, we're still iterating a bit. One thing that I've tried to look at and that we look at a lot or, we do a lot of social monitoring. So we take a good hard look at mentions across Reddit, Discord, X, et cetera, and have seen some pull. For us, I think the aha is when you can bring your own ebook or EPUB or PDF file and then pick a specific voice. The voice in your head that you want to be able to read that story or that book to you.

So I think the main goal is, I know that a lot has kind of been made of the trend around onboarding quizzes and if you introduce goal setting and that type of thing, but we've actually found more success in trying to get out of the way a little bit and trying to shorten the time between like, "I have a piece of content that I want to read and I've got a voice in mind, and can we help you find that voice quickly?"

And we do see screenshots and video and screen caps of people with a specific voice being like, "Oh, I didn't know that Michael Caine could read me Dickens or could read me my chemistry PhD homework or whatever." And so I think that that's kind of becoming a unique thing for us and one that we're beginning to put a little bit more fuel on on the consumer side.

Jacob Eiting:

It's all fair use if a user does it, right? But yeah, those are the best moments. I mean, I guess that just comes with giving people powerful tools. I don't know if you call this earned or just virality, but like I talk about people posting our screenshot of our app, right? It wasn't a thing we were trying to make happen. We just did our job and got people to value really quickly, and then it just creates your natural virality and growth that way.

Jack McDermott:

One thing that ties the two apps together a little bit and part of ElevenLabs is just the importance and the centrality of voice as a medium and a tone and a quality to it, and so I think the more that we are... We're currently doing a great deal work around a redesign to make the voice orb experience feel more emotive, right? What would a voice look like and how can we bridge the gap between what you hear and the way when you're going to choose a voice or being able to create a voice clone that would represent your own voice? Can we begin to experiment in that specific area because it is a pretty magical experience to be able to create new audio in voice tones. And so I think that that's been one core loop across voice agents, across our core app and across the consumer side as well.

Jacob Eiting:

Yeah. When do I get custom voice portability to other LLMs? I want that because I tell you [inaudible 00:46:39]. I'm so much more tolerant of the standard sort of 5.2's behavior and answers when I'm reading it. But once it starts talking to me, the person gets so annoying when they start messing stuff up, you know what I mean? And I need to tweak this.

It's so funny, I mean, this is sort of a tangent, but my wife heard, I had this standard, this is ChatGPT we're talking about, but the standard ChatGPT voice agent versus she has the Australian one and she's like, "Why is yours so annoying?" And I was like, I'm still thinking, whatever. She's like, "Listen to mine." And I was like, "Yeah, you know what, yours does sound less annoying for some reason." But I mean, it's an example of what you're talking about where voice is such a important ... It goes beyond just text, right?

It's like it humanizes information and it's the most natural interface for these token brains in a box, right? It's a very natural way to humanize them, which I don't know, just obviously tons of surface area for you guys to invent new and interesting experiences off of, which is pretty cool. How unhinged would it be if I could clone my own voice and make that my ChatGPT, just like talking to myself? That would be very strange. That would upset somebody.

Jack McDermott:

Well, yeah, I was just going to say, we did a quick analysis on the consumer side and we can look at just what's the frequency of voices that people choose and are there clusters or do they tend to congregate? Because most apps would allow you either no selection or two or three options. We have over 600 voices on our consumer app and it allows you to pick your own voice narrator to read on great audiobook content and no single voice has more than 10 or 12% of use. And so the tail is extremely long. The median amount of use is probably close to one or 2% when you think of just how people perceive voice and want that type of selection. And my preferred voice is not the same as David's, is not the same as anybody else. And so I think that's one thing that we're seeing.

David Barnard:

That actually kind of leads into the next question I was going to ask. I'd be remiss to not voice what everybody in the industry of concerned about, and I think ElevenLabs I'm sure is thinking about this as well. But do the big three LLM companies, OpenAI, Anthropic and Google, do they just subsume everything? Are all apps just become a little interface inside of ChatGPT? Is all voice subsumed?

And I think that's a good example of your 600 quality voices. And I mean, same thing with Jacob, I'm not a huge fan of the voices of some of the other LLMs, but ElevenLabs, that's where you lead and that's where you stand out. And it seems like that's one of the differentiation points. But yeah, how do you think about competing with these other labs?

Jack McDermott:

So I think there's a few things that come to mind. One is I think that we have continued to increase speed over time, and so we have continued to focus on voice, speech, and audio being our expertise and have continued to double down there. And so there's lots of great models in video and image and of course in text, right? But we have continued to double down each turn to be able to move fast and keep our edge and stay ahead when it comes to voice.

I think that we're all aware though that to continue to add and gain value, you need to do more than just one thing, and so I think that's part of why we've tried to build. Take our Voice Agents platform as a good example, right? It's not just pick your voice and talk. We've built in custom evals and workflows and integrations that allow you to begin to concentrate a lot of your work around that specific voice experience.

I think the same holds true a little bit of what we're doing with the apps. If you get greater lock-in and greater relationship when it comes to the brand that you have, if you've got a creator suite that extends into mobile, if you've got a set of enterprises that build a great text content that can then reach consumers, part of the approach is that you got to begin to build a bit of an ecosystem and have a lot of interlocking parts if you want to be able to continue to advance.

Jacob Eiting:

Again, one of the many ways I've been wrong in the last two years, and I think a lot of people have, was that like, I think I would've guessed like, "Oh, you'll just be the best voice model and that will be your edge." But you can kind of extrapolate far enough that Amazon ... Yeah, I think of Amazon is like the sort of like Costco of computer services, at some point they will have, Amazon Web Services will have a model that's probably pretty good and maybe not as good as the very top edge, whatever, but it's probably good enough, and so you might lose that edge.

And then it feels like, at least from where I'm standing two years into this crazy hype cycle, it does feel like what you're saying, it's like the advantages accrue into the application layer, right? It's like, what and how...? And that's where Amazon Web Services is never going to be able to clone a voice model. They may be able to clone a passable voice model, but they're not going to have all the little application layer tricks. And I was saying before the call, I played with the audiobook stuff last year and you were saying that there might be more to that coming. Who's going to build the best platform for using a voice model to do audiobook generation, right? Who's going to do that?

Maybe some people, but now you've compounded an edge that honestly it's like stuff that can be tokens in, tokens out, gradient and descent and can be taught that way is going to become commoditized. The stuff that's not is interacting with the people actually using these tools, understanding their human needs. I don't know. I don't know how you guys go. I'm constantly going back between my job is over forever and humans will actually be the dominant software creators. I still don't know. Every day I go back and forth, but that's kind of ...

I mean, following what you were just saying, Jack, that kind of promotes a human-centric version of the future of software development, which is like talking to the people and figuring out how the models actually fit into their needs might become the edge, right? Because everybody will have the models, right? Everybody will have the ... And then that's on the building side too, co-generation and all those things.

David Barnard:

And also the brand side. I mean, the fact that Jacob and I both recognize ElevenLabs as being the best voices in the industry by far.

Jacob Eiting:

Certainly I'm not going to go to AWS first, you know what I mean? I've heard of the good one. I'm going to try the good one and then maybe we'll look at the cost or whatever. But yeah, that certainly is going to give you some sort of advantage long in the future.

Tanmay Jain:

I mean, it goes back, and the media is compounding, right? Every time someone thinks about voice or speech and audio, ElevenLabs is the name that floats to mind immediately, right? And sure, the mod providers will go and make their best effort to release something, but our mix of focus, of relentlessly beating the drum and launching often and pushing the edges faster than anyone else in the market and compounding that on the earned media and compounding that with a smart growth model that really enables that and really empowered engineering teams, all of that just stacks up so that you have this moat. And I don't know what you call that moat. Maybe it's old school, we call that brand or ...

Jacob Eiting:

Brand almost seems too simplistic, you know what I mean? In 2026, yeah.

David Barnard:

Well, I think that's a great place to wrap up. Well, I've got three questions though I've been asking every guest now, and either both of you can answer each one or one of you can take it, but what's the most impactful experiment or change you've implemented this year? What's the biggest win you've seen this year that you think people could take a lesson from?

Jack McDermott:

I think the biggest win that we've had on the app that I lead is really to not lose sight of the fact that yes, AI is everywhere, and if you spend time like us trying to evaluate benchmarks and per credit spend and tokens and all of these things, when you go to meet a consumer who's trying to think about, "Hey, I want to consume a certain amount of content and audio," they're often not going to start from the point of tokens and credits and that type of thing.

So the biggest win is just being able to convey or being able to communicate in simple pricing plans and simple storylines that begin to abstract away from the AI complexity, which is real. There are, unlike in the past with some SaaS apps, there's real cogs, there's real costs that you need to be smart about, but that shouldn't mean that you communicate that way out to a consumer.

The biggest win that we had on our side was just we stripped out an entire pricing tier, we switched over to be able to communicate with the amount of hours that you spend to be able to play great content and beginning to abstract away from a lot of the fine-tuned details that go into an AI model, I think has just been the biggest win on our end.

Tanmay Jain:

I think for us, it's just been simplicity as well, a version of simplicity. And that's come up in a few different ways. Working on the app that I do, we try and bring over the best of web to the mobile app. And so often we talk about distilling it down and not just porting the whole thing over and squishing it down into a mobile form factor. I think it kind of reminds me of the yield days where everyone had these supplementary apps that were either really, really bare bones or they were just basically the desktop dashboard squished into a tiny form factor. And we think about how can we make it super simple and pick the best of what is working on web, what's really resonating with the audience and bringing that down.

And then that went through as well with our pricing, the way we offer I think upwards of five plans. And on mobile, it just didn't make sense for variety of reasons across Apple and place or limitations. But even just as a mobile app, we want it to be very simple in the different pricing structures we offer. And so we offer one price, we test with that and we test with different incentives around it, but it keeps things very simple and it means as well that we hold ourselves very accountable when we've picked the next thing to work on.

David Barnard:

All right, a fun one. What's your biggest fail of the year? What did you screw up that was an incredible lesson that, to your comment earlier, Jack, that provided a lot of information, even if it didn't provide a huge win.

Jack McDermott:

First impressions count and you need to spend a lot of time to be able to get onboarding right in the form of being optimized and being performant, which can mean a lot of things. We probably have spent too much time trying to smooth out and trying to optimize within onboarding. And ultimately some of the things that I would've done in hindsight is like, what's the base case? What happens if you just have a signup form that goes directly into your zero state product? What is the incremental amount of retention or amount of engagement or amount of monetization that this flow is going to unlock when it comes to grip and consumers that you serve?

And so I think without that really trying to assess that scale of incrementality, you're sort of just betting blindly compounding, right? You're trying to double down on something that you've already tried to double down on. And so I think we over rotated a lot in that specific area because there's a lot out there. You read tons of blog posts about this quiz and this onboarding funnel, but I think if you don't compare it to a base case of, just strip it out entirely for 10% for five days and see what's the opportunity size there. And so that was a big fail. We probably spent too many cycles, too many experiments, and ultimately just being simple, clear, and direct about who the app is for and trying to get you to that aha with fewer screens has proven to be the bigger win.

Tanmay Jain:

Yeah. I would say mine is more about my former past as a founder, which is making your price cheaper won't drive meaningful retention if people don't care about the product. Sounds super obvious when you say it out loud, but you go, "At this price point, maybe we've maxed it out." You do the typical YC thing, first we charge 100 bucks a month and then 200 and 400, and then somehow you end up charging a couple grand. And then you see suddenly it flattens and maybe people are churning at the $1,000 price point.

So you go, "Ooh, okay, they tell me there's budgetary constraints, maybe dropping the price will mean that we're able to retain people better." Maybe you retain them an extra month, but if they just don't care about your product, they're not going to stick around. They'll be happy to walk over fire or walk over coals to use your product, even if it's super expensive. And that sounds super obvious in hindsight, but in the moment you can kind of get really academic about it and mess it up.

David Barnard:

All right, last question. Growth would be easier if all my competitors gave up. Sorry, [inaudible 00:59:39].

Jack McDermott:

I was going to say, if there weren't over two million apps on the App Store, I think one that I was going to say is, and I know that we've spent a lot of time on this too, Tan, is just the number of tools that are used in a modern growth stack in the right way has began to proliferate, which is good. Getting all of the tools to talk to each other and to kind of have have true north when it comes to, especially around attribution that's well written about and well known about, but I think at the end of the day, if tools could talk as well as our voice models can talk, if they could talk to each other, then we would save a lot of time.

Jacob Eiting:

That may not be as a high in the sky wish. In another year, that's obviously when pigs fly, but this might be the year these things actually do start talking to each other.

Tanmay Jain:

I mean, my engineer was like, "Why can't we just cursor this? " And I'm like, "Well, this MMP code base isn't in Cursor and the Meta ads API isn't in Cursor. So, sorry, dude.

Jacob Eiting:

But I think, I mean, we've hijacked the lightning round, but I think that might be a big ... This past year we saw agents come to code, and with some tooling I think more and more products are going to become native so that people can bring their own agent and they can coordinate across agents and either they'll have their embedded agent or you'll be running Claude, or whichever agent you want, on your local and it will be doing this stuff for you. So it's not a crazy wishlist item I think anymore. We're maybe not far.

Tanmay Jain:

Yeah. I would say my response is a version of what Jack said, which is we went through this recent experience and maybe there's just recency bias to our answers, but we had three or four different people that understood 40% of this vast growth stack that we needed to implement and we all understood a different 40% and we all described it slightly differently. And the amount of times it's just talks to be like, "Oh, well, this thing's not coming through in this platform. Why is it not? Okay, let's talk to the infra team." And then it's like, "Well, how do I translate this question they asked me to the infra team?" So I think with growth and the vast array of tools you have to use and the places you have to communicate, I wish there was just like a mega translation app.

David Barnard:

All right. Well, this has been so much fun. I know you guys are hiring, any specific jobs you wanted to shout out or anything else you wanted to share as we're wrapping up?

Jack McDermott:

We're growing as a mobile team. So I mean, a lot of people think of ElevenLabs for the API and what you build on web. We also have Native SDKs to be able to build voice agents and voice experiences into Native mobile and our team's growing. So we're hiring across iOS and Android, as well as Web, FullStack, as well as Growth Folks, so would love to connect with those who are on the pod.

David Barnard:

Awesome. Thanks so much. And we'll drop a link in the show notes for folks, but thank you so much for joining me on the podcast. This is a lot of fun, a really fun conversation.

Tanmay Jain:

Awesome. Thanks for having us.

Jack McDermott:

Thanks for having us. This was great.

David Barnard:

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 to join our private community.