You're listening to Startup for the Rest of Us. I'm Rob Walling. I'm your host this week and every week. And in this week's show, I talk with Jordan Gall, the founder of Rosie at heyrosie.com. Jordan and his team have been building this AI answering service for your business calls. And he has pivoted the company from what was an e-commerce headless checkout. And this is all after raising quite a bit of money.
in venture capital. And so you will hear Jordan in this interview mention that they had a few million dollars left in the bank. And he kind of flippantly says it. And I think that's funny because when you're in that game, he meant I only have a few million left in the bank.
But obviously having that much money in your bank account gives you some optionality. And that's what Jordan takes advantage of, as you'll hear in this interview. As you listen to it, what I don't want you to think is, oh, well, Jordan is making it work because he has a bunch of money in the bank. because money doesn't solve your problems what solves your problems is trial and error gut feel execution finding that gap in the market
being committed, doing a lot of things quickly and having most of them work out. And that's what Jordan Gall has done both with this company and his prior effort, Cart Hook. You'll hear that in the first three or four minutes, the story of growing that to millions. in ARR and then getting squashed by Shopify. It's a great conversation. We cover all manner of topics ranging from getting started to what it feels like to build an AI wrapper.
moving very quickly, and what it feels like to have product market fit. Jordan has signed up to be one of our founding coaches of the SaaS Institute. That's at sassinstitute.com and it's for B2B SaaS founders doing a million or more in ARR. He's a phenomenal founder and we are lucky to have him as one of our coaches. If you're interested in potentially...
being part of a mastermind, getting one-on-one coaching, having an amazing community, head to sassinstitute.com to check it out. It is a premium paid coaching program that we are rolling out through Tiny Seed. And lastly, before we get going, applications for MicroConf Mastermind Matching are open right now. And those are open for all levels of revenue, not just a million and up. Applications close on March 31st. Head to MicroConf Mastermind. Jordan Gall, welcome back to the show.
Rob, it's good to be back with you. It's been a few years, not the first time, not the second time, maybe the third, fourth, I don't know, but it's been a while and it's great to be back. Yeah, you've been on at least three, maybe four episodes, so really appreciate it. Today we're here to tell the story. of Rosie. And it has been a journey, sir. You were telling me offline, you are four years between product market fits.
With two different apps. So you want to give a summarized cart hook for folks and then kind of catch them up. There's Rally and then there is Rosie. And then we'll talk about, I want to find out where Rosie is today, where it stands, team size. You don't have to share revenue.
anything that gives people an idea of what's going on and then we'll dive into other stuff so all right cool so so the reason you and i like almost like giggle a little as we talk about the dramas because you followed along You invested in Cardhook, and so you've been getting investor updates for years for me. Probably 10. I guess 10 years. Not every month, as any of my investors know, but regular, consistent anyway. So, yeah, Cardhook had product market fit.
Cardick was a checkout for Shopify stores before that was a thing. We were one of the first. And it flew. So there was a point in time where it was adding 30, 40, 50K in MRR every month. No advertising, word of mouth, just flying. It went from a millionaire art to... two-ish and then six. So, you know, very quick growth, bootstrap-ish. We took a few bucks from friends and family, but it was effectively bootstrapped in the way it conducted itself. And then it ran into a wall.
named Shopify. So they did not want checkouts to take payment processing off of their platform and they ended Carthook's run. So I had the taste of product market fit. Felt good. Felt really good. And then I started a company called Rally, which was an e-commerce checkout for everyone outside of Shopify. And that was intended for what was...
called the headless ecosystem. That didn't quite work. We went toward enterprise. That kind of worked. We got to a few hundred K and ARR. It kind of wasn't going fast enough. And then we pivoted to Rosie. So I went through like this big dip over the last few years. We pivoted to Rosie, an AI voice agent that answers the phone for small businesses. And it might be a bit premature to call it product market fit, but it feels familiar.
Got it. So it's growing fast. And Rosie, when you say small businesses, do you mean like lawn care, maybe electricians, cleaners, where they're like out doing stuff and their cell phone's ringing all day while they're trying to do stuff? And so this Rosie acts as an AI agent.
I ask this knowing the answer because I've called your Rosie number and asked the questions. But I want the listeners to understand what's going on. Yes. There are a lot of businesses in the economy that rely on the phone. And it ranges from... I'm a painter and I'm on the job and I can't answer the phone to I'm that same painter and it's 10 a.m. on Saturday and the phone rings and I'm with my family having breakfast and I don't know if I should answer the phone, but it might be a $10,000 job.
All the way to, hey, it's after hours and we're a water remediation company. We got to pick up for emergencies and it's 11 o'clock at night. And up until now, the only option I've had is an answering service that I pay for a human. So it lives somewhere in between voicemail and answering services.
But any businesses, it tends to be local businesses, the type that you search for online and then you call on the phone to interact with. So that's who we're helping. It's like a very, very gnarly problem, which I can talk about. how painful the problem is, has led to challenging my assumptions around churn. And can you give us an idea of where the business stands today? Team size, growth, whatever, whatever you want to disclose.
Somewhere last year, I came to terms with, I guess it was toward the beginning of the year, 24. I came to terms with, hey, this rally thing, it might just not work. And in that case, what should we do? Should we just – we had a few million bucks in the bank. Should we continue to push on? Should we close the company? Should we pivot and keep the cap table the same? So I had one of those like corporate situations.
By coincidence, I was in London for a conference and I ran into a friend of ours, Justin McGill. And I'm sitting around the table with my co-founder and a few colleagues and we're talking to Justin about his experience in AI. And I have to give some credit to that conversation. It just planted a seed in the back of our minds because what we were doing is we were talking to someone who is experiencing product market fit and all the glory and challenges of it.
And it reminded us of what it's supposed to feel like. And I guess I went home thinking it's not supposed to be this hard to grow. And I think that was the moment where I looked around and I really thought, OK, I think it's time to make some changes. So we took the team from about 20 people down to six. We all held hands virtually and we said, do you want to keep doing this?
If you want to keep doing this, I'm going to come up with another product idea and we're going to keep the team small and we're going to build something in AI. And if you're in, cool. If you're not, totally understandable. So we shrunk it down to six people and I... did my search around for ideas and landed on this voice agent thing. And then we started to build maybe May or June. And then we had something functional to look at around July.
And then we started to bring on the first customers in August and then the first paying customers in September. And now here we are in January. I think it's really worth talking about that journey and the feature set and the decisions we made around that. Because in hindsight, it seems to have gone really well. And now we have been doubling every month for several months in a row.
And it's starting to get a little serious in terms of the revenue growth. Yeah, so, you know, we'll blow past a million ARR very soon. So that's, you know, that's a few months worth of growth. And yeah, I think it can fly. It's a funny thing. The I think this thing has wings is something I repeated several times somewhere around like November, December, because it just felt like, oh, this is how it's supposed to feel. It's not hard at all.
Everyone just wants it and they're pulling at you and demanding and asking and begging and want it to work. So like that feeling just gave us all this confidence and energy. I remember at a certain point in the life.
time of drip, I think it was 2014, where all of our graphs shifted, right? It's like churn went down, trial to paid went up, like all in the right direction. And I remember pointing and telling, I think it was Derek, I said, that's what product market fit looks like, like that's it. And you just said a sentence that I think... was similar. You kind of said, this is what product market fit feels like.
Right? I mean, it sounds like this pull, this market pull. I mean, it does beg the question, I ask this periodically on this podcast, because it seems like everyone has a different definition of product market failure. We kind of know, we'll know it when we see it, is what we hear often, but like...
How did you know you had product market fair? When did you know? Again, I always say it's not a binary, it's not a one or a zero, but when did you know you were hitting that 30 out of 100, 40 out of 100? It's getting stronger and stronger. What was that feeling like? It was qualitative at first and then quantitative after. So the qualitative was just the responses from people, the support chat, literally the vibe in the chat. Like you just hear people.
Oh my God, I can't believe this is as good as it is. How do I get this here? Like, can you jump on the phone right now to help me? Like this, you know, just very, very strong desire for it to work and to be part of their business. And then later it turned more quantitative where I was taking screenshots of like profit wall graphs and posting it to like rock and jazz being like, this is a joke. What is this? And then you start to, you almost start to learn more about your own product.
you go so at first when I did my projections I put churn I think I put it turn it 15 or 20 percent monthly because I'm thinking this is an AI product for non-technical people And it's not actually that good, which is one of the things that attracted me to it, because I think everything in AI will be very good. And so the things that aren't that good right now are actually the best things to build in because everyone just assumes, oh, is it a voice? Is it really going to replace a person?
Yes, actually it is. But if you start now, people laugh at you and then you kind of get in the right spot for a year from now. And the churn is like a fraction. And you know why? Because it's not replacing a human. It replaces voicemail. So it just has to beat voicemail and you get ROI. And so all these things start to reveal themselves. Like a few weeks into having paying customers, a few months later, you start to basically understand your own product.
Yeah, this is where you had a thesis that this was a need and it would replace a human and like you said, you get three weeks in and you're like, oh, it doesn't even need to do that. This is the learning. So this is what I see separating really successful founders from those that struggle.
is they can both have a thesis or a hypothesis, and they both build something and put it into the market. And the best founders, a bunch of noise comes in. There's a lot of noise. You're presenting it like it was super clear. I know that it wasn't, but you figured out. The good founders figure out.
Sift the noise away and there's a bit of gut feel there. There's a bit of conversation with team. You ask people you trust. And the founders that I see, I'm not trying to make a blanket statement, but founders who struggle.
They launch that and then they either don't listen at all to the feedback, and they're like, no, my initial thesis holds, or they can't sift through it, they don't ask for the right opinions, you know what I mean? And it seems like you really dug through it. I guess the fact that you're on your third one. Second product market fit, but the third time of doing this probably helped with that thought process, I'd imagine. Sure. I think what it helps with is confidence.
And confidence actually allows you to say, oh, I don't know what I'm talking about. I'm just wrong. Okay, so I learned this and move on. Much more so than I... I'm definitely right. And I'm making all the right decisions because of my confidence and my experience. It's more like when I'm on a podcast six months from today, it's going to sound like I'm smart and I did this on purpose. But in reality, I'm just kind of going with what people, what the market.
and customers are telling us, maybe that's a good transition to talk about the feature set, which was one of the more interesting parts of the experience. Let's do it. Okay. So I think of this as like an accordion, like out and then in. So what we did is we built a very bad version of all the features. We looked at what we thought people needed.
for an AI voice agent like a receptionist that did all this stuff, that transferred calls and set appointments and connected to your CRM, answered questions, gave directions, all these different things. And we put that out there. Then we put some energy and money behind cold email because we just wanted to talk to as many people as possible immediately. So we spent money for two months to just send out, you know, 500 emails a day. And that just gave us people to talk to.
Then we got them into the product somewhere in August. And immediately, like no self-serve, no sign up, no credit card, just like ugly. What happened was that we took the big feature set. and went to market with it, with our initial cohort. And then we immediately started to see what people actually cared about. Out of our, call it eight features, what do people care about?
it became very, very obvious that there were two or three that everyone really needed and the other ones were nice to have. And what that allowed us to do is to then shrink back down and we removed the features. We removed them from the admin entirely. They were in code, but you couldn't see them. And what that allowed us to do was basically identify what the features of the base plan should be. So a whole bunch of features.
Everyone cares about these. Remove all the other ones. Those are the nice-to-haves, the ones that people really, really need. That's our base plan. So when we started, we didn't have many features, so we only started with one plan. We removed the other ones. We just had a $49 a month plan. And what we came to realize was that what that is is an answering service. And what the higher tiers are is the spectrum from answering service toward a receptionist.
So answering service basically just takes a message. So answering questions and taking a message, that's our base plan. That's 49 bucks. And then as we started to grow with just the base plan. We then took those other features that were nice to haves and we added them back into the higher tiers. So it's like we had the feature set kind of right with some changes.
But we just removed all the ones that people weren't using right away. That was our base plan. And then we added the other ones back in, in a fuller sense, better with better understanding of how people actually wanted call transfers to work or emergency things to work or. sending a text message.
you hear people talk about MVPs. It's a minimum viable product, usually it's either limited functionality or it's just not very pretty, it's not very well built, but if it solves a pain point, people use it. Then you hear some folks today saying, you can't build an MVPs. MVP anymore because everyone's taste is too high. Times change, right? MVP started almost 20 years ago now, and so it can't be the same thesis they had then because everything's changed.
Consider that you launched an MVP, and was the accordion approach you talked about intentional? Or were you kind of like, well, we don't really know, so just launch a bunch of stuff and see what sticks? It was intentional-ish. The direction of it was, let's...
put out a bunch of features, we don't actually need them all, we can always remove them. So that was generally, but you know, the actual path it took and the details, not intentional, more reactive to what people told us. Now, we raise money. And what that allowed us to do was basically spend on things that you normally wouldn't spend in an MVP context. So when I called it ugly, it was functionally ugly. Visually, it was not ugly.
So we work with a designer called Francois from Clearly Design. Highly recommend. So all of our stuff was actually pretty. Now, the idea of an MVP, very, very different in an AI context. Another one of the things that I've learned is the differences between a traditional SaaS experience and mindset and approach compared to AI.
The reason MVPs in AI and SaaS are like very different things is because in AI, you don't build nearly as much as you do in SaaS. So if you want to build a CRM, you kind of can't launch an MVP. If you want an email, for example, you can't compete with an MVP. You're not really going to get anywhere. But AI, you're building at the very, very top layer.
the interaction layer, the app layer, whatever you want to call it. And then underneath, you're really leveraging much more infrastructure than you normally do. So yes, none of us... build hosting and servers. Great. So everyone leverages that. Or maybe you have a framework like Laravel and you're using Tailwind. So we do leverage these things, but AI, the wrapper context, you're really leaning a lot on other services.
So we don't do the transcribing, we don't do the voice, we don't do the LLM. In some ways, like what are we actually doing? It's just that the UI layer that allows a three-person painting company to have an AI phone receptionist up and running in 10 minutes. That's actually what we do. And in that context, an MVP, much easier to build.
That makes sense. That's how you move so fast. Because I was going to ask you, someone listening to this who says, came up with DD in June, had something in production in July, had paying customers or something. It's like, wow, that's really fast. I was going to ask if you moved fast because you have money, because you raised money, as you said.
But it sounds like it's just a simpler product to build. And fewer people. Going from 20 people to six people really changed things for us. And we demanded speed of ourselves. And one of the most challenging things was our... development process. So at Rally, being a checkout, your development process and QA, super, super strict because you can't mess up. You can't mess anything up in production or you cost your customers money and lose their trust.
So we had a specifically built, very tedious, careful development process and deployment process. And then when we went to Rosie, we had to say, we are going to destroy like 90% of our process. We have to intentionally set it on fire. It's a real challenge for our product leader, Jessica, but she's done extremely well with it because she wasn't ideological about the process. She was just like, well, that product needed X. This product needs Y.
Let's change. So a lot of the speed is attributable to that. And you've kind of mentioned or hinted that AI rappers... Rapping AI? I don't know. AI wrapper is like a bad term now. It's like, oh, if you build an AI wrapper, you have no defensible moat or whatever. Do you think of what you're building? Would you call it an AI wrapper? Is it something different? And how do you think about that moat? Could me and a team of people compete with you in two or three?
months it's shorthand and it's useful even if i don't actually think of us as a rapper right i think the conversation changed for the healthier with deep seek because it poked a hole in the potential value, right, the flip side of this, the commoditization of the LLMs. And so it put a lot more focus and attention on the application layer. And I think it...
evened out the debate a little bit. It's not one-sided. You can make money or get crushed all up and down the stack. So I don't necessarily think of us as an AI rapper, and there will be a lot of competition, for sure. But there's still a very honest dynamic between service and customer. If you provide value and you continue providing value, people will stick with you. So what if there's competition?
So it's a bit of a, I would almost call it a land grab right now. The more customers you get in a short amount of time, they are unlikely to switch if what you have is working, right? That's right. I forget who said it. It was a great term. I don't know if I can remember the term correctly, but it was like UI.
commitment or dedication or uh to where someone gets used to using your app and especially and not with non-technical folks like my parents are older and they don't man they learn exactly they learn exactly which button to click right so that's what you're referring to is like hey you're an or a lawn care person and stereotypically they don't want to learn a new UI. Right. I have some very funny conversations on Twitter with technical people who are like,
you know, this can be done so much cheaper. I'm like, you, yes. You keep having that conversation. Go for it. That is the same reason why if you go to our site, we have not taken on a vertical. Normally in SaaS, what would you and I recommend to ourselves and to other people? Don't try to be broad for everybody. Find a niche. The niches are big enough. Kill it for one type of customer, then expand.
In this, I think it's the exact opposite. So I see a lot of competitors doing really well, very niched, right? AI voice for veterinary clinics, AI voice for doctor's offices, AI voice for restaurants and so on. And I think it's just... Not the right time for that. I think one of the more interesting aspects of building an AI in general is that new markets are forming. And if you think about our market, the phone, you know, gigantic problem, enormous problem.
Answering the phone, dealing with it, not missing phone calls, just gigantic. It's very rare in our economy that you have that size of a problem and there are no solutions. Right, that is just rare and it only happens in this context because that solution was impossible. So before you had voicemail and you had answering services that people pay two bucks a minute for and are generally not that happy about. So now all of a sudden this third option comes around.
That's going to get filled in by competition. And I think that land grab slash demand rush. calls for being very wide and horizontal and everyone's welcome and this is easy to get started with. And so it's like, you know, against specialization in many ways. We'll see if in hindsight, that's a good call in about a year, but we'll see. I want to call out, as you said, it's rare that there are problems that are huge that have no solution.
People see money. Big companies are smart. Well, they're dumb and they're smart, but they do see big money. And they'll move into a space, and so everything's crowded, in quotes. But these technological shifts... move all of it and make it possible, right? The internet.
Did this? Like the World Wide Web, when it came about, it was suddenly like, oh no, there's this can do all that. And social media was another one. I think of even like Web 2.0, email, iPhone. Remember the iPhone coming out? Suddenly there was this huge rush. Remember Facebook ads?
Apps, that was a big rush. VR, that didn't pan out the way, you know, drones, AI, like crypto. And some of these panned out and some didn't, but they at least created this space that suddenly there was a lot that could happen that wasn't possible last year. Yes. And the key is that there's demand for it. Yeah. Because crypto feels like a huge innovation, not much demand outside of gambling. Right.
But AI is different. Yeah. So I do think about the email context a lot. If email comes out, are you building drip or are you building constant contact? Right? If it first comes out, you don't need to go niche and specialty. You just... We're MailChimp, man. It's everybody. And it's simple. I think Constant Contact might have been the first, or it's kind of the first one that's still around, and AWeber was the first one to do sequences of emails, is my memory. And we're talking like 99?
about early and they were so simple They're built in Perl and they're hosted on a server rack in downtown Manhattan in a cage. And you have to raise half a million dollars just to get the thing built and deployed. It didn't need to be drip. It didn't need to be complicated at all. There were no workflows because you didn't need them.
And if you could just communicate with an audience, you could build it. I want to circle back to something you said earlier, because I know someone listening to this is thinking about it. You built eight features, you said, and then you just like, boop.
Five gone up in the premium. Did anyone complain? Because I know that there's someone listening being like, ooh, that would be, everyone's going to be mad when I take their stuff away that they're paying for. In software, we have a lot more flexibility than we think we do.
Whether it's raising prices, taking features away, or just coming up with other solutions. So what we did, let's say for the call transfers, we just asked people, do you want to keep using it? We're about to take it away. Do you want to keep using it? And they said, yes, we just flagged it. Okay, fine. You can still see that link.
That's it. Move on. They just knew that we weren't going to support it and it wasn't going to be good. But they were like, I love it so much. I want it. And we said, cool, we'll come back to you with a better version of it in two months. Right now, it's going to go away for everyone else. It's going to stay for you. And so. In our admin, individual features literally have their own links. So we just remove the link.
Simple as that. That's actually how we reintroduce them also, is we would email people and say, do you want to try the new call transfers? And they say, yes, we just give them a URL. And then they would just go there directly from the URL.
I like that idea of, because we used to do that all the time. Feature flags were such a popular thing when we were trying to, I say popular, we used them all the time in Drip to figure out. For testing and for this type of stuff, there were features we built. RSS to email is, I hate that.
feature. And we built it because there were some early power users that were like, oh, if you just had this, I would switch from XYZ. And I'm like, I will never use it. I don't endorse that feature. It's a pain in the ass. It's brittle as hell, but I'm going to build it for you five because I knew that they would then.
talk about it. So we have a feature flag. And I bet probably tens of thousands of users of Drip at this point, and I bet there's like a hundred that have that enabled. And that's the thing, you're going for... tens of thousands of paying customers. I mean, that's your goal, right? Because your price points, tell people about your price points. Sure, we're at $49, $99, and $199. And we do have, we have some weird issues around minute usage.
Because our cogs are directly tied to minute usage. Usually I don't even like to think about cogs in the software context because, you know, normally card hooks like $5,000 for AWS for $500,000 in revenue. This is not that. It's more linear.
And I don't like it because people think of us as minutes. They say, do my minutes? Roll over. As soon as I heard that, I was like, we got to get away from this minutes thing. We want to focus on value, not the number of minutes. The pricing has been good to us so far. So in December. we launched our self-serve onboarding and that was the turning point. So up until then it was do we have it right? Do we have the feature set? Do we have the price? Do we have all this other stuff?
But it was ugly to help to get people onboarded. And we purposely did not build the onboarding because we assumed we don't know what the onboarding should be. So once we understood the base plan feature set. We looked at it and said out of these base plan features, what is absolutely necessary to get any value at all?
And we built that into the onboarding. What is the minimum number of things you need to do to just get value from this thing? And that's what our onboarding is. The other part of our onboarding, I learned this from Justin McGill, actually. was to take the value that's in the admin and drag it out into the marketing experience so that it starts there. Instead of saying, here's a wall that you need to get over to get value, AI...
has this opportunity. You see with ChatGPT, you see it with MidJourney, you just come into this product, you type something, boom, you have value. So we needed, I felt we needed to. provide that type of an experience. So if you go to our site, the onboarding is actually part of the signup process. So it's not if you want value, get over this hurdle of creating an account. Instead, it is come in.
Give us a bit of information about your business. So Google Business Profile is our primary path. Our secondary path is your website. Our third path is your just, you know, give us a business name. But the primary path gets taken 80% of the time. You put in your Google business profile. We pull your information. It's very structured data in Google. And then the next thing you see is your AI voice agent.
and you have a few clips, and you can just hit play, and it'll say, thanks for calling Pet Busters in Illinois. How can we help you today? So you get a little taste of the value up front, and then we ask you to create an account. And then when you get to the onboarding, we've already ingested your website, Google business profile, all your information, your business hours, all this stuff. And you're like, whoa, you know, I'm almost done. So that onboarding is kind of what changed the trajectory.
We just put the other features at the third step and just keep exploring. So you get through the onboarding, you get the value, you get a phone number, you can call your agent and you are blown away in less than three minutes. You are a painter somewhere in the suburbs of Illinois and you sign up on your phone while you're in your truck. And within three minutes, you're like, holy cow, I have this thing. I made this thing. It knows my business and can answer my phone. And that.
creates enough desire to get through the last part of onboarding, which is actually forwarding your phone calls to us. And as someone, you who built cart hook that were and rally that required the onboarding was not three minutes to see value no the onboarding was extensive and it was often getting developers and right right right so you've seen both sides of it i see both sides of it with tiny c companies you are
in an incredibly luxurious position right now. I mean, I'm blown away. Like three-minute onboarding? That's amazing if you can do it. Do you wake up every day and pinch yourself and think, this is it? This is great. So this is one of the few things that I can actually say was completely deliberate. And it was a reaction to pain. Cardhook was pain. Rally, in terms of onboarding, pain. And one of the requirements in this product was self-serve.
And I think that has served us well because voice and AI is really, really powerful. And when something's really powerful, you're tempted to bring all that power to the user. And that's what creates onboarding friction. So a lot of our competitors are like sign up for a demo or build this video.
workflow and choose your AI model and we were like absolutely not under five minutes to value and to onboarding so that was very deliberate and it does feel incredible part of the product market fit sense came after we launched self-serve and we watched completely non-technical people sign up, get their phone number, make a phone call, and then put their credit card in. So we put the credit card on the other side of testing your agent.
One of the more important experiences we had was around pricing. So maybe it's worth kind of like taking a little detour into that because that kind of surprised us in general. When we first started, we had a seven-day trial. with a credit card required. And for the first few weeks, we looked like geniuses because everyone was converting. Seven days goes by, your credit card's on file, you convert. And it felt good. And then we looked under the hood.
at the usage and i was like we are building our castle on quicksand not even sand like like wet quicksand and it's because people weren't using it so what we did is we switched from a seven day trial to a 25 minute trial And what that did is it went from time to usage-based. So now everyone that converts is an activated user that's already converted their behavior. Which is partly why our churn is so, so low. Because no one's converting that isn't using it.
That's the thing. Usually with a lot of SaaS credit card up front, the first 30 or 60 days, the churn is a lot higher and then it drops way down because people are using it as an extended trial. Yes, they want to try. Fine, I'll pay you the 50 bucks because I'm so interested in this. Let's see if it works. We flipped it on its head and that has made everything much, much healthier.
And so when we did that, we took the credit card requirement away from the onboarding and we put it on the other, the end of the onboarding. as an optional step. And so to see someone come in, self-serve, make a phone call, and then put the credit card in means they were impressed enough that they said, I don't want this to go away. I'm going to put my credit card in, even if it's not required.
And then 25 minutes goes by and then they convert and all of a sudden we have like a real user as opposed to, hey, seven days has gone by and now we have more revenue. Right. And for folks listening, 25 minutes is not a linear time. It's minutes of talk time, right? It's someone calling in and it's 25 minutes of interacting with Rosie. It can be one day or 30 days. I didn't care. I thought the numbers would kind of...
to even out, and so it would kind of reveal itself to effectively be a six-day trial or whatever else. Now... Here's the interesting thing. You said building a healthier business by doing this because you're making sure that the people who convert are only those who are actively using it.
Didn't do that. If you just did the credit card in seven days, you'd have more revenue right now. You'd grow faster, significantly. As a founder, how do you wrestle with that? Is it just long-term? You know long-term that's the right decision.
Or how do you internally think, boy, let's say you said you're doubling every month, you could be quadrupling every month, literally. And that is pretty tantalizing, right? That's actually what we see with some of the big Silicon Valley companies, right? The payment processor, not Bolt AI.
where they raise all this money and they're trying to da-da-da, and then they just crash and burn. How do you reconcile that as a founder who's ambitious, you want to grow as fast as possible, but you're like, well, I'm going to not put the pedal to the metal because I think it's an unhealthy business.
So maybe two things. I think the most important thing is we saw it as an experiment. We did not say we're changing the way the whole thing works and we're taking this big gamble to build a healthier business. We were like, let's just give it 30 days and see if that works out better. So that made the decision to go into it lighter. The second part of that is I have seen what Churn does.
And I have Googled maximum churn given growth rate. And you know, it's not good. Yeah, it's not good. If anyone doesn't know what I'm talking about, there is a formula for your maximum MRR given your growth. and your churn, and you do not care about it at 50K MRR. At 300K MRR, you really care that your maximum is 350, because here we are, the wall has arrived. At Cardhook, we found the wall.
because we were churning at 15% a month. So we were skyrocketing, but churning like crazy. I called it a washing machine. And if you recall, what I did is I went to the support team and I said, what percentage... of your interactions are with people that will not become paying customers. And they were like, at least 50%, at least.
And I was like, that's an absolute disaster, not only for the support team, but also because we're wasting time on people that don't care as much and we're not giving time to the best people. So that really gnarly decision. at Cardhook to force a demo and only take on people that we thought were right and raise prices at the same time, that taught me you actually are better off building a better business as opposed to a faster business. So I was more open to it than we saw it as an experiment.
And it worked. And then there's one final step in the process that just came out, I think earlier this month in February, that has again changed the trajectory again. And that was what we call premium UI. So up until the beginning of this month, you looked at the feature set on the marketing site and on the pricing page, and then you signed up and you just didn't see the premium features.
You would basically have to ask us, hey, that call transferring thing or this custom training, can I hear more about that? And then we would say, yes, that requires this higher tier. So we had no mechanism for people to discover the higher tiers. And instead of just jamming it in there, we very deliberately went with the designer and a product team and said, how do we make this not only incredibly attractive, but also self-serve?
And so now you go through the onboarding and that third step, the keep exploring, is now expanded. And as you navigate the admin, there are individual elements with a little yellow star. And you click on those and we reveal the feature and we clearly state this is a premium feature that requires these plans. Do you want to use it? Here's what will happen. And building that self-serve.
Now we have a mechanism for increasing ARPU. So now everyone comes in at the base tier, at the $49 tier. And now the percentage, we're almost at 50% of our revenue is now at the higher tiers.
And then that starts to argue for expanding the strategy around that lower tier. Do we have a free tier? Do we lower it to 29? Now that we have more confidence around a self-serve revenue expansion, now it starts to factor into our... you know, marketing plans and our advertising and our positioning and pricing and all that. I'm looking forward to following this story.
AI in that world, things are going to change quick for you again. Yes. I just sent a tweet out yesterday that encapsulates this. Extreme optimism and paranoia. I think that that is the appropriate stance for the CEO of this company is we are going to push so hard. We're going to crush it. We're going to get to 10 million ARR in 18 months and at the same time at any minute.
Someone could release something. Someone could do something like, so we have to hurry. And I think AI kind of, that's what founders in this space right now are feeling. No one should be overconfident about anything unless you're lovable and you got the 17 million ARR in three months because then you're fine. Be comfortable.
There you are, yeah. So if folks want to read that tweet, they can head to Jordan Gall, that's G-A-L on Twitter. And of course, if they want to check out Rosie, what we've been talking about for the past 30 minutes, the AI answering service for your business calls, they can head to heyrosie.com. Jordan, thanks so much for joining me. Thank you very much.
When I'm saying on this show, you know, the founders I know who succeed do X, Y, and Z, he's one of those founders that I think about. Like he is a, if you know coding, he's a design pattern for doing. smart things, executing well, getting done, moving fast, a little bit of gut, a little bit of data and being super scrappy and making it work. So it's always great to have Jordan on the show and I hope you enjoyed the conversation. This is Rob Walling signing off from episode 764.