Steve Hoyek (Journable): Building an AI calorie tracker app to over 600,000 users

Speaker A: So what we are trying to do with our app is to help users be consistent with their journey, with their habits, whether it be eating, whether it be tracking their exercise. And gamification is a way of bringing users into the app. It's keeping them coming back, it's giving them a bit of dopamine to help them with consistency. For one, because we are able to help users, this is, you know, why we're doing this. We've built this up in the first place. And two, we see the value of, uh, product led growth. Users will become users who use the app successfully and hit their goals. They become walking billboards for our product. Right. How did someone lose weight? They tracked. How did they track? They use this app, which they fucking love. Again, sorry for my uh, loose tongue.

Speaker B: We like a little bit of spice. A little bit of spice in the podcast. All right. Hey Steve, welcome. Let's get started. You want to start by giving an overview about journable, like what is the app and how, like when did you start it? Give us the uh, the big picture.

Speaker A: Thanks for having me. So Journey is an AI calorie tracker. We've been in the market for at least 18 months and yeah, we've seen some good growth. It's been a journey. That's definitely something I can say for sure. Journey. We've currently just passed around 600,000 users in total, mostly on Android. And uh, we're starting to get some good numbers when it comes to the business side of things as well. Two of us, two man team based out in Beirut, Lebanon. We've got a nice little office and yeah, we're just uh, working to grow this thing into what we need it to be.

Speaker B: Nice. Did it start out monetized like right out of the gate or was it like free for a while and then you figured out how to convert and monetize after that?

Speaker A: Monetized, straight off the bat, we were bootstrapped. So we needed means of sustenance, we needed means of uh, refueling the coffers of the business. So monetized from day one, hard paywall. And uh, we got some early traction. We got some early users after a product hunt launch and we had a couple conversions who went for the annual subscription. So that was all the validation we needed to really double down and full focus on ongoing this business.

Speaker B: So walk me through, like, who is your ideal user and what is their path towards using the app? Like how do they find out about you and then once they install it, like what's that first initial flow look like? And then at what point does the paywall come in?

Speaker A: So I'll answer this with a bit of the backstory about how drainable came to be. So myself and, uh, my co founder, we've both used calorie tracking apps for significant periods of our lives. It took me five years to lose calories, weight on my journey. And in this process, I was using my fitness pal and some competitors rigorously. You know, I would use them every single day. And so we have a really good, let's say, um, feel for the space. We have a feel for the kind of user that we want to attract. A problem that we both faced was the complexity and time that's required to track your calories and to manage your intake. You know, it's difficult enough to eat to be in a calorie deficit on a daily basis, let alone needing to spend 10, 15, 20 minutes per day inputting these things, especially when you're eating, when you're at lunch sitting with other people. So we wanted to solve this. We wanted a very simple and easy way to track your calories at minimum effort, minimum input, with all of the results, or at least let's say 90% of the results. So that's how Chernobyl came to be. When LLMs came around, we knew that this was it. We knew this was the opportunity. And so we went ahead and built the Chernobyl as the simplest and easiest and quickest way to track your intake. So that's our ideal customer profile. It is the user, it's the person who's on a nutrition journey, be it weight loss or weight gain, who wants to track, but who wants to track with a minimum amount of effort, with around 90% of the accuracy and 90, uh, percent of the results. So that's how. That's who we're targeting and that's who our app is resonating with the most. The way we acquire them is largely through ads, and we're recently seeing a huge surge in organic growth through word of mouth. But the paid ads is our primary acquisition channel of how we target these exact customers.

Speaker B: And so what's the flow like when they first install the app, what's the first thing they do? Do you ask them to take a picture of their next meal, or is there something else? Do they put in their goals first? Like, how does that look?

Speaker A: We've elected to keep the onboarding as frictionless as possible. So you download the app, you open it immediately, you're presented with the main functionality of the app, the core, the chat screen, similar to your, you know, your, uh, Gemini and your chatgpt apps. There's no onboarding flow, there's no tutorial, there's nothing. There's just a chat window, and you figure it out. We have the same approach, and we've seen success with this frictionless approach where users generally in the space, are used to our competitor apps, which have rigorous and lengthy onboarding screens, which help build confidence, and then they're hit with a hard paywall. We elected to take another approach. We have a free tier, which is around 20 entries. So 20 food or exercise entries, either by typing into the app or sending a picture, and then we have a hard payroll straight after that. So users get to use the app, they get to experience it. They don't need to answer 110 different questions about their lifestyle and their goals before, uh, the wow impact or the wow effect of using the app and noticing how quick and easy it is to use. So that's how it goes.

Speaker B: So that's interesting. The chat UI is at the center of it. So is it like when I want to take a picture of a meal or something, am I actually sending that in the chat to the LLM?

Speaker A: Basically correct. The chat. Our main window is a chat window. It's very simple, it's very minimalistic. It looks almost identical to your, let's say, Gemini window or your chatgpt chat. So the chat box at the bottom, uh, camera image, upload, and, uh, be very simple, very straightforward. And that's what our users love about the product, which is the simplicity, the straightforwardness, and the speed, really. You know, you just open the chat, send a message, continue with your life, as opposed to clicking and tapping through 110 different screens, searching through databases. Recently we were discussing with our, uh, Google Play representative who showed us some very interesting data between our app and our peers. And it showed that our app had, let's say, a third of the screen time of our peers. So users would engage with the app for a third of the time. Now, generally in a vacuum, this is a negative metric, right? It's, It's a poor, it's poor performance. But he obviously had the foresight to tell us that in our case, this is an awesome metric. It means users are spending a third of the time to achieve their goals than our competitors. So although generally you would want more people, more hours, more minutes of screen time on your app, in our case we want less. So we have lots of engagement, lots of users opening the app multiple times per day, but, uh, for way shorter time periods. So that's our USP and that's what we're leaning into.

Speaker B: Yeah, I mean for you guys the golden metric has got to be retention, right? Because it's more about like, are people using this every day for weeks, months, years? Not are people like constantly engaged? Like if you were a social media app or something where you just needed people's eyeballs all the time.

Speaker A: Bingo.

Speaker B: So what about all the other parts of the app? Like I've seen that you have like a kind of calendar ui, you uh, have a kind of a goals section. So how does that tie in with chat? Like are those parts populated by your chat experience? Like the LLM just goes and like fills all the data in or do you have to manually go and put in your goals in like somewhere in the, somewhere else in the app?

Speaker A: Yeah, good, good point. So where we can use the LLM, where we can use the chat function to populate the other areas of the app, we do, but there are many cases where the non deterministic nature of the LLM require us to have some custom functionality, let's say. So setting one's goals. We have a questionnaire after you send your second message, uh, or your second entry to the app, we have a pop up that tells them to update their goals so we can understand a bit more about them. They click through the questionnaire, they set their current weight, their target weight, so on and so forth. But we are definitely leaning into the AI aspect of things. We're working on a couple features in the pipeline now where we're going to be providing AI generated insights about their diet, their past day, their past week, their overall nutrition, giving them a health score from 1 to 10 on each food and their day, and giving them some insight, helping them learn. So this is all leaning into the AI aspect of things whilst also providing value by means of education to the users. Generally we have two main value, uh, let's say we can create value for our users in two main ways. One is the simplicity and ease of them tracking and doing what they know. And the second is education, teaching them, using AI, using our LLM to educate users about how to make better health choices, better food choices. So these are definitely things we're leaning into.

Speaker B: Yeah. What do you think? I'm curious. This is a little bit of a tangent, but I'm, I'm very interested in this idea of now that we have LLMs and people are getting used to kind of chatting with them as a mode of like interacting with software if like going forward, these sort of Onboarding flows and like product documentation where you have to learn what parts of a product do I need and all that stuff. If that just goes out the window and you just basically use a chat interface to say like this is what I like to do, tell me like what basically adapt the software to my needs. On the back of that, like you just explained how when someone starts sending messages, eventually you do show them that same type of questionnaire that you might see in onboarding in your competitors because it's information that you need like what's someone's weight, what's their target weight, what's their goals? Like you do need to collect that at some point but in your guys case you do it in a much more interesting way which is like it gets presented to the user in the natural course of them interacting with the app and like first sending some, like getting some value first before you, you show them that. And I wonder if that, if you think that'll be like a general principle of software like design in the future or like you just like the parts of the features that you need are revealed to you over the course of the experience of like talking with the LM and using the app.

Speaker A: That's an interesting one. You know, I think zooming out a bit without looking specifically at LLMs, but generally at software patterns and UI UX patterns, I think we have been moving in that direction where features have been surfaced to users in a timely manner to ensure the users one are not overwhelmed with features and they don't need to go through some lengthy onboarding process. So this I think has been a uh, general trend. I think with LLMs it's even more magnified. Although when we were initially discussing how to get users, how to onboard users, you know, soft onboarding as opposed to a hard onboarding, the discussion point of using the LLM and using the chat to kind of guide them through the app, we felt it was, we were a bit fitting the technology to the problem as opposed to finding the optimal solution to the problem itself. So we uh, reverted back to your traditional onboarding screen. Just we propped it up at a, you know, opportune time, you know, to make sure the user has had an opportunity to understand the app, the functionality. And then we start asking some follow up questions to help customize, personalize the app to the user and you know, give them confidence and kind of build some stickiness. Right. You know, once the user has some data in the app, they've logged for a couple of days, they have their goals, they can track their weight over time. Know, it builds a bit of, you know, it, it increases retention. Right. So all that said, it's an interesting one. We are definitely leaning into using the LLMs more to customize and personalize the user experience. You know, like once they set their goals, their insight will be customized to their, uh, specific situation, whether they're losing weight, if they're on a keto diet, if they're vegetarian. So all of the recommendations, all of the insights about their micronutrients, their macronutrients will be personalized to their current situation. Um, it's just something we're kind of tiptoeing around at the moment because we don't want to cannibalize our core user experience. We don't want to cannibalize the simplicity so we can have a shiny new piece of innovation in our mobile app. Right? Yeah.

Speaker B: And people, I mean, there's something to be said for like, there are UI UX patterns that have been around for a long time and people are used to them and they just kind of expect them like having some sort of just form where they can put in the information as opposed to like having to go back and forth with the, with an LLM to like do the same thing. So, yeah, that is the flip side of that question.

Speaker C: Um, just before we move on, at, uh, what point do people log in or do you keep the free users kind of anonymous? Is that, uh, the paywall that people actually hand over their details?

Speaker A: Good question. So something we've been thinking about a lot recently. We currently have no need to log in where people can stay anonymous in the app. They're anonymous, they can onboard input all their data, still remain anonymous. They can hit the paywall, pay purchase, use the app, uh, for a year and still remain anonymous. At the moment, we have zero requirement for users to log in. Now we are changing this. We're still not going to enforce it, but we are going to nudge the users to log in because we've noticed quite a few users have lost their data due to not logging in, switching mobile devices. They retain their subscription, but not their data. And yeah, we just don't want to. We want to prevent users from losing their data. So we will nudge them after a couple entries to log in to provide us with their email. But, um, you know, at the moment we don't use their emails for anything. We understand that there is a marketing benefit to having their email. So this is something we might change in the future. But again, frictionless onboarding experience. Give them value first and then um, we'll solve the other problems.

Speaker C: Sure.

Speaker B: I wonder is there a way to store their information against their like subscription ID or whatever if they have a paid subscription so that at least if someone is paying and you have that identifier, then when they switch devices like you could still fill in that their information and like have it there?

Speaker A: Absolutely there is. And we've built an internal tool to help restore users data based on their subscription IDs. Um, productionizing it is a bit more complex than you would think. So we are discussing right now whether we should enforce emails. I uh, want people purchase subscription or whether just nudging them to log in and create an account is the approach mostly to prevent data loss. Just something we're considering now. Easy problem to solve. We just need to put uh, it into action. Yeah, yeah.

Speaker B: Okay. I'd like to talk about the user acquisition side of things first and I'll spend a little time there and then go into like retention and like long term how will you keep everyone around and stuff. But to start with these acquisitions. So you mentioned that it's mostly paid ads and I have a couple of questions there. I guess we'll start with was it always paid ads from the very beginning and like chicken and egg, you guys bootstrapped and probably didn't have a ton of initial investment because you just bootstrapped. So how did you like first get the cash to start investing in ads? Did you just have to like invest your own money for a while and wait for it to uh, become profitable or how did that path look?

Speaker A: Yeah, sure thing. So we launched, we soft launched our app, we had it on the stores, we started getting like a couple downloads per day or per week. Nothing, no large numbers at all. Then we hard launched, let's say by launching on Product Hunt, by launching on a couple different fora, you uh, know, online platforms. We basically put the word out there and see uh, to see if we get any hits. We got some success from Product Hunt. We got um, I would say couple hundred users and two of them subscribed. So we had a 2% conversion rate off the bat. Very, very bare bones MVP version of the app and we had two subscribers. So this was this effectively de risked investing some of our own money into the app. It de risked. It validated that this is a problem that people have and our solution is something that people are willing to pay for. So we kind of took that approach where we would um, run a few tests. We would validate de risk. The first, first we de risk the product and we de risk the team, we de risk the company. We, we kept going through this process of de risking different things and at every stage of validation we would invest some more of our own money. So we, yeah, to start off, once we got the first couple users we knew, okay, let's find what the ratio is. Let's start learning about acquisition costs, cost per install, cost per subscription, cost per impression. And so we ran ads on different platforms. We learned a bit about meta ads, we learned a bit about Google Ads, Apple ads, all of the above. You know, we went on Reddit, we went on LinkedIn, we went on um, TikTok, Snapchat. We tried every platform to kind of get a feel of our average acquisition costs. And we started fine tuning, optimizing and eventually found a formula whereby we are profitable by running certain ad campaigns on day 30. So 30 days after running a campaign, the campaign has paid itself back. This has been our North Star since we started running ads, which is finding profitable campaigns on day 30. So obviously we're a subscription app, but we completely discount all revenue beyond 30 days from users. We discount monthly renewals, we discount annual renewals. We assume that users, revenue, users LTV is their revenue on day 30 because we're bootstrapped and we need to refuel the engine right. So we can't think like some long term SaaS apps. We can't base our LTV on multiple years renewals or multiple months. It's all day 30 revenue. And that's been how um, we've operated up until now and how we've been able to sustain and refuel our engine, uh, of growth.

Speaker B: So you said you experimented with a lot of different ads platforms. Did you do that all at once before you had profitability or did you start with one or two? Because I imagine like you, each ad platform needs some sort of minimum spend before it's got enough like data for you to actually see what's going on and validate the results. So I imagine if you tried to do all those at once, it was either it was probably wicked expensive or you didn't have enough data. Like, how did you guys do it?

Speaker A: Jason, we made many mistakes along the way. This was one of them. We launched, um, many campaigns on many platforms, all with insufficient budgets, way below the threshold of actual performance or of learning. And then we started to learn and optimize. But uh, yeah, if, if, um, anyone watching this is in the process of starting to get some traction and launch, they should focus one platform at a time and give Each campaign, sufficient budget to learn to scale and to actually, uh, get a feel of the performance.

Speaker B: So when you, when you were doing that and you realized, oh shit, like this is not actually enough to have data, did you just like stop spending on a bunch of platforms and focus on one or two? And if so, like, how did you pick the one or two to stay on? And how did you like, feel confident that that was the one? Because the budget was low.

Speaker A: Right. You know, it's, it was a long process of discovery, long process of optimization, trying different things on different platforms, understanding these ad platforms. You know, my co founder and I were fairly new to the distribution and marketing game, so we had to uh, deeply understand how the platforms work, the, let's say the process. So for example, on Meta Ads, creative is the name of the game, right? On Google Ads or Apple Ads. When you're running ads on the Play Store App Store, it's a bit less on the Creative, it's more about the App Store page or the App Store listing or Play Store listing. So we had to deeply understand these platforms, understand our capabilities, our strengths and experiments. And so we just kept experimenting. We eventually landed on the major players, you know, Google Ads, your Meta Ads, Apple Ads, so on the stores themselves because we knew we had a, uh, good product, our product was selling itself. So we knew that running the store's native advertising placements would be successful for us. And we landed on Meta because Meta Ads is just, you know, the biggest, let's say acquisition platform which is not native to the store.

Speaker B: So which one is the biggest? Is it the Play Store ads and then Meta, or is Meta number one as far as your spend and everything?

Speaker A: Definitely. Google Ads, Play Store Ads is, um, our biggest acquisition platform. We are growing on Meta. We've tried many different things, but as of now, our Play Store, our Google Ads campaigns are the most successful we have. Most of our ad spend is going there and most of our, let's say profitability is coming from those ads.

Speaker B: And so I assume you tried Apple as well. Did it work? Not quite so well there because the cost per install is just that much higher because Apple users in general are a bit more valuable and like more expensive to acquire. Is that just the economics didn't work because of that or what was the reasoning there?

Speaker A: You know, it's an interesting one. Obviously the conventional wisdom is that it's better to launch and to grow on iOS because the users are worth more, you know, higher average value per user. But we found a similar conversion rate on Android and iOS for our app and four times the acquisition cost for iOS. So it just didn't make sense. You know, we have four times the performance on Android than iOS now this could be a function of our creative not appealing to iOS users. It could be a function of many things. But we tried and we tested and we burnt lots of money on running Both Android and iOS campaigns on Meta and iOS was just four times more expensive for the same LTV, the same conversion rate, same revenue per user. If let's say we were a mobile game or we were an app with multiple in app purchases and the LTV per user would be different within the same market, then that's a different story. But frankly, you know, having just one fixed price for a monthly or annual subscription meant that we were not reaping any benefits of the um, dichotomy of LTV per user that other apps see between iOS and Android users. So yeah, we stuck with Android. We grew a lot there. We, we are starting to make Some moves on iOS again, but one step at a time.

Speaker B: Yeah, yeah. Well, I mean especially if you, since you guys are bootstrapped and you're optimized for like what is the fastest and biggest ROI we can get, like in 30 days, what's the, what's the highest ROI we can get? Then it seems like the Play Store is just way better for that because you, the conversion rate is the same and it costs you way less to get each, each subscription. So it makes total sense. Um, Charlie, you're going to say something.

Speaker C: Yeah. On the, you mentioned that the, as the Play Store is kind of your biggest, biggest angle right now, what would you say to people is like the most important thing for getting those ads to work then because you mentioned like the Play Store page, is there any specific things that you know, you've kind of tried that have definitely worked and getting that ROI higher?

Speaker A: Definitely. Um, Play Store assets very important because the default assets that your Google Ads campaign, specifically your Google Ads mobile app campaigns use are your Play Store assets. So that's your icon, that's your screenshots, specifically your first two or three screenshots. Optimize those A, B test your app icon. You can find huge results by just changing the color of the icon, for example, the background color. So you know, we, no one knows how these things work, but it is your icon shows up in the most places, the most placements by far. Abtest the order of your screenshots, the background color of your screenshots, you will find significant difference in conversion rates by Doing that, that's the thing I would say otherwise, just listen to the campaign, listen to the ad platform. You know, if you have some copy in your campaign, some headlines and descriptions which are performing well, use ChatGPT to find five alternatives which, uh, which sound similar. If you have something which is performing poorly, replace them. You know, just continue to optimize, listen to the recommendations of the platform, test different creatives. You know, don't be afraid to copy, don't be afraid to, um, to see what works for your competitors. You know, open the Play Store, uh, look at the sponsored ads, look at the different placements, see what they use, see what they do. And uh, you know, they, especially the big competitors, they know what they're doing. They've been doing this, they've been a B testing for many years. So just use the benefit, uh, benefit from all of the lessons they've learned over the years and just um, create similar assets, you know, like large images with large text with a back, a very simple background seems to be a winner. Everyone does it, so you should do it too.

Speaker C: Yeah. Because I know Meta, uh, on Meta, you can basically just see everybody's ads in the library.

Speaker A: But awesome. Best resource.

Speaker C: Right? But is that the same for Google Play?

Speaker A: Amazing resource. Yeah. Um, Google Play, I haven't actually looked into it, but you have more than enough, let's say inspiration on the Play Store itself. So just go open the Play Store, go to your category. In our case it's health and fitness and you'll start to see sponsored ads, different placements or obviously the icons, they show up everywhere. But then you can also see some other placements, sponsored placements, wide images, square images, copy. Yep.

Speaker C: So I think one thing that people get kind of get a bit stuck in or rut in is with the testing. They know they're going to need to test different versions, but maybe they don't have the budget to really give these things enough leg room to work or they kind of switch around too quickly. Would you have some kind of guide or a gauge of um, how long you would need to leave these things to really see if they're working?

Speaker A: Yeah, I mean, patience is a valuable virtue when it comes to user acquisition campaigns and testing. I would say at least a couple weeks, ideally a month and more. That's in terms of length, in terms of budget. Listen to Google's recommendations. If you're running a acquisition campaign. So you need the uh, 10 conversions per day, whatever your conversion action is, if your conversion action is installs, 10 installs per day. Generally cheap, right? Anyone has the budget to run an install campaign to get 10 per day. If you're optimizing for in app purchases, ensure you've got the budget to get 10 purchases per day. If a purchase is costing you five bucks, you need at least uh. If a purchase is costing you five bucks, then you're in a really good place. If it's costing you 50, then you need at least $500 a day of budget per campaign in order to find good results. Because the algorithm definitely needs 10 conversions per day to optimize to learn. Frankly, talking to someone from Google, they even suggested 15. Even though the official documentation says 10. They said, you know, he's conservative. We're also conservative in the business. It's good to be to err on the side of more data than less data. So design and structure campaigns and your campaign objectives such that your budget will meet the minimum required threshold to ensure the algorithm can optimize and can learn. If Your budget is $300, find an action Create an event, uh, in Google Ads in your stack which allows you to make 10 of these events per day instead of five purchases. For example, if you need to go halfway through the funnel, maybe a trial start campaign or an onboarding completed campaign, do that. Once you've got the budget to run a more, let's say higher or otherwise lower in the final campaign, then do it. If you have the budget to run a subscription campaign, go for that. But make sure you've got enough conversion events per day to help the algorithm learn.

Speaker C: Yeah, I think that makes a lot of sense.

Speaker B: So since you're at some scale now, does that mean that your Google campaigns are all the conversion action is actually the subscription or are you guys still using the install as the like? Where are you at on that conversion action?

Speaker A: We are as a. Our Google account manager would say we are sophisticated in our app advancement. So that's effectively the um, how the degree of advancement of a mobile app on Google Ads. We run all of our campaigns as target return on ad spend campaigns. So again this is because of our business does not look the same as everyone else's business. We are bootstrapped. We really need profitability and frankly high profitability in order to fuel the engine of growth. So we run target return on ad spend campaigns with a very generous return on ad spend. So for each, every campaign needs to be very profitable for us to run. Has taken us a long time to reach this point. We experimented with every type of campaign and as we grew we started let's uh, say graduating our lower level, lower level campaigns. So we went from install campaigns to trial start campaigns, onboarding completed campaigns down to purchase campaigns and eventually we reached return on ad spend because all of our data is going straight to Google Ads from Firebase.

Speaker B: So it sounds like the path then is actually clear and actually a bit more optimistic than I would have thought for, for someone starting a new app because it sounds like you can do install and if your installs are your conversion actions, then you actually don't need a huge budget, right? You only need like 10, 15 installs a day. Call it 15. I think installs on Google Play can be what like 50 cents or a dollar or something like that. Like they're not, they're not too expensive. So it doesn't cost much money at all to run that. And then once you have the install is coming in, you can start potentially talking to those users, like figure out who they are, see if you can like get some of them to convert to paid and then gradually you can increase your spend and so on until you can graduate from the install to the trial start or the onboarding completion or whatever. And just like you did like move down the funnel, um, and grow the Google Ads things. That's a somewhat optimistic message, I'd say.

Speaker A: No, Bingo, you nailed it. And anyone can do it. And this progression, this chronology which you've just outlined is exactly verbatim written on the Google Ads, let's say learning path. If you go, you want to learn about Google Ads, you want to learn about mobile app campaigns, they very clearly show this progression of you start with this type of campaign, you get these lessons, you graduate to this type, once you've got the budget, you get those lessons and eventually you reach the final, let's say the final boss of Google Ads app campaigns. Where, which uh, is where we are now. It's all online, it's all there. Highly recommended. Go to the Google Resources skill shop, I think it's called, where they uh, host their learning content and take uh, the Google Ads course.

Speaker B: M and does the return on ad spend campaign that you run or campaigns, does that, is that a 30 day cutoff like you, you mentioned, or is that like lifetime? How does that the math work?

Speaker A: You have the option of setting your conversion window. We have it set at 30 due to the nature of our business, I believe it can go up to 90 days. I believe Google recommends 30 as and frankly it depends on your type of app. But for where we are, 30 made sense. Um, we don't want to be Too reliant on recurring revenue. We want to be reliant on day seven and day 30 revenue. We need companies, uh, coffers to constantly be replenished. So that's how we settled on things.

Speaker B: So let's talk about the uh, difficulty of bootstrap, uh, businesses specifically around the sometimes fragility of, of user acquisition. Are you concerned that as the tides of Google Ads and the kind of auction based nature of things changes that you might get out competed by some VC backed calorie tracker that just blows money with a, you know, two year horizon on roas and just like crushes the, you know, drives the price up super high or does that not stress you out too much?

Speaker A: It's the number one thing on our minds on a daily basis. We are very aware that we're on, what's the expression here, Borrowed time, borrowed ground effectively. Um, we know that at any point BC backed competitor can flood the market and will drive us out. So we are working very heavily. If my co founder were here, he would be nodding his head right now because he's been pushing this agenda for a while. We've been working very strongly to diversify our acquisition channels. We cannot be reliant fully on paid ads. We cannot be reliant on a single platform. So we've been working to grow our organic acquisition by virtue of referrals, word of mouth, putting content out there. We're also discussing hosting a podcast similar to yourselves, um, around health and fitness. So we are looking at uh, uh, you know, affiliate marketing, influencer marketing. We're really trying to diversify our acquisition channels and trying to find another profitable channel because as of yet our paid ads is our only, let's say working and functional profitable channel.

Speaker B: Yeah, let's talk about that a bit. So have you, have you started to lean any particular direction as far as the new direction to go or like what new channel you might want to add next?

Speaker A: We have launched our affiliate program whereby people can earn commission for referring users to our app, uh, users who subscribe eventually. And we're very generous with our commission structure. We give 50% of all recurring revenue to our affiliate partners. And we've already had some people um, you know, earning some significant um, uh, you know, returns by sharing our content, by sharing our app on their social media channels. We even had one creator who just put one or two stories on Instagram and made 500 bucks within within the first month. And a lot of these are on a recurring basis. So affiliate marketing is definitely something we're leaning into. Currently launching an outreach campaign to uh, onboard more affiliates. We've also worked to diversify different acquisition channels, diversify away from Google Ads and launch profitable campaigns on other platforms like Meta, like Apple. And then we're also exploring some other smaller alternatives. But mostly um, affiliates is where we're putting the most focus at the moment.

Speaker B: Yeah, I mean affiliates is interesting because it's more like a tool that can feed into many channels. Right. Because like if you have the affiliate model, then someone can go and post on TikTok. And now you're doing influencer marketing just via your affiliate program. Right. And in the same way like you know, someone who owns a newsletter can join the affiliate program and like plug your app and their newsletter and now you're in the newsletter marketing game. So it's like uh, it's almost like a little leverage to get into multiple different channels. But Charlie, I'm curious what you think about this. I feel like this type of app lends itself really well to your kind of standard influencer marketing, like TikTok stuff because like there's certain things that do really well with that. Things that are about like self improvement or like looking better or like being healthier, like these sorts of things. I think you can have a lot of like viral, viral hooks, I think on the back of that.

Speaker C: Yeah, I was just going to say like as we've learned, like it can be really awesome, you can go viral. But then essentially repeating that on a predictable basis is really ah, the challenge. But yeah, curious what your thoughts are.

Speaker A: We classify, I think it comes down to semantics because we classify influencer marketing as paid acquisition. Obviously it's uh, less, it's more sporadic, it's less um, predictable and uh, you know, you could go viral and it could flop. But because we're paying directly for the content to be created, we classify it, you know, in the same bucket as um, as let's say, uh, paid ua like meta ads or Google Ads, whereby affiliates, where we have very little upfront investment and again being bootstrapped. We're very careful with every dollar we spend. And we are more in the business of deterministic, let's say risk taking as opposed to trying a couple influencer campaigns, throwing money at it and hoping that one of them sticks and makes us a return. Frankly if they don't make a return, then our business is in a poorer state than it was before we spent this budget. So we're very careful with every dollar spent. Although we do see, like you mentioned, you know, our category, our kind of App, uh, is something which influencer marketing works very well for and we've seen many competitors take this route. We've seen many competitors find great success with this route. The non deterministic nature of it and the nature of our bootstrap business means that we've erred away, we've shied away from this path. Frankly. This is something which uh, my co founder has been pushing and I've not uh, been on the same page about because it's, you know, if obviously we're in the business of taking risks, but if this risk doesn't pay off and the business is in a way poorer state and we have less firepower to fuel our growth engine. Right. It's a comeback to that point with our payday campaign. So at some point we need to pull the trigger and just go for it. Uh, maybe that point is soon, but we'll see. Focused on affiliates for now. If we find success with affiliates, then influencer marketing, influencer campaigns paid become less of a necessity.

Speaker B: Yeah, and affiliates give you a way to test the waters almost because you could email or contact like a bunch of influencers and be like, hey, we have this affiliate program and like you're welcome to use it and like make some money because like these kind of hooks do really well on the channels that you post on. So like you could give it a shot and maybe they try just with an Instagram story and hey, that you know, they're making 500 bucks a month off their Instagram story. Like you guys can watch who is posting and how successful are their posts like through the affiliate kind of compensation uh, scheme. And if that is like going well then you can be like, okay, now let's look at all the other category of influencers that just want cash and are a bit afraid of the whole affiliate thing. Maybe we can try some of those now that we know that the model in general works and you know, go from there.

Speaker A: That's a good point. Didn't think about that. But they are very closely coupled and yeah, we can take learnings from here and apply them there. Makes perfect sense I guess.

Speaker C: On the affiliate program side I've heard there can kind of be very like 99% of users kind of just don't send you anything and it's the 1% that really sends you something meaningful. Do you find that's true in your case as well? Absolutely. Ah.

Speaker A: A lot of people like the idea of earning commission by just sharing something on their socials. But then to put it in into practice, some of them they just half ah, asset. I don't know if I'm allowed to say that, but um, yeah, no worries. It's the ones who put in the effort and who uh, it's usually it's the ones who actually believe in the product. Right. Those who use it, they try it out for a while, they think it's awesome. So mostly the inbound affiliate partners are the ones with the highest, the outbound ones which I reach out to, they generally don't go anywhere. So um, it's kind of making me rethink our model and whether we should reach out to our own users and see if any of them would like to onboard those affiliates as opposed to reaching out to influencers on social media.

Speaker B: Okay, so you, you're focused on day 30 when it comes to acquiring users. That's all well and good, but long term you need to retain users for longer than 30 days. So how is the like, what is the retention picture like?

Speaker A: Yeah, it's an interesting one. We have recently shifted a lot of our focus from acquisition and conversion to retention. Previously the name of the game was acquiring users, it was the funnel. Right. Top of funnel, mid funnel, low funnel, Conversion optimization. Mhm. We ran all these experiments with our paywall, the timing of it, our uh, pricing, all of the above. Right. We didn't leave a single stone unturned. But in doing this we largely neglected retention because we were just optimizing, we needed our campaigns to be profitable. Now we see the value in retention. We obviously knew there's a lot of value in it, but now we see the value in optimizing for retention. So we had, we've always had let's say quick win, low hanging fruit, push notifications, things like inactivity notifications, customer reminders that users can set to bring them back to the app. But now we're really honing in on uh, retention and we have a couple things in the pipeline. One, users logging in. As we discussed before, we want to nudge users to log in and we can subsequently use their emails for marketing for uh, unturned campaigns.

Speaker C: Right.

Speaker A: Bringing them back to the app, maybe offers, if they turn at the paywall and they don't pay, giving them a discount. So email marketing is one for attention and unturning users two is a bit more personalized on the push notifications. So at the moment we just have some basic logic that you know, wide sprays push notifications to users based on certain criteria. We want to get a bit more granular with that. We have quite a few ideas Things that we can do, uh, pieces of info, the timing of the notifications. So that's another thing. And three, and my co founder is really excited about this is gamification. So we'll start small with this. We want to add uh, a logging streak for users, you know, people who come to the app on a regular basis, who log their entries, uh, people who log their weight on a regular basis. We want to add the concept of tasks, things that they can complete, onboarding tasks and then, you know, recurring tasks. Something we're really, really excited about, really. And we definitely see lots of value in building these habits and users because to zoom out a bit and talk about our space and not just the, the tech and the business, at the end of the day, most of our users are here for weight loss, right? And weight loss is an interesting journey. It's an interesting problem whereby the biggest factor and the biggest determinant of success is consistency. So what we are trying to do with our app is to help users be consistent with their journey with their habits, whether it be eating, whether it be tracking their exercise. And gamification is a way of bringing users into the app. It's bringing, it's keeping them coming back, it's giving them a bit of dopamine to help them with consistency. You know, the technical term for us is retention, but for users it's consistency. And we really want to lean into this and help users even more. So it's fine and dandy if someone downloads the app, sees the value and purchases a subscription. But in practice, if they're not getting the value out of the subscription, if they're not coming back to the app, logging their weight, logging their food, losing weight, meeting their goals, then we wouldn't have provided them much value. We would have been a gimmick. And we don't want to be a gimmick. We want to be a product that actually helps users, that helps them with their journeys. For one, because we are able to help users, this is, you know, why we're doing this. We built this up in the first place. And two, we see the value of, of uh, product led growth. The end users will become users who use the app successfully and hit their goals. They become walking billboards for our product. Right. How did someone lose weight? They tracked. How did they track? They use this app, which they fucking love. Again, sorry for my uh, loose.

Speaker B: We like a little bit of spice, a little bit of spice in the podcast.

Speaker A: So um, yeah, so we definitely see the value in retention as a KPI, which we're focusing on now and we have a couple sprints coming. Full focus, engineering and product wise on retention. So super looking forward to that.

Speaker B: Yeah. So I'm curious, like, how far along are you in the process of designing and kind of ideating the gamification experience overall? All because I imagine like there must be some urgency, right? Because I would guess that New Year's is probably a pretty big time for you guys, right? Because like everyone sets the New Year's resolution and then losing weight is probably like the most popular New Year's resolution of them all. So is there, is there. Do you feel like you're on a big time crunch to like ship this by New Year's or not so much? And then again the like, where are you in the process of designing all of this?

Speaker A: You're absolutely right about New Year's. You know, holiday, uh, season, like people go spend time with their families, eat lots of food, uh, people always want to lose weight by New Year's. Huge. For us, you know, it's when we effectively launched our app the first time around. It's last, uh, end of December 2024. Yeah, 2024. But frankly we're not uh, we don't feel the sense of urgency because once we set our minds to this uh, feature or the set of features, you know, we work very quick and we were very, um, agile. In the next couple of weeks we, we will have our gamification gamification features launched in three weeks. It would be live to all users who have our app. So we don't feel the time crunch before New Year's. Although that said, we definitely take an agile approach to this whereby we're going to start small, we're going to start with a streak, m, maybe a notification that is related to their streak. Then we'll start growing from there. But we are very careful with our product in that we do not add a single feature to the app unless we are 100% confident it will not cannibalize or decay the user experience. 1, 2. We need to be 110% sure that it will not complicate the user experience because at the end of the day our value proposition is simplicity and speed. So we're very careful with this. We'll start very small with a streak and then very carefully start to trickle in new features one step at a time. New gamification aspects.

Speaker B: Yeah, so it sounds like you'll be building this soon, so I'm curious. So this streak, obviously we think a lot about streaks because we've built A trophy, the streaks feature for other folks to use, among other gamification features. And the streak thing. We've spent a lot of time looking into the psychology of streaks and, like, why they work when they work, when they don't work. And the big con of a streak, let's say, is that streaks are really good at improving retention until someone loses the streak. And then it's really tricky because when they lose that streak, they now are faced with the reality that, like, the next time they come into the app, they're going to be looking at a big fat zero. And it feels really bad. And there's this kind of, like, psychological thing, like, maybe this app isn't right for me after all. And you can actually kind of get burned on that. And I think in your case, that might be even potentially exaggerated more, since it's weight loss, which is something that can be even more emotional than maybe other types of things, like you'd imagine, like language learning doesn't have the same sort of emotional a baggage. Maybe that weight loss can have.

Speaker A: Interesting. We thought about it briefly and discussed it briefly and then assumed that if the conventional wisdom is that streaks in apps are good, then we should likely launch it and see how it goes. Right. We can learn the lessons along the way, but we have not delved, uh, into the details of it. I guess there's your duolingo approach of streak freezing, or like, recovering your streak while watching an ad or something along those lines.

Speaker B: Yeah.

Speaker A: We also have an easy out, which is users can return to previous days and log their entries on a previous day. So if put comes to shove, they can kind of recover their streak by logging something per day just to prevent this emotional backlash of losing their streak. Not something we thought about. We've just kind of assumed that streaks would be, um, a net positive. I'm curious, what have you seen in your data of the usage of Streak? Does it have a net positive impact on retention or could it have some negative impact?

Speaker B: It does have a net positive impact on retention. So what we've seen with streaks is we've launched streaks for different customers kind of incrementally, like one part of it at a time. So, for example, we first launched streaks with our first customer as email only. So, like, it didn't even show up in the app, but you get an email that would say, like, once a week, that would show your progress, and it would show, like, your streak kind of during the week, and that already had a measurable bump, small but a measurable bump. And then we you know, for other customers, we had the email, but then also we displayed it in the actual. In this case, it was a web platform, but it was the same thing. Like, we displayed it there and the bump is a little bit higher there. But then the most, the biggest impact by far is sending someone an email or notification when they're going to lose their streak to give them the chance to go in and extend it. That one is the biggest one. Like, when we first started sending those emails, we would send them out like four, four hours before the end of the day and the user's time zone, right. So you have to be also careful about time zones with these types of notifications. But we'd send them that notification and say, hey, like, you know, you haven't logged in today. You should extend your streak. And I think it. Charlie, was it like, I think it was one in five. It was 20% of people who got that email ended up extending their streak that, that day. Um, not even taking into account who opened the email. That's just like, of all the ones the emails we sent, like a fifth of them actually went and extended their streak. So that, that was the biggest lever, actually, is making sure that your notification system is like, really hooked into the streak feature or else people will just forget. And if they forget, no matter how good your streaks, like, look in the app, if they've forgotten, they've forgotten. Right. And it's too late.

Speaker A: Interesting. Okay. Really, really good to know. Actually, I'm just going to note that

Speaker C: down because, yeah, sure, the streak is great as a mechanic and as a thing, and it's very famous, but it's not actually the thing that causes the retention, I don't think. I think the streak as itself is just a mechanism, as an excuse to contact people. And I think that is actually the retention mechanism that's more important than the actual UI and whatever you build in the app, uh, like, that's great that people hook onto it, but the reason they hook onto it isn't because it looks nice. It's because it's about the notification and getting people to come back. So I think that's actually the point of the shriek.

Speaker B: Uh, there is a commitment, uh, aspect, uh, as well. I think if you just send people a notification every day saying, hey, use my app, like, it wouldn't work as well. It's the reminder. But it's also the fact that you've asked the user to commit. Like in the case of. It only works with certain types of apps. Right? Like, if your app is Like, I don't know, an accounting or personal finance app or something where you don't actually need to like log in every day to do something there, then the streak doesn't make any sense and there's a net negative. But for apps like language learning, for apps, uh, anything that's like personal improvement that requires a regular opening of the app and like taking some action, then it works in part because of the reminders, I think, and in part because the user is committing like to that goal and they're being very explicit about it. Like, I want to log in every day, I'm committing to doing that. And Duolingo has run a bunch of tests on this where they ab test even the language by which they ask the user to commit to the streak, like the specific words that they use and they find that it impacts the retention and everything. So, yeah, I think it's those two things together, the notifications and the commitment from the user.

Speaker A: Super interesting. This is one of those insights which you get when you're obsessed over, obviously in your case, your solution of, uh, gamification. Right. It's one of those very specific insights which very few people will have access to because we generally see the vertical of the whole mobile app and not the horizontal of gamification across multiple apps. What's another one of these insights? I'm curious, uh, apart from Streakk, is there something else which is a bit counterintuitive?

Speaker C: Yeah, leaderboards is an interesting one. M. I'm not entirely sure leaderboards would work for your use case, but it's something that we do for a lot of the customers. And everybody thinks you create a lead world with all your users in because that gets the most reach. Everybody's in the same bucket, everybody can see where everybody else is. But if you think from the user's point of view, uh, if you're in a million and one oneth place, that's actually really demotivating to see. You don't want to come in and see that. And you've barely moved this month. Like, what the hell? There's no point in me even trying this. So the conventional wisdom on that just really doesn't apply. We found. Well, we actually have limited the leaderboards feature in Trophy to a thousand participants. So you can only actually create a leaderboard with a thousand people in because we don't want people creating leaderboards any longer than that. And uh, the wisdom there is like actually break the leaderboard down by some useful attributes. So it could be location or it could be language in Duolingo's case, but whatever the useful breakdown attribute could be, actually bucket people up into smaller, uh, categories.

Speaker B: Yeah, the leaderboard one is, I think we talked to. I can't remember which of our prior guests in the podcast it was, but it was like basically they added a leaderboard and they put everyone in it and it actually had a negative impact. Like, it had a negative correlation with retention and all their good, you know, the metrics they wanted. Because it was super demotivating. Like, you get put in this massive thing and you just get to see how shit you are compared to everybody else above you, and it's really, like, it's really no good. Right. So you need to, like, bucket people up in leaderboards and, um, group them with like, some grouping that gives them a chance at being in the middle. Duolingo does this with leagues. Like they have this, this XP feature or whatever that goes up as you, as you complete lessons. And they just, if you start to do, you know, worse, they just put you in a new league where you're. Now, you're in the middle. So it's like you're never, you're never really feeling, like, demotivated because you're at the bottom of a leaderboard.

Speaker C: So.

Speaker B: Yeah, that's definitely a big one as well.

Speaker A: Interesting.

Speaker B: Have you thought about any other particular gamification features outside of the, the streak? I know you mentioned something along the lines of like, tasks or achievements maybe. Have you guys given that some thought?

Speaker A: Yeah, we have a few things on our minds. We have the concept of achievements, badges, you know, different milestones. So let's say your weight loss goal is, uh, 10 kilos in total, right? You're at 90, um, 5 kilos. You want to go down to 85, then, you know, breaking that up into four different milestones and getting a badge, uh, you know, 10% of the way, 25% of the way, so on and so forth. Also logging streaks, those translating into badges as well. And, you know, badges effectively things that you can share, representations of your progress, I guess. Um, maybe also progress, uh, from like being able to collect all of the badges is also another thing which brings people to the app. Badges is one milestones kind, uh, loosely coupled to badges. And then the other big front we see is social, you know, social gamification. Like you mentioned, logging with friends or accountability with friends. I think a huge aspect here is accountability, right? Being showing other people or giving other people access to your consistency stats, letting other people see what you're doing, being able to keep you and you know, a friend or family member accountable to each other by logging?

Speaker B: Yeah. Is that something that people frequently do in a group or like, they try to do together to, like, motivate each other? Because if it is, then I think that is a great. Like, any time you can get people to commit to something together, they're way more like the retention will be way better than if they were doing it on their own. Because now they feel that social pressure of, like, if I quit, everyone's going to know that I quit. So I can't, you know, I can't do that. I got to, I got to stick it out. So is that something that, that folks do already or is it mostly a, uh, kind of individual solo journey?

Speaker A: It's mostly an individual solo journey as opposed to people tracking and logging with their friends. That said, there are many instances we've seen from our users who reach out to us with these requests where, let's say a couple, they both are tracking and they both use the app. They purchase subscriptions together, and they both use it and are consistent. Um, we've also, we're very aware of, and a lot of our users who work with dietitians or nutritionists or personal trainers who would like to share their weekly reports with their health professionals. Although generally it's not something you would do in a group of friends, you know, a couple people logging together. There are many instances where sharing your progress, your data and finding accountability, uh, with other people is done through our, through our app, uh, generally in the form of sharing their weekly reports, their monthly reports. Although, um, we are trying to find a way to build this inside the app as well. So maybe being able to track someone else's progress live, or two people sharing their progress with each other live. We just need to gauge the general demand of this and whether, again, coming back to our North Star, whether it would keep the app simple or whether it would increase complexity beyond the threshold that we're comfortable with.

Speaker B: We didn't talk too much about the monetization and the actual, like, checkout funnel and like the pricing model and all that stuff. Have you always just been, is it just a monthly subscription or you can buy annual or is there like a weekly options? How have you thought about this and have you kind of experimented with it at all?

Speaker A: Yeah, um, we currently have two products, monthly and annual subscription. So you either pay by month or you pay by year and get a significant discount, uh, around 60, 70% discount depending on the market. So two thirds of our users go for the annual and we Again, nature of our business mandates that we push people to the annual so we have more upfront revenue which can refuel our uh, our growth. We've experimented a lot. We've experimented with pricing. We have localized our price to every single market. We loosely follow the Spotify index where our price, let's say in the States is uh, for an annual subscription is $40. In Cameroon it's $10. Right. In India it's 13. The list goes on. Right. So we have this localized pricing which allows us to be operational and grow our subscription base in every country in the world. And we found success in this, especially in some markets where the price is exactly right. And now we're experimenting with pricing globally as well. So price increases, price decreases, really finding that sweet spot of profitability without sacrificing volume because we all appreciate that volume is very, very, it's a strong determinant in organic growth and word of mouth growth. So we have looked into weekly product, uh, weekly subscriptions. We did not launch that because various reasons, you know, reducing complexity and also it's a product is not something that a week is uh, will do much. Yeah. Generally users who go for weekly subscriptions want to test out the app. We already have a generous uh, free tier which allows users to understand the functionality of the app, get uh, the wow effect and move on with their lives. All that said in terms of monetization, we've worked a lot on optimizing our conversion rates, our paywall, the timing of our paywall hitting users in the window of intent. Because generally with weight loss and um, our category of apps, there is a window of motivation and window of intent for users. It does not last very long. If your free trial is too long then users will not convert as much. So our approach has been to find that um, the sweet spot where users convert and once they convert, keeping them on the app, getting them back, working on retention and making sure that they are getting the value out of their subscription.

Speaker B: Yeah, I'm shocked by the 2/3 annual. That's, I mean that's great. That means the business is much more durable than if it was flipped the other way around. Right. Because you have that annual instead of monthly like that. That's real ARR. Not like, you know, some folks post their ARR, but it's not really a

Speaker A: annual ARR times 12, it's a vanity metric. ARR times 12 subscription apps is a vanity metric. It does not really. Yeah, but uh, we post it as well. We know it's a vanity metric. We're very transparent about it. But so.

Speaker B: Well, in your case it's more accurate because a lot of this revenue is actually annual, at least recurring. I guess we'll see in 12 months. But. But annual, at least it is.

Speaker A: Yeah. I will point people to the awesome resource by revenuecat, which is the state of subscription apps report where you can see what uh, these metrics are, what you know, the uh, general trends in the industry are. And you can see that generally a third 30% of annual subscriptions in health and fitness renew for a second year. So what your ARR is, let's say all of your subscriptions are annual, 1.3. That is your actual ARR for two years and then one point, um, uh, four, you know, an extra 10% for the second year. And we've seen the same numbers in our retention and our renewal rates as well. So yeah, ARR is a vanity metric. If you want the actual numbers, we can probably wrap one together based on the revenue cut source report data which uh, is very, very, very accurate.

Speaker B: So the LTV for most apps then is basically like 1.4, like ish of the uh, of the pre annual price, right?

Speaker A: Bingo. For, to be very specific, Health and Fitness.

Speaker B: For health and fitness.

Speaker A: Yeah.

Speaker B: All right, I think that's a wrap then. That was a great chat.

Speaker A: Thanks guys. Yeah, ah, no, that was, that was really good.