How the World's #1 VPN App Reached 1 Billion Downloads – Tanuj Chatterjee, Super Unlimited

David Barnard:

Welcome to the Sub Club Podcast, a show dedicated to the best practices for building and growing app businesses. We sit down with the entrepreneurs, investors, and builders behind the most successful apps in the world to learn from their successes and failures.

Sub Club is brought to you by RevenueCat. Thousands of the world's best apps trust RevenueCat to power in-app purchases, manage customers, and grow revenue across iOS, Android, and the web. You can learn more at revenuecat.com. Let's get into the show.

Hello, I'm your host, David Barnard, and with me today, RevenueCat CEO, Jacob Eiting. Our guest today is Tanuj Chatterjee, CEO at Super Unlimited VPN.

On the podcast, I talk with Tanuj about the product-driven growth loop behind the number one VPN app in the world, why they intentionally leave money on the table, and how the prettiest design often loses in their AB tests. Hey, Tanuj, thanks so much for joining me on the podcast.

Tanuj Chatterjee:

Thank you for inviting me.

David Barnard:

I said joining me, but it's joining us. Jacob, nice to have you here today as well.

Jacob Eiting:

I'm just here, David. I'm here to bring what I can, interject philosophy as needed, but I'm excited to be here.

David Barnard:

We had a lively conversation before hitting record, so one of these days we'll have a Sub Club after hours and talk about life and business and everything else.

Well, I wanted to kick it off today talking about your app, Super Unlimited. And I did not know, kind of embarrassed to say this, but I had not heard of it until you and I met at an App Growth Summit, and I looked it up and I was like, "Holy crap. This is the number one result for VPN and just an incredibly successful app."

So as we get started, I wanted to give you an opportunity to just talk a little bit about the app and the success it's had to help people understand and kind of frame the conversation.

Tanuj Chatterjee:

Yeah, sure. So thank you for the kind words. If you search for VPN, it's the number one app on the App Store. In the productivity category, before all the ChatGPT clones, they came in, we were always on the top 10, along with Google. The only other products were from Microsoft and Google.

So, proud of what we have done over the four or five years. As we talk through the podcast, you see a lot of it is very product-driven growth, but at this time we get over a million downloads a day. Most of them are organic.

Cumulatively, we have crossed well over a billion downloads. It's also a category where the whole pie is increasing in size, and our slice of the pie at the percentage is also increasing in size. So we are in a good spot there. Yes, we have not done too much PR marketing branding yet, but that will come over time.

David Barnard:

A million downloads a day. That's almost an unfathomable number. I didn't even know apps were still getting a million downloads a day, except for when they go viral, or ChatGPT, I'm sure, had a few days like that, maybe a few million. That is just an insane number. And then you said most of that is organic. So I wanted to jump to that next. How did you end up at the top result for VPN and how do you defend that?

Tanuj Chatterjee:

Yeah, we didn't start being on the top. We didn't start trying out to be the biggest VPN. My background, just FYI, is engineering and product. Before getting into the VPN world, I was CTO at Rakuten where we grew the company from $30 million in revenue to billion dollars in revenue in the US. A lot of lessons learned around the consumer DNA then. And given my background, I think very bottom-up rather than marketing down, sometimes to a fault, but it has served us well.

So we basically looked at the app, started looking at the reviews right away, started looking at our competitor's reviews right away to see what's missing in our app, not in terms of feature functionality, but, "What are people complaining about?" And we started fixing those things. We started removing friction. And then we started reading the competitor's reviews in terms of, "What are the challenges over there that we can fix and resolve?"

We basically built an app which was very, very high-quality. I say to people sometimes, "VPN apps are commodity," as you know, right? It's very easy. The open source scores glow in the internet, so you can find and build a VPN app in two weeks. The first 85% is very easy, but after that, every 1% is hard work.

"How do I connect a user who's in Poland to the best server? What's the right server? What's the right protocol?" And then there are blocking and tackling that we'll have to do. Some people might block you because an airline might block you thinking that, "Oh, you might be a bot." "How do I give them very, very good service quality?" That's really hard work. So that's what we focused on a lot.

It's a premium app that helps because there's no friction. You just come in, no login, no very, very heavy paywalls that are hard to get around. You just start click on connect and start using it. Our engagement time is literally half a second, half a minute. The first time they come in, they'll have to do a very quick setup, but there's the second time they come in, they come in and they click on the connect button, they go off to their Netflix or to their browsers and all of that.

And that's what we want. We don't want people to stick around too long, unlike a lot of other apps where engagement... We define engagement differently. It's not time spent on the app. It's actually the intention to come back to the app.

David Barnard:

And it seems like that is kind of feeding the algorithm. It's like people download the app and then it's so easy to use. You get these massive downloads, and then that just keeps this perpetuating cycle. Are there any specific things you've done and experimented with? I know Ariel at Appfigures talks a lot about rating velocity. Have you moved the rating prompt earlier or later and seen moves at all in your ASO?

Tanuj Chatterjee:

Yeah, ratings are super important, but there's a time when to ask for ratings. You can't ask it too many times and too off one, and at the wrong time mainly. We basically ask for ratings after we've delivered value to the user. That, multiplied by the number of downloads, the number of users who are coming in and using the app, and then coming back, that helps overall.

So yeah, so rating is important, very important to us. When people come in and look at the app store, not only we are number one, they also see two million plus very high-quality ratings. That definitely is impactful, but we don't try to do too much gymnastics with trying to get more ratings. I think they come organically, we ask for them, but we don't overdo it.

David Barnard:

Speaking of not doing too many gymnastics, your screenshots don't seem especially fancy. There's a lot of trends in screenshot design now of bigger text and gradients, and span two screenshots with the device image. But I looked at yours and they're pretty simple and straightforward.

The biggest thing I noticed was that you use logos like HBO and Netflix. And so you're kind of telling people, "Get access to these services no matter where you are."But is that intentional? Any lessons for you on screenshots? Have they just not mattered? Have you experimented all with them?

Tanuj Chatterjee:

Yeah, screenshot has been the biggest mystery for us. So we have tried. So of course, for different regions and all, we try different screenshots, localization and whatnot that everybody does, but our screenshots, and you've been very polite, they're a bit stale. They look a little stale. And we have tried modernizing screenshots. And I have to say, 80% of the time when we do AB tests, they lose.

And it's like, "Huh, people just like to see what they were used to seeing." So in many cases, we have gone back to the original ones that we had. In some cases, like I said, 20% of the cases, they do work slightly, but we have not seen any meaningful difference when we experimented with screenshots.

We also don't, because we are number one on ASO, we also don't do massive changes at the same because it's very risky. So we have done tweaks, we have done new screenshots, we have done AB tests, we've changed the colors, we've changed what the content, we've done all of those things in a very organized, methodical way, but most of them have not produced results that we were expecting.

Although, you would think that they're better and more modern and should be more compelling in 2026, but it doesn't work. So again, I think the data is proving us wrong, which is fine. It is what the truth is.

David Barnard:

Yeah. I just got done with a live YouTube paywall roast. So we spent an hour on the Sub Club YouTube channel live roasting paywalls. And one of the things we talked about was you sometimes quote, unquote, I don't mean to be disparaging of your screenshots, I don't think they're ugly, but sometimes the super fancy design just doesn't perform as well. And I don't understand it. My bend is always like, make things prettier. I don't understand why-

Jacob Eiting:

It's not that the technical elite have some sort of out of step actual understanding of what people want or like. No, it's not that. It's that people have no taste, David.

David Barnard:

People have no taste. Well, next I wanted to dig into SEO, 'cause I've had so many people, I mean, maybe less so the last year, like now it's more about LLM optimization, but so many people over the years have asked about breaking into SEO.

And I know it's something, a project you all worked on. What are your lessons from there? And then to tack onto that, have you seen the work you did do in SEO start to pay off and showing up in LLM responses?

Tanuj Chatterjee:

Let me go back a little bit. So till beginning of last year, we were only a mobile app. We were just a one-page site because there was no need for it. So last year, late last year, we launched our Windows product. And to get ready for that, we had to start building a web presence. SEO takes time. So we said, "In 2025, where do we start?" So what we did was we started with programmatic content.

So we worked with this company called Daydream. The founder, I knew the founder from before, he had actually worked in the VPN space before, so he already knew the domain. So I said, "Okay, let's help me out." So yeah, we started seeing results there, but the conversion was still not that great. So it's still a work in progress for us. It's still early. I'm not an expert there.

I know SEO in general has suffered. Every consumer company that's reliant on the web traffic is suffering a little bit. Both SEO is down a little bit because the first results that you see are Gemini and whatnot in Google, as well as even paid search marketing is less effective because the real estate is gone. Part of the real estate for the first page is gone because Gemini takes it over. So, a lot of people are struggling.

So I don't think we will spend time on SEO, but we won't over index on SEO. I think one of our biggest channels for growth, even to send them to make them start using Windows, is through our big mobile user base. So we'll spend a lot of time on product marketing within mobile. SEO, we know it'll take time. We are still working on it. We have content people, we have SEO people, but that's not the top channel for us.

David Barnard:

It's funny because I've read the case study about this and it's doing like 1,600 app downloads per month, which for at your scale, that's just such peanuts, but for a lot of apps, that would be a pretty meaningful improvement to download. So it's all about scale there.

But I did want to follow up with that last point about LLMs. Have you found that because of the SEO efforts that you've put in, have you started showing up higher or what's your LLM? I forget what it... Either of you know, what's the LLM term for SEO?

Jacob Eiting:

GEO, right? Generative engine optimization. It's odd. It should just be like LEO, but that's low earth orbit so it's [inaudible 00:12:04].

Tanuj Chatterjee:

It's interesting you asked. I was looking at Google Analytics the other day for our web. There is traffic coming in. There are downloads coming in from there. It's not in the top five channels, but it is coming, so it's a good sign that we are starting to get more visible.

Jacob Eiting:

It's pretty interesting. I think we see so much higher on the DevTool side, but it makes sense. I feel like choosing a VPN is... You know exactly what you're going for at that point versus... This is my naivete about the VPN market, but it's like I don't suppose people are being like, "How do I tunnel my track?" You know what I mean? They know they need a VPN 'cause somebody told them, "If you want to do this, you need a VPN." Right? And so they should search those three letters.

Tanuj Chatterjee:

Yeah. And also I would say there are bigger VPN players out there, not bigger in terms of downloads, but there are few VPN players who are premium only, while we are premium.

So their downloads are smaller, but their LTVs are much bigger, and they came in from the desktop side of things and they've spent years focusing on SEO. So if you're looking for, "Oh, I'm in Bolivia and where do I watch the Winter Olympics right now?" I bet you that you'll find a website from them.

Jacob Eiting:

Even on consumer, 'cause as I could say, the VPN market's really funny because things like Tailscale and Citrix are also in the VPN market, but don't feel like direct competitors to you, right? No.

Tanuj Chatterjee:

Some of them are more and more catering to the B2B2C side. We are purely consumer play. Yeah.

David Barnard:

I did want to dive a little deeper into the freemium side of things, 'cause it does seem like this is a big part of your growth engine, is having that freemium tier, which allows you to get so many downloads and so many ratings, which then gets more downloads. How have you balanced the monetization? And then as a VPN, how do you manage costs for those free users?

Tanuj Chatterjee:

Our free-to-paid conversion is quite low, and I'm okay with it because we have focused on the top of the funnel a lot. There are a lot of companies, smaller competitors in the freemium world who would try to get more aggressive in monetization. And I think that's not our philosophy. We want to give you very, very, very, very good quality, free product.

And the reason it's low, not because the product sucks, it's because our free version is that good. So you don't want to throttle or take things away from the free version. There's a reason we have the word unlimited in our name, where if you want to come and you can't pay, that's fine. You probably just will have to watch an ad or two, and then you can use it unlimited. We have dozens and dozens of countries and cities that you can connect to. It's just fine.

I would rather give people more value before asking them for more money. So, it's a denominated game. So, part of the reason you also alluded to before, that most of the users are organic, partly because it's a numbers game for us. Our LTV, because it's freemium, because of a big majority of the users are freemium, our LTV is not that high. We can't compete with Spotify's of the world where their LTV and retention is amazingly high.

But we are adding a lot more now. Talking about adding more value, we are adding Windows. So not only you can protect your mobile phone, but you can protect all of your devices, Windows, Mac, mobile phones across the family. We talked about adding a second phone number app so we can throw it in the bundle. I think that's where we are going. But the main thing is our top of the funnel is our superpower. Don't mess with it, give people a fantastic free VPN experience.

David Barnard:

So do you manage the ad revenue to break even on the free users or do you even lose a little money on the free users? And how do you make money on ads? Are they full screen interstitials? Do those pay reasonably well?

Jacob Eiting:

Do they get injected into your webpages?

Tanuj Chatterjee:

No, no. So it's limited. So I mentioned our average engagement time is 30 seconds on the app or less. So you don't have much of a window to show an ad. So to your point, yes, we have interstitials and we have full screen ads, and we have some native banners and all, but we don't try to optimize it for profits necessarily. Some countries, high rev countries, US, Western Europe and all, they're profitable.

And in some cases, even the match rates are low, the CPMs are very low. It's totally okay with us. So I came from the e-commerce world. We used to get spikes on Valentine's Day and Mother's Day, and Black Friday and Cyber Monday. Our spikes usually happen because of geopolitical issues. So I'll give you an example. I think it was last year, Turkey banned Instagram for 11 days, and we got 15 million downloads in 4 days.

And if you think of Turkey's population, it's 85 million. And if you think of 50 million downloads and think of the internet population using it, it's probably 30 to 40%. So, do we make that much money in Turkey? Probably not. There are countries where there are more issues that are poorer. Myanmar, for example.

It's having a civil war, but we just think it's the right thing to do. People should have access to information. People should access to free information. They should not be constrained by just propaganda. So we just provide service even if we lose money. As long as, yes, we are running a business, but we are not trying to over optimize this.

Jacob Eiting:

Are you on ads? Do you defer showing ads until you've determined somebody's not a likely purchaser? Do you have any orchestration between showing ads versus not and stuff like that? Or are you thinking about it that granularly yet?

Tanuj Chatterjee:

For the first download, if you just install it, we don't show an ad right away, right?

Jacob Eiting:

So yes, I guess is the answer, is you do some sort of smart and be like, "Okay, we're going to defer to see if this person's going to... They might convert right away and we don't want to degrade the experience."

Tanuj Chatterjee:

It's not just for conversion. It's also, we want them to see the value of the app. So even the paywalls are not very hard. There's a big X sign where you can cancel out very easily. So, we are not trying to over monetize it. Just the main goal is come and use the product.

Jacob Eiting:

Yeah. I guess becoming their VPN app is probably more important than monetizing on that [inaudible 00:17:54].

Tanuj Chatterjee:

Money will come. Money will come. Everything else will come.

David Barnard:

It sounds like part of how you manage this is just to keep costs really low so that you're not bleeding out by serving all of these freemium users. Any lessons there about just keeping costs?

Jacob Eiting:

This is my naivete on VPNs, but you all have to maintain that tunneling infrastructure. Right? Or do you rent that out from somebody?

Tanuj Chatterjee:

No, we maintain it. Yeah, we spend million. So David, you're right, managing the cost is key, but not at the cost of service quality. So, every data packet for a VPN goes through our servers. So, if people in Turkey or some of the lower revenue countries, if they're using it, sometimes you lose out too, and it's okay.

It's okay with us. What we can control a little more or what we want to control is marketing costs. That's why we spend in marketing and we will continue spending more, but that's where we can control things. We won't make any compromises on service quality.

David Barnard:

So then, you're a business, how do you make money? What are the features that you put behind the paywall? How do you entice people to actually subscribe when you do provide such a fantastic freemium product?

Tanuj Chatterjee:

The extra value that we can provide with a paid product is, of course, no ads, top tier customer support, but also a lot more locations and access to higher 10G servers and whatnot. But now, starting late last year, multi-platform where once you subscribe, you can use the VPN for all the platforms.

So, we are upping the value game. That's how we make money. By the way, we are growing very nicely. We are very profitable. The investors are not pushing me otherwise, so it's a good position to be in.

David Barnard:

Yeah. What an incredible position to be in. The deeper we get into this conversation it sounds like there's so much dry powder, with a billion downloads and such a huge freemium base, that as you do find secondary product market fit, find new features to add, things like that, you've got so much dry powder to really build the business over time.

Tanuj Chatterjee:

Yeah. Yeah, we are looking at long-term because we are, unlike some of the other VPN companies that are desktop first, some of the bigger competitors that we have, we are mobile first. And so, one of the apps we recently launched, and we are cleaning it up and all before bundling it or promoting it heavily is this eSIM app. So think of this. So a lot of VPN user that we have, think of the cohorts.

There are a lot of people who are travelers. They just travel to Europe, but they cannot access their favorite content or they cannot access their bank and all. Or immigrants that want to watch a show, like a Korean immigrant in the US, they want to watch a Korean show and whatnot. So, there's a lot of access-based issues that happen once you cross the borders.

We think eSIM is a fantastic market and intersects very well with our user base. Just last year, 500 million people traveled to Europe. If you think of it's a very big part of the affluent global population who can afford to travel to Europe. Are there a lot of these people who are using our VPN products? Yes. Do they need eSIMs? Yes.

Jacob Eiting:

I was in Europe last week, and for the first time I did it, and compared to the pay $10 a day, get ripped off, slow access, it was pretty good. I paid $20 for the whole week. I had pretty decent access. It wasn't perfect, but it was pretty decent, and I didn't have to worry about... And you don't have to go to a shady shop and get a car to slide into your phone. It's pretty cool.

And I'll tell you, I think there's opportunity in the market, because the two I tried to use, the names you've heard of, the first one, their website didn't even work on mobile, it was so bad, which I was like, "It seems like this should probably work."

I think it failed because I was on my phone on bad service, which I'm like, "Pretty sure that's a test case you should probably have nailed, is was like your website shouldn't require JavaScript and be so crazy." But anyway, I buy it. I think that's a great strategic direction.

Tanuj Chatterjee:

Think of other values. So we have this massive customer base. What else will they value? It's not, "What else can I sell to them?" I don't even need to sell. "Here's what we have. Do you want to use it?" And they're privacy conscious users, they're access conscious users. I think with a large denominator, even a small change in the multiple of conversions will help a lot.

David Barnard:

Your denominator at a billion downloads, orders the magnitude larger than most denominators in the subscription app space?

Jacob Eiting:

It's a blessing and a curse 'cause once you're at those mega numbers... I mean, I guess if you're playing a flywheel game where you're taking the current volume and you're adding new things, that's fun. I think once you get to that scale, it can be kind of demoralizing 'cause you're like, "Well, there's just no more people."

Tanuj Chatterjee:

That's the part of it. The downside of being the number one is where will the future growth come from for the top of the funnel?

Jacob Eiting:

At the end of the day, it's return on capital. Where's it coming from?

Tanuj Chatterjee:

It's a very First World problem to have, but it is a problem nevertheless.

David Barnard:

So you mentioned earlier, we focused a lot on the organic side of things, but you mentioned that you do do some marketing, and it's something that it sounds like you're maybe looking to do more of so more people are aware of-

Jacob Eiting:

I was going to say, if you start selling eSIMs, you're going to need to get to know some ad buyers in airports. It's a very on-prem advertising.

Tanuj Chatterjee:

We are also learning. So eSIM, "Okay, we have so many users, just show an ad, and let's see how it goes." And unless people are traveling, there's no intent. Why would you buy an eSIM right now? So it's like, "Okay, good to know." Come summer or before summer, can we offer some promotions at the right time?

Not every phone is eSIM compatible either. Again, do we not bombard people with an eSIM ad, just be very deliberate about who we want to show it to, who'll actually get value out of it? So the ads is another one of those things you talked about ads. I would rather leave 10% money on the table rather than bombard people with that. There's a lot of trust that you destroy with overdoing monetization.

David Barnard:

And back to growth side of things, what have you done in paid marketing or has it almost been completely organic?

Tanuj Chatterjee:

Well, Apple and Google, it's for search ads. Yeah.

David Barnard:

Have you found that brand defense especially important or what wins have you found on search ads when you're already the number one result?

Tanuj Chatterjee:

Number one result on VPN. There are other keywords in there, too. Sometimes you also do it for brand defense, like you mentioned. There are other keywords. We focus on specific keywords, specific markets where we spend dollars on, and we look into the ROAS very carefully, and you can't go beyond a certain amount. We also see sometimes some of these ads will cannibalize our organic channels anyway, so there's always that risk, but it's okay. So we are trying to manage it carefully.

David Barnard:

It's fascinating at your scale that you do spend on that. And then philosophically, you're still trying to get a return on ad spend. You're still penciling it out versus just trying to be everywhere, because you already are everywhere with that number one placement.

Tanuj Chatterjee:

Five years ago, we were not number one everywhere. We were probably number one in a couple of locations. In iOS now, we track 71 countries, and right now for the VPN keyword, we are number one in 67 of those and number 2 in the rest.

 In Android, we are not as successful. We are number one in the US right now. We are in the top five or top 10 almost everywhere, but it's not that we are number one everywhere. So we do see that as an opportunities like glass half full or glass half empty. "Okay, where can I use some extra marketing push to kind of get up there?" Because I think the brand definitely helps.

David Barnard:

I wanted to transition and talk about some of your product-driven approach to growth. And one of the things I found fascinating was that you put your support team inside the product team. How does that work?

Tanuj Chatterjee:

The head of support reports to our head of product. I want to mirror in front of the product team right there. Our growth comes from product decisions and not just from marketing campaigns. Sometimes we know that, "What are people struggling with? What's not clear? What are people asking about over and over?" So that could be a UX issue, that could be a communication issue or an education issue, or an onboarding issue. Service quality. "I'm in Uganda and I'm starting to see a lot of service quality tickets." Product should know first right away.

I have been at companies where product and support was a totally separate organization, and support will talk, do a talk once a month, and show their KPIs and how many tickets they've sold. It's fine. But the pace at which the feedback loop goes into the product side of the engineering side is a lot less. So because we are so focused and anal about product quality, we wanted the support team inside the product team itself.

Jacob Eiting:

I mean, we essentially not exactly have that at RevenueCat, but support's been an engineering attachment at RevenueCat since day one. And it's both for cultural reasons, because it's engineers asking most of our questions, so it makes sense, talent reasons, but I just think that...

And the same with QA. I think QA, I don't know if you guys are doing it or any special, but when I got into the industry, it was very normal for QA to be in the other building kind of situation.

And I think now QA teams don't really exist 'cause it's all been merged. You have junior engineers. You have people on the team, and maybe it's support sometimes. That's kind of our closest thing to a QA team, it's like the support engineers involved in the product development, and that seems to work so much better.

Tanuj Chatterjee:

The fewer dependencies you have, the better. And if you have -dependencies, it cannot be the next building over, right? You should be able to

Jacob Eiting:

Well, we don't have buildings anymore, so we solved that problem.

Tanuj Chatterjee:

Right, right.

Jacob Eiting:

Everybody's in another building in 2026.

Tanuj Chatterjee:

Mentally even another building, right? Yes.

David Barnard:

You mentioned early in the podcast, and it was something, again, that stood out to me about your philosophy is the bottoms-up, not marketing -down. Any other top lessons there or anything else you wanted to share about that approach to product development and marketing?

Tanuj Chatterjee:

For us, for our domain, service quality is massive because there are always a lot of corner cases. Think about it. You have a phone, 3G network in Nepal, and you're on wifi in your hotel, and then you walked out and elevator for some time it disconnected, and then you are on a 3G line or a 4G line and then something else happens. In a bus, you're traveling, you switch from one tower to the other tower.

VPN in particular, as opposed to most of the other tech world where it's a quick request and response, it's a very long-lived connection. So technically, there are a lot of challenges that come with that, and our customers are on the move, so a lot of focus on that.

Jacob Eiting:

Getting all those edge cases right, it is the whole product. The last 10% is the whole product, probably. I think you said something to that effect at the beginning, but I think that's what sometimes differentiates great products and commodities, is do they just always work versus sometimes maybe work?

Tanuj Chatterjee:

So it's not just engineering. It's the operational discipline, that that is important.

Jacob Eiting:

Meaning reacting to all the issues, actually taking the time and investing and finding them-

Tanuj Chatterjee:

Anticipating, reacting. Our QA guy actually walked with his phone inside an elevator, down, he had a debug build so you could see all the logs at that time, and he did the whole thing a few times. So, that's how you get better.

Jacob Eiting:

This is just a mobile thing. I think in a lot of app developers, especially now we're seeing so many app developers coming into the market. And you kind of always assume good wifi, beefy connections, whatever, which is a rarity. Or certainly, it's not going to make your app great. You can be okay, but if it's...

I was using a fairly new app. I won't name names, but you've heard of it, and it's in the AI category, and their offline stuff just didn't work for me. I downloaded it and then I got offline and opened it up on the plane and I just got a login screen.

I can see that's a second order case. And maybe that's forgivable for our first year app or whatever, but in mobile, you have to nail that. You have to nail the offline case, the online case, the middling case, the changing countries case.

But I'm just kind of wondering, again, it goes back to VPN being a commodity, what brands could potentially just attach a brand to it? Because sometimes commodities, like Coca-Cola, it's the brand. It's the sugar water, whatever, but it's mostly the brand you're buying.

So are there brands that could just buy them being the brand they are, like DuckDuckGo is maybe a good example? They're even very niche. It might be like Symantec or something like that, but I think that's a 20-year-old brand now I don't think that, or whatever that doesn't really have the cache that it maybe used to. But are there big brands like that and they just are not investing at the level that you are on the product?

Tanuj Chatterjee:

No, there are bigger brands. And without naming names, there are bigger brands and bigger companies, and they come more on the desktop side in the VPN space.

Jacob Eiting:

But then that's a different problem. A desktop, you're not on the bus. You know what I mean? So is it your thesis, is that just by them not being mobile first... By being mobile first, you can just have an advantage because-

Tanuj Chatterjee:

We have an advantage, desktop people pay more.

Jacob Eiting:

Oh, sure, sure. Well, it's a trade-off. It's like scale low LTV, but there's a niche there. Right?

Tanuj Chatterjee:

We have to go and tackle and understand that market a lot more. The users are different. The mobile user case I just explained, they're a little more forgiving because they do experience these issues when walking-

Jacob Eiting:

Right, right. "I'm not using this to connect to my company's VPN," so it's different?

Tanuj Chatterjee:

Desktop users are less forgiving. Yes, the technical challenges are not that the person is moving and shifting to 3G. You can assume that the connection is strong, but let's say if they're gaming, the reaction time needs to be very fast, otherwise... You cannot have any latency in gaming.

Jacob Eiting:

I was going to say, I think that would be the only time I've used, well, except for work, a desktop VPN was when I was trying to do LAN games with people, and it was just easier to have a virtual LAN that you were on.

Tanuj Chatterjee:

So the challenges are different. Bandwidth is very high. If you did downloading big content or streaming in 4G and all, you need to make sure the bandwidth is top tier, the latency is as low as possible, which is the same-

Jacob Eiting:

And those are all things you control 'cause it's like you have the edge nodes, you have the relays, that's all infrastructure you have to monitor and pay attention to?

Tanuj Chatterjee:

Yeah. We control, but it's still a good engineering problem because you are connected to some servers for a very long time, for hours in many cases. So the load balancing happens in a very different way.

Jacob Eiting:

Yeah. This is off the Sub Club beat, but can you just host that? You guys just run that in a standard cloud infrastructure or do you need specialized infrastructure?

Tanuj Chatterjee:

No, no, no. We have special. So, I'm getting a bit geeky with you right now, but-

Jacob Eiting:

It's great. They can edit it out. This is just for us.

Tanuj Chatterjee:

All right. Our control plane is in the standard well-known cloud infrastructures. Our data plane is not because we lose our shirts because the bandwidth costs-

Jacob Eiting:

Right. Yeah, 'cause it's not designed for just tunneling a ton of traffic-

Tanuj Chatterjee:

It's not designed for that.

Jacob Eiting:

So you have to go to a layer lower to rent that?

Tanuj Chatterjee:

Yeah. Where we have control, we get bare method machines, we control it, we wipe it out, and-

Jacob Eiting:

Oh, okay. Yeah, that's very interesting. Which is also another advantage versus any other... David was talking about cost inputs. When you have that scale, you can afford to make the investment to control your data playing at way better rates. And that allows you to have a different retail economics than somebody who might have to rent it all from AWS and pay ingress, outgress or whatever?

Tanuj Chatterjee:

Yeah. So those cost optimizations we do. We don't do cost optimizations that will affect service quality. That's the last thing we will touch. We'll not touch it ever.

David Barnard:

This is great Sub Club content. I don't know what you're talking about, Jacob. We got a little nerdy there about control planes and data packets, but it is super meaningful.

Jacob Eiting:

People on this podcast make software still, I think. You know what I mean? This is part of how the sausage is made.

David Barnard:

And it's super meaningful to the business to build those cost advantages. And as subscription app businesses, this is part of what you have to be figuring out. And even in this conversation, I'm thinking about my dinky weather app, and part of the reason I don't have a freemium tier is because their data is really expensive.

Jacob Eiting:

David, invest in weather stations. You need a weather station in every place on the earth.

David Barnard:

Well, no, what a lot of apps do for weather is they do actually just use the National Weather Service data, which is generally pretty good, and it's free. And there are ways that we can control our costs so that we could offer a freemium experience to users. So it's inspiring to me to go this deep because I think a lot of people should be thinking along these lines.

Jacob Eiting:

RevenueCat's had a free experience and we designed it to be supportable from the beginning. All of our cost scaling is such that if somebody's not paying, if an end user's not paying, it doesn't really cost us much to support. And that takes design choices. There's ways we could have designed it that wouldn't have been that way, and that's allowed us to maintain this really nice freemium tier.

Tanuj Chatterjee:

No, but you open up the funnel right there, 10X. And it has to provide meaningful value. Right?

Jacob Eiting:

Yeah. And then the value is we can give the product away for free if you're small and just say like, "Hey, just use it. Don't worry about it. We'll get you later. Don't worry about it."

David Barnard:

In the era of insane inference costs, this is a very important topic as well. It's like-

Jacob Eiting:

Yeah, yeah. Every app that's AI driven is kind of really having to think about this stuff heavily.

David Barnard:

Can you run an open source model? But I think, Tenuj, you keep hammering on don't sacrifice service quality. And I think that is one of the challenges currently, is that the leading model providers like OpenAI and Anthropic and Google have set the bar so high from what kind of a response you could get that a lot of people got super excited about Apple's on-device AI stuff.

And my pushback the whole time was like, people just expect so much out of AI and you're not going to get that from a local model. And it's going to be a while. And then as soon as the local models get better, the leading models will get even better. But there are opportunities to cut costs by using open source models, by using different hosting and things like that. But if you do that at the expense of service quality, if your product degrades because you're trying to cut costs, that's not a good experience.

Jacob Eiting:

Yeah, this is really prescient in the AI moment for all these apps that are shipping AI features, too. It's also important to think about, 'cause I assume you all did not have your data plane in house on day one. Right?

Tanuj Chatterjee:

Yeah, yeah, yeah. So day one, you don't build for Google scale or for our scale. Right?

Jacob Eiting:

Fastest. It's all about the fastest, the best.

Tanuj Chatterjee:

You do the fastest. Get to market fastest. You can optimize the costs later.

Jacob Eiting:

Exactly. And then that's what I was going to suggest, is have a path. Just understand that you can. And it's like when we're building AI stuff internally at RevenueCat right now, it's like, "Don't worry about it. We'll spend the big, it's going to be expensive, it's going to be a loss lead or whatever." Two things are going to happen. The models are going to get cheaper in 6 months, 12 months, and then we're also going to get smarter about using them. So the costs are not going to go up from here. So the marginal cost, at least.

Tanuj Chatterjee:

Time to market is huge. Initially, your volume won't be that big. If the volume gets bigger, it's a good problem to have.

Jacob Eiting:

This is a classic early entrepreneur mistake when you over-obsess about, "I'm going to lose my shirt" or whatever. It's like, this is not going to be a problem. If it is a problem, it's a good problem. And also, I don't think AWS has taken anybody's house yet for not paying a bill. So typically-

Tanuj Chatterjee:

Don't optimize too early. Find the product market fit, grow the funnel. And there's so many levers that you can pull for cost without sacrificing service quality.

Jacob Eiting:

Yeah. I just love the philosophy of focus on the product. This is RevenueCat MO, is customer recession, "Just talk to customers, see what they want." That's our only guiding principle is, "Talk to them and see what's broken ,and then apply our gifts to try to serve them."

And ultimately creating surplus value, Which then you mentioned the money will come, and it will. If you create surplus value, the actually capture part is kind of the easy part. If there's so much surplus value created, you'll find a way to share in that. And there's optimizations. You can 1, 2X, 5X that one way or the other, but unless you focus on the first thing, which is the value creation, none of it else is going to matter.

David Barnard:

Well, actually, let's take a step back. So my understanding, correct me if I'm wrong, Super Unlimited was spun out of Rakuten. How did that all happen? And then we can tie it into the early days where you were just using off the shelf cloud hardware, and then when did it start making sense to build out your own infrastructure?

Tanuj Chatterjee:

Yeah, Super Unlimited was spun out, but not from Rakuten. So I had left Rakuten. I joined this company at that time that was called AnchorFree, and we had a few VPN products, so you can look it up.

But during while I was there, we had bought this product called VPN Super Unlimited Proxy. Got enough product, nothing like what it is right now. And I was there as I was running engineering at that company, I was their CTOs.

And then when AnchorFree was bought by a larger company, I decided to step away. My investor said, "Tanuj, don't go anywhere. Here's the product that we bought. Can you run it? Let's just try to grow it." "That sounds fun. Let's try it out."

I had managed engineering. I'm an engineer. I'd done engineering and then managed engineering all my life. This was something interesting. I was like, "Oh, what do I have to lose? Let's try it out."

And yeah, so I was not the founder of the product. It came from somewhere, but we grew it substantially over the last five years. So five years ago, beginning of January in 2021, we spun off Super Unlimited as its own company.

David Barnard:

That's a crazy story from app acquisition to spun out to the top VPN app with a billion downloads. That's crazy. Early on you were just on commodity hardware. When did things start picking up? What was the inflection point? What happened that really started to move the needle? And then when did you start building out your own infrastructure?

Tanuj Chatterjee:

It was not like one golden moment where everything changed. It was a lot of things along the way. Unfortunately, there was no one silver bullet, but like I said, we looked into reviews, what's keeping us away from being the number one product. A lot of it came back to service quality for us. So, we spent a lot of time on service quality. And thankfully, that's my DNA and that's how I think, too.

We started building a team. We started focusing on service quality. We started some basic things about ASO, localizations and whatnot, but mainly service quality. And then things started going up. In the last five years, as you know, a lot of geopolitical things also happened, where suddenly we'll get a massive spike. So as we were starting to go up, we will see a massive spike from country A versus country B.

And what happens is once the spike is gone, you hit a new normal at a higher level. And we were, like I said, in many cases we are not making money. We're losing money. There are many countries right now in this world where we lose a lot of money, but it's totally okay with us. That was totally okay with us even then. We said, "Okay, let's build a product that everybody wants to use and we are proud of."

And that flywheel took effect. We kept improving service quality. We spent some time on the air, so those things we did, but I'm sure there are companies that are doing a much better job than us on that one. But the thing that it's not important, it's important, and we spent some time on it. Did we do the best we could? At that point, yes. Are there better ways to do it? I'm 100% sure, yes.

So most of it was product, give the service to anybody who needs it, don't over index on monetization, money will come, all those things that we talked about earlier. And slowly and surely it went up. Talking about ratings. So we also broke down the ratings by different countries, for example. "Okay, this country, the ratings are quite a bit lower than US, let's say. Why is that?" So understanding and digging into the details helped us a lot.

Jacob Eiting:

One thing I've learned from this is that VPNs are a good inverse hedge against world order. So if Ray Dalio is right, everything's breaking down, I should just buy VPN stocks 'cause apparently that's a good hedge.

Tanuj Chatterjee:

So what we did was we expanded. We added a lot of new vendors, we upgraded our vendors.

Jacob Eiting:

Also, I imagine you have to internationalize too at some point. You need nodes in every country for good latency and stuff like this?

Tanuj Chatterjee:

So we have tens of thousands of servers in dozens of countries right now. Yeah. Coming back, we do use cloud for spikes. The downside of bare metal is it cannot handle surges that well. We can add a little extra capacity, but think of the Turkey kind of a situation where we added... We were practically serving Turkish population.

Jacob Eiting:

There's no solution to the spike problem. You either are way over provisioned and you're losing your shirt and you can't react faster. Even if it's a wave that takes days, it takes time to provision.

Tanuj Chatterjee:

So then you use cloud providers for surges.

Jacob Eiting:

Soak up the excess. Right.

Tanuj Chatterjee:

Soak up the excess very quickly. Bare metal, you have to order it and it takes a week for them to deliver, and nobody has 500 servers sitting around. Yeah.

Jacob Eiting:

Exactly. Well, somebody has to. So it's a financial operation problem to see who holds that, which is like going back to the thing, David, about building your own weather, parsing or running your own local models.

It's like when you add in all the costs of the hardware sitting there and not being used, at least if you're buying tokens from Anthropic or whoever, you know they're fully utilizing their hardware. So it's a much higher utilization rate. So you might be surprised that your amortized costs aren't that different.

Tanuj Chatterjee:

Yeah. Now, if you're building a company or a mobile app, which is not a VPN, just go with cloud. And prove fastest way, do it, and then optimize later.

Jacob Eiting:

Everything's portable, too. The nice thing about TCPIP, it's standard. You can do it on anything.

David Barnard:

One of the other things I did want to dig into was the 80-20 rule. I know that's part of your philosophy, but you've also talked again and again about service quality. How do you not over perfect everything and still actually ship? How do you balance and decide what's good enough when your North Star is service quality?

Tanuj Chatterjee:

We cannot compromise on privacy or trust, and we don't compromise on service quality. What we can compromise on is features. So, I'll give you an example. The algorithm itself for a VPN user, which server to connect to which protocol to use is very complex. And the algorithm that you need or the rules that you need to connect you in the US will be very different from somebody who's connecting from, let's say, Saudi Arabia.

We have built a lot of domain expertise to understand what we need to do for which country. We were building this algorithm and it would've taken months and months to build it right. So we said, "Okay, you know what? Do we have this algorithm what we think is right for these five countries? Let's just launch it." So, we used feature flags and we used it just for those five countries. And for the rest, we can build over time.

We could be wrong. We could have been wrong. In which case, we have to iterate on it and got it right. We could have introduced some bugs, so iterate it right now and get it right. So that's how we do it. We will do it in smaller chunks rather than build a big massive Taj Mahal and launch it.

David Barnard:

All right. Well, as we're wrapping up, there's three questions I now ask every guest. So let's start with, what is the biggest win of the last year? An experiment that increased revenue, 20%, anything like that? A new paywall? What was the biggest win of the last year?

Tanuj Chatterjee:

It's less about an experiment. It's more about us building and launching our Windows product. As I mentioned, it's a First World problem, but the downside of being number one is where's the new growth coming from? Where's the forward-looking growth coming from? And this is what we are betting on. This is one of the big things that we are betting on that'll drive us future growth.

We are already starting to see a lot of takers. People are coming in onto our website and buying it directly, to our surprise. Our website didn't even even exist, so our brand is known. We have to start doing a lot more product marketing this year. So, building a product that we are happy about, that we are proud of is great.

It took more time because we had a product. We were not very happy with some of the service quality issues. I'm sure there's still issues that we will find, but finally it was at a place where we are very happy about. We launched it, and we are very excited about us showcasing it to our users for future, for our forward-looking growth.

David Barnard:

I don't think many companies in 2026 would say that their biggest win was launching on Windows, but that's a pretty cool story.

Jacob Eiting:

Don't sleep on Windows, David. Don't sleep on Redmond. It's a lot of companies famous last mistake.

David Barnard:

All right. And then how about what was the biggest fail of the year? Experiment that went off the rail?

Jacob Eiting:

Not shipping on Windows soon enough. Sorry.

Tanuj Chatterjee:

I think not a big one, but a lot of little ones. We talked about the screenshot experience. We are working towards on improving our design and UI. It's a bit stale, still. It has improved a lot over in the last 12 months.

Sometimes, as we talked about, we tried new ASO screens. 80% of them failed because people didn't like it. The previous screens were much better. We just thought, "These designs are not that great, but it's what people like." So that's what I would say. If it's the biggest, I don't know, but it's one of the things that surprised us, and lessons learned.

David Barnard:

No, that's great. That counts biggest fail was changing your screenshots to a prettier design. Again, not a failure you would expect.

Jacob Eiting:

Never change anything. That's kind of vibe.

David Barnard:

And then the last question is growth would be easier if?

Tanuj Chatterjee:

If we figure out how to bundle our offerings without friction. We won't get it. So we have Windows now. We just launched a eSIM product. We are acquiring, and almost at the end of acquiring a second phone number. Great products. "How do we bundle? What kind of message do we do? How do I reduce friction in terms of people using the other products, too? Who do I target? How do I target?"

So I think there's a lot of things and a lot of experiments we will run. We won't get it right the first time, 100%, but I think there's definitely a big cohort of users who want to use these other products. We have to be able to get the answers right there. So it'll be easier after we fail a few times.

David Barnard:

All right. Well, as we wrap up, anything you wanted to share with our audience, any job postings you wanted to shout out or anything our audience can do for you?

Tanuj Chatterjee:

There are a lot of job postings. All of them are in Asia or Europe, so.

David Barnard:

We have a big audience around the world, so yeah.

Tanuj Chatterjee:

Because our product and engineering teams are mainly in Asia and Europe, so that's why we wanted there. We are looking for an eSIM product manager. We are looking for a growth product manager, growth manager, or a director. eSIM product manager, ideally somebody who worked in the eSIM space.

We are always looking for engineers and QA people. There are at least 15 job postings. I don't remember all of them right now, but if you are excited about this, if you're in Asia or Europe, please send an email to jobs@superunlimited.com, and I promise you, I will personally take a look.

David Barnard:

Awesome. Thanks so much for joining us, Tanuj. This was a really fun conversation. A lot of great insights.

Tanuj Chatterjee:

Nice to talk to you. Thanks, Jacob. Take care, David.

David Barnard:

Thanks so much for listening. If you have a minute, please leave a review in your favorite podcast player. You can also stop by chat.subclub.com to join our private community.