How to choose a category to launch an app

Listen to the episode

After working as a UA manager and a product manager for a few years, Slava Kononeko gained significant experience in launching and marketing apps. Now he is a co-founder of Onelight Apps, a company that developed several apps in a Photo&Video Category with more than 30m downloads in total.

In the 12th episode of our SubHub Podcast, Slava told us how he came into the mobile app business, what UA looked like back in 2014, and how to launch an app MVP.

Read the most interesting quotes below (translated automatically with DeepL)

Intro

Nikita Maidanov: Hi everyone, and welcome to the SubHub podcast! The podcast is supported by Adapty service, which allows you to quickly and easily integrate your app subscriptions and run A/B tests on them. Today we’re on our 12th episode. And our guest is Slava Kononenko, cofounder of Onelight Apps. And, in general, a man with eight years of experience in mobile marketing and product management, who has a lot to say about how mobile applications work. Very interesting experience, very cool guest – let’s get started!

Vitaly Davydov: Slava, hi!

Slava Kononenko: Hi, Vitaly!

Vitaliy Davydov: You really have a very interesting background. When I studied your biography, I was quite impressed. You worked in top companies which deal with cool applications, have a lot of volume. It’s interesting to ask where it all started in the first place. Tell me about how you went from a pretty strong technical background to procurement, to marketing… What was it like in general? What did you study?

Slava Kononenko: It all happened quite organically. I graduated from BSU, the Faculty of Mechanics and Mathematics. Research and production – the Department of Algebra and Information Protection. Globally, when choosing a university – if we’re really digging deep – I chose from what I was interested in. I was interested in mathematics. In Minsk, we do not have such a great choice of universities: only three. I chose the one that was stupidly closer to home! As for the department itself – it was pretty boring to study a lot there. But, since it’s not a bad school, at some point it became clear to me that only the technical part I lacked. That is, it was interesting for me to be fully involved in the process, to invent something. Purely to develop the code I was not very interesting. 

I had a pet project, a dancing school, which had to attract traffic in some way. At that time I was beginning to make HTML sites, buy advertising in AdWords, VK. There was no point in buying from Facebook at that time, because only businesses or older people in Belarus probably used it. This is why my first acquaintance with traffic was when I had to buy advertising and promote something. That is, it was a local project, but the first acquaintance was there.

I was lucky, I guess, in a way: our department did a test course on Internet marketing. They called it Digital Marketing for some reason (apparently, it sounded more effective and salesy) with Artox Media.

There was another course on SEO StartUp Labs. This was a company that did, as I recall, data collection for financial scoring. That is they were reselling user data to banks, lending institutions, etc. This was just another course, but it was purely for development: EPAM Labs in C#. These three courses – they moved in parallel – from the third or fourth year I began to take an interest in this.

Apalon

Vitaly Davydov: Tell us briefly about Apalon. What does it do in general? And what was it like when you wanted to get there?

Slava Kononenko: At that time it already had a fairly strong portfolio. The company was making various essential apps, i.e. utilities. The concept was to make products that everyone needed every day. That is, for the widest possible audience, without any targeting or adaptation. We need to look at the weather forecast every day plus or minus in any country. People use a calculator, they want to install wallpaper on their phone… The point was to make these applications as simple as possible, but necessary every day.

Vitaly Davydov: How many of them were there?

Slava Kononenko: I do not remember exactly. The portfolio was already quite large at that time. At the very least, there was Wallpapers, Weather Live before it became mainstream, by the way. NOAA Radar was a weather app mostly for the US market. Pimp Your Screen, a calculator, and there were some other utilities but I don’t remember anymore. There were about ten apps for sure.

Vitaly Davydov: And how much did the company earn at that time?

Slava Kononenko: I can’t say!

Vitaly Davydov: Well, ok. But obviously it was more than $1 million!

Slava Kononenko: I think so.

Vitaliy Davydov: And you went there – what did you do? What position did you take?

Slava Kononenko: I came in 100th, i.e. I was already 100th person in the company. And I was hired as user acquisition management, i.e. traffic acquisition. There was only one user acquisition manager at the time, so I was second.

Vitaly Davydov: One user acquisition manager for a hundred people!

Slava Kononenko: In fact, yes. There were some related positions that also could bring in traffic one way or another.

It was cross-promo manager, ASO/marketing managers… But there was only one person involved in direct purchase at the time. Accordingly, my task was to buy on social networks: Facebook in the first place, a little later Instagram, Twitter, Snapchat. It all started with Facebook.

Buying traffic in 2014

Vitaly Davydov: Can you share some of the first insights that you got, or some figures that you managed to achieve? It’s interesting to see this kind of retrospective: with what efforts what could have been achieved in those years? That was five years ago, right?

Slava Kononenko: It was 2014. In general, it was funny! Firstly, till that moment I didn’t work with products, which were represented in the world, where you could get traffic from any location, where you could get it with positive ROI. Of the insights, the cool thing was that at that time there was still a market, first of all, for premium apps, which has been pretty dead for a long time now. The second was how cool Facebook is as a network. Even back then, it was clear that it was unreal quality traffic with huge opportunities for targeting, for segmenting users. So it was really interesting. And the third thing was how cheaply we could buy traffic back then. I mean, in principle, we had CPI of 7-8-12 cents in the States. It is clear that we are not talking about all projects, not for all geos, not in any period. But globally traffic was still cheap.

Vitaliy Davydov: Was it for paid applications or for free, freemium?

Slava Kononenko: No, of course, there were no such prices for paid ones. It was mostly for the freemium. Although no, in fact, there were probably the cheapest CPIs on Wallpapers.

Vitaly Davydov: Well, if you have a paid app for ten cents, it’s certainly a fantastic channel of traffic!

Slava Kononenko: Globally, of course, ROI was not that huge. Respectively, the traffic volumes were not at all like the ones most companies buy now. Neither in terms of volume, nor in terms of cost of budgets per month. But on the whole it was probably not far from the beginning of a stronger user acquisition, as I think.

Vitaly Davydov: And tell me some figures on volume, for example, purchasing at that time. For example, now companies, who really earn, easily turn around hundreds of thousands of dollars a month, or even millions. How was it then, seven years ago? How much has the volume of purchases increased in comparison with those times?

Slava Kononenko: The volumes have grown considerably. The way major publishers buy now – it is really millions a month. And not necessarily millions – it may be dozens of millions a month. At that time there were no such figures, it was up to a million, 100%, depending on the period. It was quite a long time ago. In the initial period it was noticeably less, and then it was more. So I will not name specific figures anymore. But compared to how it is now, and how much more competitive the environment has become, completely different states of the market will be.same already – Instagram, Twitter, Snapchat. It all started with Facebook.

Application launches and niche analysis

Vitaly Davydov: Tell us then about the launch of new products, maybe. How did it happen? Based on what considerations, from what research could you understand what was better to launch?

Slava Kononenko: At the time, there was a certain algorithm that, in principle, worked. Now, in my opinion, it is not very applicable. But there was a pretty clear step-by-step process. There was an analysis of the market: what size, what trend – positive, negative – over the last 30, 90, 180, 360 days. To understand where the market was going within that niche. Is there a leader in that niche who has a monopoly. Because if there is a hard monopolist, it is quite expensive to take a certain percentage of the market, because the strategy of both Apalon and Easybrain has always been to produce products that will be leaders in their category. Because the leader takes most of the profit, most of the market. And if there’s already some hard monopolist in that market, that could be a problem, a blocker in deciding whether or not to make that product. An analysis of whether we can offer the market something of quality, that is, whether we can be somehow better than the players that are already in the market. And then there is this summary.

This, of course, is not all of the points that were in the works at the time, but the logic was something like this. It was necessary to find a big enough niche, which had a positive trend, where you can earn an interesting amount of money for the company… At the same time it was clear what we could do: there were certain competitive advantages… And here we go!

Vitaly Davydov: I understand that you can’t tell the whole framework how it was set up. But, for example, tell me what you need to look at, at least approximately. If I was running an application five years ago-obviously, it was easier to run an application five years ago than it is now. But still, maybe what tools did you use to do that? ASO tool, I don’t know…

Slava Kononenko: As for tools, I think that the market hasn’t changed much. Rather, only new tools have appeared. Sensor Tower was one of the pioneers. We were probably one of the first customers of that service. Now there’s also AppMagic. I guess there’s a whole other list of tools that allow you to analyze size, number of installations, revenue, if it’s not advertising revenue, roughly product metrics, what might be – retention for a session, etc., ASO positions…

I guess the way it used to be isn’t really relevant anymore. Rather, you can look at how it is now. That is, in your chat room there are often questions about how to procure traffic to a period tracker, or to a wallpaper, or to a scanner… And you have to understand clearly that in niches where there is a hyped auction, small players have basically nothing to do. In niches where there is venture capital that allows many to buy with zero or negative ROI, where it’s important to grow quickly, small players probably have nothing to do, again. For example, Calm, HeadSpace: Meditation is a very heated niche, heated with money. Accordingly, it is impossible to become a top-5 player there. Why the top 5? Here, as in any market, or as was once the case with web output, if you’re not on the first page in the results, you’re not going to get traffic if you’re talking about organic. If you can’t get in the top 5 for a key, you’re not going to get organic. Paid traffic – the economics don’t add up because of the expensive auction. We get zero in the output! Or, more likely, development costs, content and stuff like that.

So getting into obvious solutions that everyone can see is not the best strategy. You have to look for either new trends, which are growing, or at the junction of several niches, where there are no clear leaders, monopolies and big publishers…

For the reason that they probably aren’t big enough, and bigger publishers won’t make much money there. But for the small indie developer or publisher, it could still be interesting. And, obviously, since a lot of things now depend on paid traffic, if not all, it should be a niche where the traffic is either inexpensive or it is clear that the product will have an extremely high LTV, with a high retention. And then that would allow even expensive traffic to be bought. I would probably draw this algorithm.

Vitaly Davydov: These are, by the way, quite interesting thoughts. It turns out, according to you, for example, if I have 10-15-20 thousand dollars and I want to make an app… There is absolutely no point in launching a new meditation app. Right?

Slava Kononenko: In my opinion, yes. Because this niche is being developed by venture capital. Accordingly, there is a huge number of publishers that have a lot of money. And it will be extremely difficult to outbid them. The volume of the audience is finite, and those who have more money will simply take it.

Vitaliy Davydov: Do you believe that you can come up with some new cool mechanics that can beat the mechanics of Calm, or Headspace, or other projects?

Slava Kononenko: I don’t believe it without venture capital. Simply because your mechanics will be copied tomorrow, and there is nothing you can do about it. The speed of copying now is such that Chinese publishers release a Supercell game faster than the Supercell game comes out of soft launch. So the speed in the market is such that you just don’t have that time, when you find something that you’re going to do a lot better than the competition, then scale up the marketing, and the output will be good. That’s possible, but highly unlikely.

How Onelight Apps came to market

Vitaly Davydov: Let’s take your specific case with Onelight Apps and talk about how you made this company. On what grounds did you come up with the idea of making the products you are making now and how did you choose them?

Slava Kononenko: It should be noted that the idea of the first two applications was not mine, but that of my partner. That is, I did not make this company by myself. If to speak basically about directions, I have been interested in photo and video-apps since Apalon. One of the test assignments was to suggest an app that would fit the Apalon portfolio at that point, if you had all the resources for it. And just at that point, I was proposing to do a video editor. But you have to understand that this was before the heyday of the golden age of subscription, subscription modeling. Photo-apps have their own peculiarities. It’s a short session, low accumulation, for example, of ad impressions, not the best. At that time, such apps did not make much money. Because you can’t make much money from a paid app, you can’t show a lot of ads, and in-apps will not buy much from you. So the whole niche-it didn’t make as much money as it does now.

So, going back to Onelight Apps, at some point in Easybrain I was responsible for subscription projects, for music apps, for entertainment. And with products, I did the best I could. And I had already run out of ideas on how and what could be done significantly better. It was clear that the music apps weren’t even close to casual games. So there was a desire to try something new. And at that moment my current partner Yura showed me his first app: it was a scheduler for Instagram. And he shared with me the idea of our current flagship: it was FLTR presets. And it became clear to me that, first of all, I understand how this can be scaled. Secondly, I was already working with Yura on other projects within Easybrain. He was, at the time, a remote worker on the team I had found for publishing at Easybrain… Such a complicated chain! So I already knew who I was going to work with, I believed in his ideas, I saw how I could scale them up, and so it was a two-way match. And we started to do what we’re doing now.

And if you answer your question about why we started making these apps, it was exactly the same algorithm. There was a clear problem among users: there was a demand for this service or content. These niches were not big enough for big publishers. Or there were additional risks for them in the form of Instagram’s service or the need to use additional products like Lightroom, for example.

Therefore, the same Apalon or some other publisher, in principle, would unlikely go there. But for us, it was a good place to start. We had room for maneuvering, a lot of small indie teams with limited marketing, or which, in principle, do not buy traffic. That’s why we were able to take the lead in these small niches.

How was it possible to successfully launch an application in a competitive category

Vitaly Davydov: Do I understand correctly that the product itself allows you to create presets for Lightroom, for mobile, and this is a core value?

Slava Kononenko: Initially, at that time there were no apps with presets for Lightroom in the App Store and Play Market. In the beginning Appa didn’t allow to create something, it allowed to quickly download a ready preset, transfer it into Lightroom and quickly process a photo in two clicks. Now, of course, it’s a more comprehensive product. There are filters and new kinds of content that are not related to the presets themselves. But in the beginning it was a mobile client for quick search and download of the presets.

Vitaly Davydov: Look: the connection is really not obvious. You say that in order to use your product, the user must already have Lightroom installed.

Slava Kononenko: Yes. But Lightroom, I have to say that it is probably one of the coolest photo editors for mobile devices. That is, it has a huge user base. That said, Lightroom itself at the time – it didn’t have a ribbon – didn’t allow you to quickly find ready-made presets. They were mostly sold by bloggers. They sold them through Instagram, through the web, through some third-party solutions. And going to the App Store or Play Market, downloading an app and scrolling through the feed in it, there were no such solutions. And that’s what we did.

Vitaly Davydov: For me it is a little bit unclear. For me, photo and video industry is just a hell of an ocean! There’s so much there! Starting with the fact that there is VSCO, Lightroom and a bunch of other players… How do you find a new niche there? Is there some kind of methodical research process that will allow me to find something underutilized in a finite amount of time? Let’s say I’m a person who likes pictures. I think my idea might work as an app. How do I know that this thing will really work?

Slava Kononenko: I think there are several options here. First, it is a good understanding of the audience. As I said before, the ideas for these two apps were not mine, but my partner’s, who is into photography and photo processing. And he saw (there was a hypothesis, of course) that there was a need among users for such a service or for such content. That is, it was already evident. Users were interested and were already using other solutions. They were looking for presets on the web, buying them from bloggers. For example, they used Photoshop or Photos on iPhone to plan their Instagram feed. There were already some crutch solutions which people used one way or another. But there wasn’t a more convenient solution. We came up with a more convenient solution based on the fact that we already saw a need for it. There definitely has to be luck somewhere. But you can also watch for trends.

The last one, probably in Belarus, was Vochi, which made cool video effects. The technology itself was interesting there, not so much the app they had. But the trend that has been seen in recent years is that, first of all, everything is moving away from photos to video. Second, to content stories, which is Facebook’s purchase of Masquerade. Before that, Snapchat bought a Ukrainian company that worked with Masquerade. Now Pinterest bought Vochi. You can see a definite development vector to where the niche is going.

To answer your question about “over a finite period of time”: you can try to look at trends, i.e. those apps that appear in the App Store. Which grow organically, or not necessarily only organically, but where there are no other strong competitors. But at the same time it is clear that the project is developing, and it has a user base. And probably somewhere at this junction you can try to catch the market, catch the trend. Or we could look for new mechanics, like we did with coloring books. We are already moving away from the category of photos and videos. But originally there were ordinary coloring books, then there were coloring books by number, and then there were pixel-based coloring books. That is, the mechanics were changing, which made it possible, for example, to attract cheaper traffic or, on the contrary, in larger volumes. Somewhere it simplified the mechanics itself, and because of this there was a retention. That is, a consistent upgrade and development of one trend, which made it possible to form five different niches in the Play Market and App Store.

Is it possible for apps to survive on organic traffic

Vitaly Davydov: Do you think that Prila, even in a not very competitive niche, can survive on organics alone right now, or is that unlikely?

Slava Kononenko: You need to understand what you mean by the word “survive”. In principle, it can. If, for example, we understand that the team has a small burn rate… Or it’s just one developer and another designer who made the app, and they have an expense account maintenance fee per month and another $100 for something… Then, of course, such cases are possible, and there are a lot of them. I personally know a couple of teams that make good money, enough to live on. But the team there consists of a developer, a designer, and someone on outsourcing or part-time. To develop as a business on pure organics – hardly. On some, maybe, volume of n applications it is possible. But I don’t really believe in such a story, because organic traffic is finite, that is, it is difficult to manage. Algorithms change, competitors come out, so scaling organic traffic is difficult. Plus, the business still does not have the same margins as the indie team, the expense account is higher. Therefore, to grow, you need to grow faster, and this can only be done through scalable channels and paid traffic.

Paid traffic

Vitaly Davydov: If we talk about the paid traffic channels that we have now: Facebook, Instagram, TikTok, Snap… Plus in-app traffic on all sorts of advertising networks like AppLovin, etc. If we’re talking about apps specifically, which channels would you name the most interesting right now and why?

Slava Kononenko: I think that I would divide them into categories. Now TikTok not so long ago surpassed Google in the title of the most used Internet resource. So, TikTok is definitely a trend of the last few years, which will continue to grow and develop. Apple Search – due to the fact that Apple has really messed up the mobile market with its history with IDFA and SCAD. So it seems to me that this is one of those channels that will grow, and the competition is going to grow, because of the quality traffic, and there is information on users that Apple directly gives away. Facebook, which isn’t going anywhere. It has been slightly (or in some places not slightly) clipped its wings, but still Facebook, WhatsApp, Instagram own a significant share of the mobile market. So there are few people who can compete with them in terms of quality of traffic.

And if we’re talking about games, it’s a different set-up. Often networks such as AppLovin, Unity, RemSource can give no less quality traffic in large volumes than Facebook.

Vitaly Davydov: What do you think about YouTube?

Slava Kononenko: I think that YouTube is an interesting story for brands, for big companies in terms of using it as a content platform if it fits organically into their business direction… Or for collaborations with bloggers, mini, micro, any kind of influencers. So it’s an interesting enough story, but not for small teams.

How much money does it take to launch an MVP

Vitaly Davydov: I get it. The question that I think plagues absolutely every person who wants to launch an app is, “How much money do I need to spend to see if I can quit my job and make money with just this app? What should the cost of the experiment be?”

Slava Kononenko: There’s probably a different expense tag for everyone. Here, as in business, everyone has their own burn rate, and everyone needs different amounts of money in order to exist comfortably or uncomfortably. Plus, in my opinion, you don’t need a large amount. It’s such an interesting time now, and it used to be: you can test a lot of hypotheses with a limited budget to see if there’s any potential at all. So you don’t even have to quit your job or invest all the money you have to test one hypothesis. This is where Taleb’s anti-fragility approach comes close to me. You can allocate the amount of money that you have, or collect it over time to invest in the idea, test the hypothesis, and if it does not work, move on. That is to act in these iterations. To start the project, especially if there is expertise in development, 5 000 – 10 000 dollars may be enough to build a normal version of MVP or core functionality for the first version. That may be enough.

Plus, you have to realize that a limit is a good thing. When teams or people have unlimited resources, it usually all leads to building a spaceship that goes nowhere!

Vitaly Davydov: It’s true, yes! It’s the problem of venture money, it seems to me, including, initially.

Slava Kononenko: In many ways. I.e., limited resources are always a good thing. Because it allows either better prioritization or faster death of the projects that have no place to grow.

What metrics to watch for at the startup?

Vitaly Davydov: If you are building a project from scratch again, give me an example of a few metrics you would keep an eye on from the very beginning. Here’s the first one right here.

Slava Kononenko: I think that I would follow it anyway. This is a classic first, seventh, and thirtieth day. Plus, if we are talking about subscriptions, this is the renewal rate, i.e. this is the retention of extensions. The actual conversion to trial or payment if we are talking about subscription-based projects. And the cost of attraction is either CPI or the cost of attraction for an event, i.e. activation of trial or purchase of an application.

Vitaly Davydov: You listed about seven metrics. How much are they enough to keep track of the app? Do we need any other metrics, if we are talking in the abstract?

Slava Kononenko: Generally speaking, they are necessary, because the funnel is a little bit longer. It would be better to watch the entire funnel from the beginning to the end. If we speak about the beginning, it is CTR of creatives, conversion, conversion of a landing page… The way the user reaches the main screen, or at least the pewall screen, what happens to him next… This is a very subtle issue. It depends on the product: is it aimed only at earning money, or at increasing the number of active users? That’s why it’s a trick question.

How to choose a subscription price

Vitaly Davydov: The question is tricky, yes… By the way, if we talk about the purpose of the app. How can the purpose of the app and, for example, the subscriptions that the company sells be related? What recommendations would you have for choosing a pricing for the app?

Slava Kononenko: There are probably two questions put together here. Because everything can vary a lot depending on the purpose of the company. From the way the purchase is made to what monetization model is chosen, to what prices and durations of subscriptions. In the CIS, most people think of it as a classic business. “For so much we attracted a user, so much we got from him. The difference is our margin, that’s the profit.” In public companies, for example, earnings are based on share price. And they have other important metrics that allow them to buy with negative ROE, but still double or triple the audience in one year, and the value of the shares will quadruple. So the economics basically dance on other metrics. And that’s another argument I use when I say, for example, you don’t have to do the 105th meditation, because Calm and Headspace will eat you up.

When it comes to choosing the cost and length of subscriptions, there are, again, a few factors to go off of. If you have a cash gap, how quickly do you need to turn your money around? One-year subscriptions are actually sent by Apple in two months. Because you got the first payment – accordingly, the larger amount the user brings in has already been paid. With shorter subscriptions, the tail is longer, so it takes longer to get a refund. You can start with factors such as your budget and available resources, and play with the duration of subscriptions based on this. As for the cost – there’s no one who has the nerve. Still, we try to be fair to our users and not to set exorbitant prices. Moreover, we do not have weekly subscriptions, for example. This was done at the expense of LTV, but it was a strategic decision, in my opinion. Because it is important for us to build a more long-term relationship with users, rather than charging them a hundred dollars, getting one star, and that’s it.

Onlight Apps: Team, Plans, Goals

Vitaly Davydov: I also wanted to ask you a little more about Onelight: how many people do you have, what plans do you have next, etc.?

Slava Kononenko: We have a great team, 25 people. Cyprus, Minsk. Originally we’re from Minsk, but now we’ve got an office in Cyprus. Our team is small, as you can easily notice, if we compare with those who are commonly known. The Vochi ones, I think, are 50+. But we don’t grow the team fast on purpose. We spend quite a lot of time on recruiting, on selecting people. It is important for us that everyone is comfortable working with each other, that we stay fast, that everyone is interesting to work with. Therefore, every person who came to us was interviewed with at least one funder, and sometimes with two.

As for the plans for the future – we plan to grow within one vertical. That is, we do not want to switch between a huge number of niches, to blur the focus. We are focused on one vertical line. We will grow both horizontally and vertically. With those applications where we can improve and scale something up, we’ll keep working. Applications that do not show potential will be paused or have already gone. And we plan to release new applications. I think that at least one or two more products should definitely hit the Store next year.

Vitaly Davydo