10 Lessons From 10 Years Of Designing, Creating and Marketing Mobile Games

I launched my first Unity mobile game back in April 2013. At the time, Unity was starting to gain traction as the framework of choice for mobile game development. I happened to choose it because the developer I outsourced to suggested it.

Since then, I have self-published 10+ games; all created using Unity. In addition, I’ve gone on to work for a major mobile games publisher managing million-dollar ad budgets.

Here are ten lessons I've learned from designing, creating, and marketing mobile games.

A fair warning: it’s long.

But I promise if you read it all the way through, you’ll save a few years by learning from the mistakes I’ve made.

1. Be Inspired, Don't Copy

Prior to launching my first game, I was studying the App Store for months. I read books, watched YouTube videos, and generally blocked out everything else. I was a bit extreme and told my friends, “If it’s not about apps, I don’t want to hear it.”

When others zig, zag instead.

I was looking for an app idea that would be engaging but also can generate substantial revenue. I had a limited budget of ~$5k to break into the market.

I tried a ton of apps and games, looking for that.

Then it happened.

4 Pics 1 Word launched in February of 2013 and went straight to #1 in games. I knew this was the product I was looking for. Both game design and UI were insanely simple. And it had a very straightforward business model.

I had no problem just taking the business model as is. But I was a bit more concerned with launching an exact clone. I was new to the app store, so by the time I figured everything out and launched a game, the market would be saturated with clones, and mine wouldn’t stand out.

Instead, I applied the game design and business model to a more niche category. I created The Bible Trivia game. It was significantly better than the competitors out there. There were games that were fun but inaccurate. Then there were ones that were accurate but weren’t fun. Mine was accurate and fun. As a result, it stood out.

It launched on April 4, 2013; to my surprise, it made money on day 1. People were spending money on in-app purchases.

The game ended up doing pretty well, considering it was my first attempt. Several years later, I sold the game in its entirety, turning a 40x ROI.

When others zig, zag instead.

2. Everything Takes 3x As Long As You Think

This is true for anything related to software development. Plan for this or plan to fail.

That first game I created indeed took longer than I anticipated. I thought for sure I could have an idea to launch in one month, but it turned out to be closer to two.

I actually received a prototype of it in a week. So what caused the delay?

Well, first, I think I underestimated the complexity of creating a game. Even though I was a software programmer, I didn’t touch a single line of code in the game. My estimation, even by someone who’s technical, was a bit off.

Second, I thought of new feature ideas that came up during the development process. They may be scope creep, but I realized they make the product better. For me, if a feature makes a product better, I generally go ahead and implement it. That extended the original development time.

For example, instead of just having four images and having the player guess the word, I had different types of questions. I had audio questions- that was really different. I paid a British voice actor to read the clues. That extended the original development time, but it made the game experience so much better.

3. Making the Game is Easy. Getting Discovered is Hard

In mobile game development, I’ve found creating the game is much easier than marketing it. The real challenge is figuring out the discovery problem- who’s going to play it, and how will they find it?

Software development seems to be a binary problem- yes or no. Marketing, on the other hand, seems to be more trial and error. Maybe it will work, maybe it won’t.

In mobile, the main discovery problem comes down to “will the app even get installs?” Getting featured by Apple or Google is not a strategy, but rather hopium.

Larger studios usually have a war chest of funds ready to deploy to buy the installs. They do several soft launches of the game. Each soft launch can cost anywhere between $15-30k just to buy the installs. They use the soft launches as a way to gather metrics about the game. Then they keep improving those metrics until it’s ready for a worldwide launch.

Smaller studios usually have a growth hack strategy. This is what my friends and I did with Hexar.io, as I’ll talk about later.

Another way to solve getting discovered is by picking a different channel altogether. Maybe the game shouldn’t be on mobile, at least not initially.

I recently worked with a game company that had a mission of creating edifying games for kids. I believe and still do that it’s a very noble mission and one that I very much agree with. Their team was heavy on the engineering side but very light on the mobile marketing side. They insisted they needed to launch a mobile game because they wanted to make money via subscriptions.

However, they hadn’t thought about the discovery issue until six months ago. By this point, they’ve been in development for two and a half years. And only last week did they launch their app. I believe they’re facing a significant roadblock ahead in getting installs. Also, they have to deal with privacy issues in particular because their app is geared toward kids.

Now, what if instead of launching on mobile first, they launched on Roblox? Would that solve their discovery issue while staying true to their mission? I believe so. And as I have discovered recently, Roblox is a fantastic platform for driving organic players to games.

I learned that solving for the discovery problem oftentimes is about understanding the different distribution channels.

4. Phases To The Mobile Games Market

Over the years, I've observed several phases of the mobile games market. Each phase presented unique opportunities for new entrants into the industry. New studios rose if they positioned themselves well during these phases. And older studios died or became obscure as a result of not adapting to the changes.

If someone can spot these opportunities, they can take advantage of it.

Around 2015, I noticed a significant shift toward paid app installs. At that time, I was leveraging App Store Optimization (ASO), the SEO equivalent of the app store, to generate installs.

I did keyword research and optimized the title of the app to get as many impressions as possible. More impressions led to more downloads. But things were changing. I noticed ASO wasn’t working as well as it used to.

I also began to see companies buying installs as a way to launch their games. And I started to hear a lot more about CPI < LTV (Cost Per Install < Lifetime Value).

This became the dominant theme in the years to come. Data-driven companies like Machine Zone, Zynga, and Scopely capitalized on this shift and set up their operations to take advantage. They were very good at predicting the lifetime value of their players and knew exactly how much they were willing to pay for an install.

Around 2017, another trend caught my attention: the rise of hypercasual games. The company Voodoo rode this wave and emerged as the dominant hypercasual game publisher in the subsequent years. This trend was about creating simple, addictive, and easy-to-learn games that appealed to a broad audience. The game sessions typically lasted only a few minutes, even shorter than casual games. My friends and I rode this trend, as I’ll explain later.

Now, I'm seeing yet another trend influenced by privacy restrictions. Apple made it difficult for advertisers to track paid mobile app installs. As a result, influencer marketing is becoming more and more the primary way to bring in installs. This is not just happening in mobile but elsewhere as well.

In Roblox, the primary way for developers to drive players to their games is through influencer marketing. Some very astute studios, like Wonder Works, create games that make it easy for YouTubers to create videos for.

It’s almost like instead of marketing to the players, they market to the influencers, who then market to the players, who then go play the game. It’s a totally different model.

Here are the phases that I saw in mobile games. Time ranges are estimates.

1. Paid games (start of the app store - 2012)
2. Beginning of free-to-play (2012 - 2015)
3. Paid user acquisition (2015 - 2018)
4. Hypercasual (2018 - 2022)
5. Influencer marketing (2022 - present)

As I’ve learned, there are always new opportunities despite the saturated market. If someone can spot these opportunities, they can take advantage of it. This leads me to my next lesson.

5. Opportunities Have A Limited Window. Move Fast

I saw a window of opportunity in late 2016 where a specific sub-genre of games had the potential to generate organic installs easily. This, I thought, would solve the discovery problem I alluded to earlier.

I noticed YouTubers playing “.io games.” These were mainly browser games (at the time), had very simple graphics, was multiplayer, and was about becoming stronger the more players you defeated.

Something about YouTubers playing them felt really off to me. By this point in my career, I had enough of those feelings to dig deeper. What I found was the following.

At the time, YouTubers thought mobile games were less than AAA console games. They thought they were less than in fun and less than in quality. They didn’t think mobile games deserved their attention in creating videos.

But for someone reason, they would play .io games. I watched PewDiePie, the biggest YouTuber at the time, play a game called Diep.io.

And his audience loved it!

The audience wanted to see more of these games, and as a result, YouTubers sought out new .io games to play, creating videos that further fueled their audience's desire for more. This cycle of demand and supply led to an exponential increase in the popularity of these games.

Then I saw my friend launch a Diep.io clone on mobile that rocketed to the top 10 in games almost overnight, with zero paid acquisition.

Seeing this, my friends and I decided to create our own .io game called Hexar.io. Our strategy to get installs was simple. We would put a version of the game online and advertise it on various .io games websites. We believed YouTubers would eventually find it and make videos of them playing it.

And. They. Did.

We didn’t play a single dime to have YouTubers play them. Our thinking was that people would see these videos, then search for “hexar.io” on mobile.

As a result, Hexar.io did very well. The game has generated over 12M downloads across both iOS and Google.

As mentioned earlier, I believed there was a window of opportunity to launch this type of game. Even though we did pretty well, I think we were slightly late to the party.

Paper.io, a similar game, debuted just a few weeks before our release and garnered significant attention. I still wished we had been first.

Just a few weeks later, we saw many clones of both Paper.io as well as Hexario.

6. There Will Be Flops. Embrace it. Learn from it

Creating a hit game on my first attempt colored my perception a bit. It had me thinking I was good, not lucky.

Unfortunately, games are a hit-driven business, and luck is a huge factor.

And over the years, I had some huge flops.

The first was Matchy City back in 2015. This game holds a special place in my heart because it’s still the best game I’ve designed. I did tons of playtests and iterated from their feedback, so I knew the right players would love it.

However, the game failed to generate installs and monetize effectively. Looking back, Matchy City lacked levels, and I knew it. I was given this feedback, but by that point, it was way too late in the development process. I needed to launch the game because I was bleeding money.

Also, I didn’t have a strategy to acquire installs. The lesson I learned from this flop was that I would never launch another mobile game without solving the discovery problem.

Just a few years later, I learned that lesson with Hexar.io. #Winning.

Another flop was Bounc.io in 2019. There were a few things that went wrong. We ambitiously took on a project that required an entire year to develop. Maybe I hadn’t quite learned the lesson of estimating something would take 3x as long.

There were only four of us, and we essentially put all our resources into making a hit, or so we thought. That’s really tough because long development cycles can be an absolute killer for small studios. We did get a Google feature but that didn’t save us, which brings me to the second thing that went wrong.

I ran tests for this game using Facebook ads. I took early screenshots of the game and used them as ads. I measured the ad click-through rate to see what the potential response would be. The results were not stellar.

I did the same test with Hexar.io previously, and they were significantly better. This should have been a huge red flag, and the right move would have been to pivot even though we already put development time into it.

It felt like we wasted time. Instead, we should look at that as prototyping, a way to probe the market to see if there’s demand.

The lesson I learned from Bounc.io is to test early and often. There are ways to do it, and I’ll get to them later.

7. Keep Doing What Works, Until It Doesn't

When we had a successful .io game on the market, we should have continued creating more games in the same genre, utilizing the engine we had developed. The strategy was clear: keep riding the wave of demand for .io games until they subsided.

Stick to it even when it feels boring.

I underestimated the demand for .io games because even to this day, there is still a demand for them, although their appearance may have evolved. A prime example is Survivor.io, a recent game that has achieved notable success.

Additionally, games like Archero incorporate elements of the .io genre. Interestingly, the creators of Archero previously developed Arrow.io, further highlighting the opportunities for building upon existing success.

Had our team continued to create and release more .io games, we might have been able to cultivate a large and dedicated user base, generating substantial revenue along the way. This revenue could have provided us with the resources to venture into more ambitious projects, ultimately allowing us to create a larger and more expansive game.

The lesson learned for me is this: once I find a formula that works, I stick to it until it doesn’t. Stick to it even when it feels boring.

We didn’t and, as a result, had to close up shop. Looking back, I take a large part, if not all, of the responsibility for that decision.

8. Certain Game Genres Require Certain Implementations

I learned this lesson after the Matchy City flop. I was told that free-to-play casual puzzle games, particularly match-3, absolutely needed levels in order to monetize well. The exact words I remember were this: “These games are fun, but they don’t make money.”

The game that comes to mind regarding this is Triple Town. The game was incredibly enjoyable but struggled to generate significant revenue. It should have been a massive success, but it wasn’t.

The creators learned this lesson because their next game, Alphabears, did much better.

Oh, it had levels. Here's a postmortem of it.

Here is another implementation I can think of. Hypercasual games absolutely need rewarded videos. It’s a great way to monetize, not directly off the players but the advertisers. It’s also really simple to add them throughout the game.

Have a game that players play for hours on end but has terrible retention after day 1? It should probably be a paid game. Monument Valley is a paid game for this exact reason. Players generally finish the game within 1.5 to 2 hours and never play it again.

9. Test Game Concept Early, Test Often

One critical lesson that I’ve learned that many indie developers and small studios tend to overlook is the importance of testing the game concept early. I would say that most indies don’t even test the game concept until a prototype is created.

What I’m talking about is testing the concept of the game before the prototype.

I think developers don’t feel they’re making tangible progress because they’re not building. This is, in fact, the exact opposite. Testing for product-market fit is the most important thing to do. I get that it feels weird to test the product-market fit when we’re talking about game design, but it works.

Let's consider the example of Hexar.io. In the early stages of development, we created screenshots depicting what the game would look like and ran Facebook click ads that directed users to a landing page.

I analyzed metrics such as ad click-through rates (CTR) and waitlist submission rates. Facebook ad CTR on average, was 1%, and our numbers were close to 2-3%. The numbers were promising and indicated we were in the right direction. All it meant was we should continue developing.

When I ran the same test for Bounc.io, they were subpar. It was definitely less than what I expected. This should have been a signal for us to stop and figure out what the problem is.

Testing for product-market fit is the most important thing to do

Watch this video to learn how Wooga does “fake door testing” in mobile games.

10. Play Game Dev Tycoon

Instead of giving one last lesson, the one will give at least four more. I played the game again recently and found that it captured the challenge of being a game studio very accurately.

Here are my key takeaways:

Creating a Game Engine is Important: In Game Dev Tycoon, the significance of developing a game engine becomes apparent. Building a game engine allows for differentiation in the market and can significantly shorten the development time of new games by leveraging the existing engine.

Certain Game Combinations are Better than Others: The game highlights the concept of theme and genre combinations. It teaches us that certain combinations work better together, leading to greater success and player engagement.

Want to make a farming-themed game? Well, you’ll have a higher probability of success if it is in the simulation or casual genre (Farmville). But what if you want to create a farming RPG? Well… you might want to rethink that.

Bankruptcy Is Always Around The Corner: There are two modes to the game: easy and hardcore. I played hardcore mode and was always one flop away from bankruptcy. As I’ve found out IRL, it’s pretty true.

Stay Small and Build a Cash Reserve: To survive in hardcore mode, the player has to stay in the first level (the garage) for a very long time.

When the player’s in the garage and not leasing an office building, the costs are lower. Also, it’s just the founder. There aren’t other employees yet, which also cuts down on expenses.

During this time, the strategy is to create game after game and build up a large cash reserve before moving on to the next level. Don’t just go onto the next level just because it’s available. This is pretty sound advice.

So if you’re thinking about creating a game on any platform, I encourage you to play Game Dev Tycoon. And yes, play it on the difficult mode.

Wrap-Up

That’s it. You made it to the end. I hope you found it interesting and insightful.

If you enjoyed this article, tweet at me @jackzliu or connect with me on LinkedIn and let me know your thoughts!