Postmortem: Reakt, our first game
At Rock Milk, a lot of what we know, we learned by following other game developers working in the open. That is why, sharing our learnings, is one of Rock Milk’s most important values. So, Reakt’s post mortem was a must.
Writing it proved harder than I expected though. Maybe because it is the first I ever wrote publicly or because it was the first Rock Milk game and ideas are still all up in the air. Anyway, it’s done and I wanna start by referring some resources that helped me put it together.
Our goals
Last week I had a conversation with a Mexican olympic qualified runner (hope he gets a medal soon) and he said a phrase that stuck with me as it embodies a lot of Rock Milk values:
There are three types of people: those that don’t know, those that think they know and those that are willing to learn
I like to think we are the later.
We were about to start a new business, and even though we are game developers for a while now, there were plenty unknowns. Can we run Rock Milk without impacting our dayjobs? Will it be fun or stressful? Do we have all game developing skills in house? How good is Godot as an engine? Do we have the skill/will to self publish? How the hell do you run a digital marketing campaign? What about social marketing?
It happens that, to answer most of these questions, we needed launched games. That made Reakt main goal very clear: launch as fast as possible.
That led to the principles behind it:
- It had to be our version of a simple proven design
- It had to enable us to test as many features as we possible could
- It had to be a real product, with everything from gameplay to monetization.
- It had to be ready fast and time box was our friend here. We ended up going for 2 months development cycle.
What went right?
We can develop games! (Constant communication is key)
Diego worked for 1h a day early in the morning and I worked mostly after putting my baby in bed. Weekends were in the mix as well. Weekly 30 minutes meeting and daily communication kept us moving forward, specially during periods only one had time to work. Trello was the organization tool, mostly due to familiarity.
It was also very interesting to see how complementary our skills are even though we are both computer scientists. Diego made all the art assets and animations. I took publishing tasks. He made the build and 3rd party integration while I coded most of the gameplay logic.
Godot is an awesome engine
Godot is a really awesome product of the open source community. It is simple to use, powerful and brings interesting solutions for common game development tasks. Highlights for the clever scene hierarchy, powerful animation system, one click mobile development build. Other cool feature is GDScript, of which I was really skeptical in beginning. Not a big fan of proprietary languages.
❤ Firebase ❤
From the get go we knew we needed Analytics and Ads in the game to call it a proper product. Firebase made it easier by providing it all into one package. We know we might be able to do better by using other ads systems, but the lower development costs totally skewed the equation toward Google’s tool. Extra kudos for the Analytics solution that automatic calculates more metrics than I could ask for.
Digital Marketing reach is crazy
We were genuinely impressed by the power and reach that Digital Ads have nowadays. We are about to reach 400 downloads in 45 days and half of those came from a low budget Adwords campaign. It is not much, but it give us enough players so we can keep testing new features in a data driven manner. We even made some Ad revenue.
Planning to try out Facebook Ads soon.
Random fact: Most of our Adwords players are from India. Real players entering the game every couple of days for a while now.
What went wrong?
Why do all projects have delays?
When we started the plan was to develop the game in 2 months and have it launched by the end of June. In the end, we published to Google Play on July 18th, which seems like a small delay but is over 25% delay.
Most of it was due to the way we designed Rock Milk. I had a couple of weeks I could not put in one hour. The same happened to Diego. We also failed to proper prioritize some must-have features focusing on the fun unimportant ones. Common mistake.
Godot have flaws (as expected)
Godot is great for gameplay code, but when it comes to building business features (Analytics, Ads, etc..) it is not straight forward as you have to code outside of the engine. We spent a good amount of the development time to learn how and integrate Firebase properly (finishing up strong with a push request to this module). We are still holding out the launch of iOS version for this reason.
The good news is that Godot 3.0 is in development and it has a feature that totally solves this issue.
Social Marketing work is real!
This is THE skill we are lacking the most at Rock Milk. We created Facebook and Twitter pages. Created some content and even worked with a creator make an Youtube gameplay video. But, we had little to no clue of what we were doing in the process. Looking for good resources to improve on this.
Final Thoughts
Overall, Reakt was a very successful first project. We found ways to launch it in a reasonable time and generate a player base to enable proper product development and new features tests while answering many of the questions we had prior.
Now we have a new problem, balancing our second game development and Reakt further improvements.
Fun problem to have though :]
If you read so far, thanks very much! If you want to discuss further, please reach out on Facebook and/or Twitter.
Matheus Almeida, best half of Rock Milk
