Key Lessons From Developing Roller Splat – Voodoo & Neon Play

In this post, we’re looking at Roller Splat from Neon Play, and why it works so well (now achieving over 49 million downloads worldwide). Before we do that, we first want to say a big congratulations to the whole Neon Play team. They’ve published a few games with us at Voodoo now and we’re so proud of them – they do a fantastic job.

We think Roller Splat is really interesting. That’s because it reflects the kind of studio Neon Play is and its mindset, which is why we’ve decided to focus on it today.

The secret to Neon Play’s success

Obviously the market for games like Roller Splat is a difficult one. It’s very hard to produce a single prototype, churn it out and have it succeed. But Neon Play succeed because they keep prototyping until it works. If something sticks, they carry on improving it, and if it doesn’t, they keep prototyping.

Neon Play also have a great culture of coming up with new ideas. Like the hackathons they organize once a month. These tend to last a couple of days, and their developers and artists get together to come up with new game ideas and then design them. They then go on to create prototypes for the ideas that stick, killing any that don’t.

An introduction to Roller Splat

The first thing we want to say about Roller Splat is that it’s not the most complicated game. It looks very simple and its purpose is easy to understand – you have to fully cover a level with paint. But it also includes all the basic ingredients we ask our developers and studios to include in their ideas and games.

Roller Splat is fairly original compared to some of our other games, in that it’s a puzzle game. If you take out things like Fight List or Twenty48 Solitaire, these are actually quite rare for us. But it’s a hyper-casual game which we specialize in. And actually, when you first play Roller Splat, it doesn’t feel like a puzzle game – it feels like you’re messing around with paint, having an impact on the environment.

It’s very satisfying and you immediately get the sense that you need to complete it, even if that’s not what you had in mind initially.

As mentioned Roller Splat is a very minimalist game. It’s not an .io game, you don’t have thousands of balls on the screen, there’s no fluff or complex designs – it’s very simple, and you get it immediately. You can work out how to play it within a few seconds. It’s also a very feminine game. That’s nothing to do with the pink paint – puzzle games like this are generally more successful with women.

How we got there

Let’s take a look at how Roller Splat evolved over time.

From left to right: MVP (minimum viable product), Version 2, and Version 3

Version 1

The image to the left is the MVP (minimum viable product – the first version of a game to give us feedback for future development) that Neon Play had in December 2018 , which came out of a hackathon.

This first version had a roller avatar and the angle of the camera was a bit closer and lower than it ended up in the final version. That meant that at times you had this kind of hidden avatar – it was harder to say what the aim was and what your avatar was doing. There was just a bit of a lack of clarity. Having said that, this prototype was actually pretty polished and tested well. But the cost per install (CPI) was quite high and day 1 was quite low, we think because of this lack of clarity.

Neon Play decided to test the game properly to see what worked and what didn’t. We then made small changes which were basically changing the camera angle and eliminating latency from the controls. We also changed the avatar from a roller to a ball – we felt that the ball represented speed and power more effectively than a paint roller. And even though to some it tells the story a little bit less effectively, we felt it really improved the physics of the game – when the ball lands against the wall, you feel like it really squishes against it, and you really feel powerful when you’re manipulating it.

Another important change is that we sped up the ball. So when you’re swiping up, down, left or right, the ball moves almost instantaneously. Whereas in previous versions the reaction time was a bit longer. Shortening that added a lot of satisfaction and reward in the gameplay. It’s really this interaction of the users with the game that was the big difference between version 1 of the MVP and the final launch version.

Version 2

Once we had a better CPI, we focused on things that were easy to improve for the second iteration. That was the controls latency, the feeling of power which came with a faster avatar and using the ball instead of a roller.

And finally we looked at user onboarding – so we made sure the first levels were short, snappy, and easy to complete. This quickly pulls players into the next section of the game which might involve more puzzle and deeper gameplay.

Version 3

The final iteration before the launch involved really building out the game. So we:

• created more maps,
• tweaked the balancing so it went from easy to difficult then back to easy occasionally,
• increased the depth and challenge of the hard levels, and
• really focused on the puzzle aspect once users were onboarded.

What we learned from developing Roller Splat

First of all, the order of the iterations is important. Working on the CPI, then the UX and finally moving on to building out the depth and long-term retention is a great way of shortening each sprint, without just having a road map that includes everything in the game at once.

Missing one of these steps could mean you end up with a game that stays just below the radar in terms of the metrics. But if you focus on quick fixes with high impact, and you do them in the right order, you’ll find that your metrics improve much more quickly.

We’ve put this into a graph for you:

The idea is that whenever you’re improving and iterating a prototype because you think it has potential – so let’s say it has a day 1 of 40 per cent and a day 7 of 10 per cent, and maybe a CPI below $0.50 – a really important thing is that for each sprint you start with high impact, low effort tweaks. So in the case of Roller Splat, it was the visuals, the basic controls, and the movement of the avatar which made the difference.

Let’s say you see the metrics improve after a particular tweak which maybe took you two or three days. You then move on to the next one, which might involve deeper changes in the gameplay. And then finally, you move on to a third iteration which, in this case, was more maps and more depth.

You can do it too

All the studios and partners we work with have more than enough ability to create games like this. Neon Play are outstanding because they’re able to do it over and over again. But that doesn’t mean they’re the only ones.

Here are our top tips for getting it right the first time:

• hyper-casual games should be simple,
• simple ideas win because they’re easy to communicate, and
• for prototypes with high potential, short sprints in the right order will get your games out and published much faster.

We’ve actually shared the 3 rules to keeping hyper-casual games snackable. You can read it here.

And finally…

Once again, big congrats to Neon Play. Roller Splat’s a fantastic game that reached over 10 million downloads globally in only a couple of weeks.

We actually spoke at GameAnalytics’ recent event, Making Hit Casual Games, where we discussed just exactly what is our process to publishing hyper-casual games. Feel free to watch here (you can get a copy of our presentation and learn more about the event here, too).