WebGL matures into a commercial-ready technology
This article discusses how our WebGL version of Heroes of Paragon turned out to be a commercial success.
In April 2015 EVERYDAYiPLAY was confronted with a serious
problem. Chrome officially disabled the NPAPI plugin. NPAPI is actually a really old technology that made Java or Unity WebPlayer plug-ins possible. At the time, 70% of our users on Vikings Gone Wild were on Chrome and a large proportion of them used the Unity WebPlayer.
This was terrible news as our small indie company with 24 employees made half of its revenue with Vikings Gone Wild on Facebook.
The Open Web community has been working hard for the last couple years to provide a safer and more open alternative to Flash and NPAPI plugins.
However, our attempt to port Vikings Gone Wild at the time, a Unity base game, proved to be a failure.
The technology was not ready yet and we took the hit.
In mid-October 2015, my dear friend at Facebook, AJ Glasser, convinces us to explore a port again.
She introduces us to talented engineers at Facebook as well as the wonderful folks at Mozilla.
Marcin, our CTO tackled this subject one evening and sent me an email at 5am titled “I have a surprise!!”.
The surprise is an astonishing 30 FPS version of our game working fully on WebGL! It was not perfect yet, however, as we were facing one major issue. A massive and totally random memory leak would occur between 1 and 2 minutes of running the game, killing the browser.
But the step forward was huge. We forwarded the issue to Unity and Mozilla who investigated the bug and provided a fix, as well as a short term alternative to us.
Because of this positive progress, we decided to postpone Heroes of Paragon release — originally planned for early November on iOS and Android — to early December in order to include a web release.
You can find here a slide deck which I presented at DevGAMM in late December 2015 describing all the issue we faced and how we got ready for launch. Note that the KPIs in the DevGAMM deck were from the first few days of running the game.
December 3rd 2015, we officially launched the game on iOS, Android and Facebook with a WebGL export. We were not fortunate enough to
be featured in the mobile marketplaces. However, we already had a community, with Vikings Gone Wild players, perfectly fit for Heroes of Paragon and we launched the game mainly by cross-promoting from one game to another.
This article discusses how successful the WebGL launch was in comparison to the iOS and Android versions of the game.
First, our install statistics:
What is striking here is the amount of organic traffic we managed to get from our Facebook version
On about 55000 installs, more than 60% alone came from the Web.
Let’s have a look at retention:
It is not surprising to see the WebGL version at the bottom of that chart given the lack of support for older OSes and browsers.
However we can note that a 25% D1 retention for the technology is a very good score.
It is important to also notice that once past the first wave of cross-promoted, quality players, WebGL retention remained the most stable across all platforms.
Let’s have a look at DARPU:
A first striking fact is that on our first month, marked by a quality cross promotion, WebGL did better in monetization than iOS.
And through mid-February, supported only with pure organic traffic, WebGL remains a strong monetizer — significantly better than our Android version.
To conclude, given the current results, I am quite happy to report that WebGL has clearly matured to a commercial-ready technology.
There is work left to improve these numbers. A lot can happen with the release of the WebAssembly technology.
But the amount of work and optimization brought to WebGL is nothing short to amazing and finally provides
a solid solution to the incoming death of Flash.
I want to take advantage of this column to extend a very warm thanks to the team at Mozilla, Facebook and Unity who made this release possible.
CEO — EVERYDAYiPLAY