Amazon Lumberyard: The next big games engine?

Amazon has been dabbling in the world of games for a while, but launching your own engine is a huge step up. Why was now the right time to launch Lumberyard?

As with anything at Amazon, it started with conversations with customers.

We have a lot of developers as customers with Amazon Web Services, Twitch, our devices, our retail platform… so we’ve had ongoing relationships with them for a long time.

When we talked to them, what we heard was a consistent theme, which was that they wanted to spend more of their time and energy on creativity and differentiating aspects of their work, and less time on the heavy lifting and undifferentiated aspects.

As we looked at that and talked to them, they really wanted a robust commercial powerful toolset for making games. They wanted to be able to make games that connected to the cloud, but without needing to go and hire a bunch of specialised engineers to design, build and deploy backend game technology.

Then, of course, any artist – including developers – looks to build a fanbase and build relationships with those fans. We view Twitch as being a great opportunity to help developers with that.

Those are conversations that have been happening for years – this isn’t something you just wake up and do really quickly. It takes a lot of people, time and energy. We’ve been at it for a while. In terms of timing, we had some closed beta customers and we obviously hire a lot of new people – as those people came in we were trying to understand when it felt like it was a good time to launch a product. Based on the feedback we were were getting, it was a ‘go’.

What are your priorities during Lumberyard’s early days?

For us, it’s about engaging with the development community and with customers to understand what they like, what they would like to see added, and which areas of interest we can use to inform our roadmap.

It’s really been about how we’ve been thinking about our post-launch activities. Game engines are complex beasts, which is why we like the business so much – we think there’s a tremendous opportunity to help developers by allowing them to spend more of their time on making and less on building and maintaining engines.

In that vein, you don’t just pick an engine on the fly, you do research on it and think about it, so getting that feedback has really been our focus given that we’re just getting started. 

Lumberyard is based on Crytek’s CryEngine. Why did you decide to use that as the foundation?

While the basis of the technology is CryEngine, we also have a lot of technology from AWS, Twitch and Double Helix – we purchased the studio a while back and they have a lot of technology they use to make games like Killer Instinct and Strider. The fusion of those three really comes down to what it is that you’re trying to help developers do.

CryEngine is known for its visuals and performance, and those things were very appealing about it. It has also has hundreds of features that developers can use – or not – at their choosing. We really liked that. If a game developer doesn’t have to build a feature, that’s a good thing.

"If a game developer doesn’t have to build a feature, that’s a good thing."

Mike Frazzini, Amazon Games

What differentiates Lumberyard from rivals such as Unity and Unreal?

Firstly, deep integration of AWS and Twitch. Those are very material; when you talk to developers about how hard it is for them to build connected games backends, it’s quite challenging.

Then, if you talk to developers about how hard it is to build and engage with a community of fans, reach is one of the central challenges game developers face. Lumberyard’s online connectivity helps that, because you end up building a bunch of community-centric features in your game that allow for a lot of virality.

Games is a highly competitive business, and to be able to offer such a toolset for free matters a lot to developers. The notion of a robust, commercial-grade toolset that’s inexpensive was something that was very clearly something that developers sought after.

The full source is critical because the most successful games will deeply customise their experience for exactly the type of thing that they want to create. Those deep customisations require source code access. That’s a big aspect of our launch.

Lumberyard is billed as a ‘triple-A’ engine. What exactly does that mean?

We’re talking to the biggest games companies, but also lots and lots of indies, and we’re very happy with the initial reception that we’ve received.

Helping them spend more time on creativity, connect their games to the cloud so that they can create community-driven experiences, help them reach an audience on Twitch, provide a growing powerful set of tools to make beautiful and performant gameplay – those are the central aspects of what we’re trying to achieve.

The response has been very positive, both from indies and very large companies. We’ve even been surprised by some of the game projects we’ve heard – although you should never be surprised in games, because it’s inherently creative.

There’s a long conversation where those companies will do due diligence to understand whether or not they want to commit to the technology.

Is Lumberyard focused on digital games only, or do you expect it to produce full retail products?

I see no reason why it wouldn’t be both. That’s ultimately up to the developers – they determine the platforms they want to ship on and how they want to ship. Whatever they want, and wherever their interest and demand lies, we’re going to work hard to support.

Lumberyard is monetised through the use of AWS. How does that position the engine in the market?

It comes down to what we see in the industry today, as well as where we think it will continue to expand, which is this concept of community-driven games, where titles that build and garner large communities of online fans tend to enjoy larger fanbases, more minutes of engagement and stronger retention – all the things that you want to create with a game. 

In terms of free versus not free, there’s gradients – everyone has a business model and we like ours. It was informed by conversations by customers, but the idea of giving all the source code away to developers and getting the full toolset with the download is a critical aspect.

If you’re going to connect your game to the cloud, Amazon would be a natural choice. You don’t pay anything extra – it’s all the standard rates when you use AWS.

We really felt like there was an opportunity there where you’re getting a lot of technology for free and when we make money when you connect to AWS, it’s in the context of your success – you’re only paying when your game scales to lots and lots of users, and paying for services that you’d be paying for anyway. 

Included in the Twitch integration is support for ChatPlay and JoinIn. Why should more devs consider utilising such features?

We’re not trying to convince developers – it’s more like inspiration.

A critical part of developers’ lives is spent thinking about how they build a community of fans and then how they engage with that community. That’s a very hard thing. We read a lot in the development community about acquiring customers, lifetime value and all those concepts which are very important. Obviously, the games that are the best at going out and garnering active and engaged fanbases are fully predicated upon the content itself. 

Twitch is this extremely large and active community of fans that love games. To have this notion of a player, broadcaster and viewer all sharing an experience in a synchronous fashion is really interesting.

The idea is: how do we create tools that bridge that experience so that a player, viewer and broadcaster can all participate in some meaningful way? It’s an area that we believe is ripe for continued invention, so our thought is that we provide tools to the community and, ultimately, the community shows us what can be done with the tools.

"We’re not trying to convince developers – it’s more like inspiration."

Mike Frazzini, Amazon Games

You have used the word ‘community’ a lot. Outside of Lumberyard, how is Amazon planning to support devs in building their audience?

Twitch has an SDK and APIs, and the platform itself is a representation of a bunch of tools for developers to either broadcast or engage with fans.

Another programme that is very interesting is something we launched last September called Merch by Amazon. The idea was, if you have a creative IP and you want to sell physical items, we allow developers to do some designs but we do all the manufacturing, customer returns and processing payments of T-shirts.

We’ve seen very positive results since that launched. As a consumer, when you can bring your passion and your gaming lifestyle into the physical world, that’s a big deal. We think that the physical manifestation of lifestyle gaming and having the community manifest itself in physical items is an important thing, as well. We have a host of things along these lines, with more to come – we’re just getting started.

What’s next for Lumberyard?

We’ve already announced some things that are coming soon, such as support for mobile and VR.

One of the most common requests we’ve hear is a way to import different asset types, so we’ve built an entirely new asset importer that supports FBX.

Beyond that, there’s a lot of engagement happening with our customers on our forums and social media, as well as on a one-to-one basis, that is informing our future roadmap.

The way we think about it is: obsess over customers and really listen to what they want and what they see as gaps. Sometimes they’re saying things that require us to make little extrapolations and build our roadmap around those things. With just a couple of weeks of feedback we’re already getting a ton of great stuff and reshaping what we already had in our plans.