See also:
An introduction about me.
I'm a product manager in Google Ads based in the London office, actually based in the AdSense team in London. As Tom mentioned, I'm working in, primarily, ads monetization for games within the context of AdSense, but have also been working on a distribution product called GameSnacks, trying to think about what a distribution strategy for games at Google could look like. Previously I was a software engineer at Google and before that I was a software engineer a company called SGI working in file systems, and before that I did research in graphics at HCI, a little bit of graphics, moderate amount of HCI. And that was a very long time ago, so I was around for the very first wave of 3D graphics. I know what VRML is, I didn't use it obviously, I looked on it with contempt, but I did know what it was.
Certainly a vibe that I've been picking up from web games as I've been working in the space. One of the things I like about it is it actually reminds me of the 90s. Even some of the games look like the graphics from the 90s. It's just in the 90s, you had a $100,000 workstation to run them. Just for context, I have a little bit of a software background, but it's very much a sort of a Unixy kind of a system perspective. I'm not an expert on web technology. Forgive me for any cluelessness around web tech during this call. I'm very much a sort of API's and processes kind of a person.
Just in terms of... I'll try to make this quick. I just want to give a little bit of background on what we're doing at Google in web games. Talk a little bit about GameSnacks, our distribution product and our monetization product. I'm going to see if I can just touch on a little bit of the philosophy that's guided our monetization products in particular. And then I've just got a slide of various questions that may be relevant to this group and maybe we can discuss and we can just talk about it.
At Google, in terms of the products we have for web gaming, there's essentially four now. We have two monetization products. We've got an ads products in AdSense, which is what I work on. That's a very high level approach to putting ads in games. Ad Manager has ads products, so that's more of a low level tag based approach. And we've now actually got two distribution products at Google, GameSnacks and YouTube.
Both of these are relatively new efforts. Well, Game Snacks isn't that new, but it's certainly, we're quite small in this space. There's big beasts like WeChat and Facebook and folks like that who have quite successful and large gaming distribution products. We're still relatively early in the space.
Today I'm just going to talk about the AdSense monetization product and games
This group doesn't need to have HTML5 or Web games explained to them. The only reason I sort of call this out is there's a a lot of things that could be called a game in terms of web technologies and web games. What we're referring to explicitly is that subset, which is basically JavaScript talking to a canvas or a WebGL, or maybe WebASM or WebGPU. It's not so much the things with rich DOMs or more websites. These are really just apps that happen to be built on web technologies. They're opaque. They're like black boxes, just like regular apps.
Everybody knows this, but one of the cool things about them is portability. These apps run anywhere there's a web renderer.
It's not really correct to think of them as like web pages, in a traditional web sense. These are just apps. And they're apps that can be distributed anywhere and a whole bunch of people on this call are in the business of distributing these apps on various services, right? So you know exactly what I'm talking about.
Very cool for low-end devices. It's big in fast -growing markets. And if you look at the actual ecosystem, it's fragmented, it's complex. It's a post-appstore ecosystem. And there's a lot of scope for improvement in many ways, in terms of the way the ecosystem is structured.
Just in the space as we see it, we think there is a transition to a streaming model for games. In the same way that, there's sort of this broad trend of media types transitioning to a streaming model, music, video, we think that the same thing is happening with games. We're seeing a lot, especially in the last year or two, a lot of innovation competition in the app stores. And so you see like folks like Microsoft and so forth getting into the app store business and they're building their app stores on the web using web content and web apps for portability and also because of the policy landscape that the web sort of opens up to them.
The idea there is that they're outside of the traditional app stores they can reach Android and iOS and a whole bunch of other devices because they're web. And that's obviously a very interesting space for an app store to be in.
We're seeing just an explosion in distribution channels. YouTube, Netflix, we're seeing lots of big mobile OEMs starting to build their own game centers, trying to disintermediate the app stores so they're trying to build their own gaming experiences and capture those users directly.
And we're seeing things like Microsoft Edge starting to fold in streamable games, et cetera.
App stores is a big trend, new distribution, a big trend. I don't think it's overhyped at this point. I do think AI and authoring of content with AI is actually going to impact the creation of this kind of content. So I think that's sort of an accelerant on those trends. Like this is this move away from app stores to web -based distribution of games is happening and it's happening and it's accelerating. And I think if you look in places like China, you see they're probably maybe three to five years ahead of us and the sort of the revenues now of what they call like web instant games over there is really starting to pick up.
And in many ways that's a more advanced ecosystem than we have. Better support for distribution of across super apps, payments better business model, et cetera.
I don't know if that's going to work. This is just giving an example of this, right? This is Microsoft Edge, has an icon, you click it. I know you got a whole bunch of games. So it's just an example of bringing collections and streams of games to users wherever they are with less of sort of a centralized, distribution point in the form of the app stores.
GameSnacks is our effort to play in this space. At its core, the purpose of GameSnacks is just to expand distribution of this content type, ideally in an incremental way. We're trying to open up new channels, new audiences that might not be currently accessible, particularly on Google First Party. So this is on actual Google products where you can't currently distribute games. But also in places where it might be hard for smaller developers to open doors. Google can help. So we also do first part, sorry, third party deals with various distribution surfaces, such as, again, some of these OEMs and Android and things like that.
There's nothing particularly surprising or special about GameSnacks. Essentially, we've got a catalog just now. It's growing gradually. We have a mix of content that we've paid for and a bunch of content that we license and publish. We source content for a bunch of different people and then we do a rev share with them. So we're as monetized across all of those different surfaces and we give folks a rev share.
We've got all the standard parts of that sort of an ecosystem. We've got a developer console, SDKs, documentation, various products for third party distributors, APIs, et cetera.
That's kind of the picture. There's a bunch of people distributing games in the marketplace. We're playing there as well. We're all trying to like, create this network of distribution across the ecosystem. I guess most people get that any questions at this point.
Okay, let me keep going because I think the discussion will be more useful.
That's the distribution half of my world. So building out GameSnacks and trying to expand the network of places that we can put these H5 games, trying to be incremental, trying to grow the pie for the whole ecosystem.
I briefly want to talk about our monetization products and just touch a little bit on the philosophy of those. One of the nice things about GameSnacks, GameSnacks is still very small, but for me and the team, one of the nice things is that we get experience actually trying to run a game distribution business.
We, in a very real sense, understand the difficulty and the pain of doing that and it's challenging and it's hard, and we get a better insight into a lot of the sort of the problems that you face when you try to build these large catalogs of games. And that's useful for us because it does inform things like the ads products that we want to build.
A little bit of that philosophy has gone into our updated APIs. I'll just briefly talk about how we structure our monetization APIs for games. And then I think we can then get into the sort of the sets of issues and questions that we're facing that we'll talk about in the chat.
But we recently rebuilt, within a couple of years ago, we rebuilt API for placing ads within H5 games. Previously, we had a technology called the IMA SDK. It's very, very complicated, very powerful. You have to put a lot of code in your games to make it happen.
We went in completely the opposite direction of that, which is we tried to build something that put as little game logic, sorry, as little ads logic in your game as possible.
We're trying to avoid a world where games have to be baked into a particular ads product, have to be baked into a particular ads configuration. As an attempt at sort of a separation of concerns, which I'll get to in the next slide.
And just in practice, what it looks like is the game has very very little logic that it implements, basically it just triggers a few events, such as this is a location where you can show an ad. Very, very simple pause the game, resume the game, mute the game. And all of the actual logic around requesting ads and pulling them down and inserting them to the game happens elsewhere on the page. It just happens in a runtime that's loaded by the ads tag. It's not in the game.
This sort of gives you the kind of separation of concerns that we were trying to implement. Things that we were trying to achieve here, we want games... we understand that games have to be constantly rewritten depending on the ad system and where you want to distribute them and that's sort of this was our kind of very very first attempt to try and make that less painful for developers.
But obviously it's a tiny part of the problem it's not going to solve the problem. We wanted to make the sort of APIs much simpler so that, just from an ads perspective, it was easier to comply with policies, simpler for devs, simpler for distributors, decouple those two, and just generally make the process of managing these large catalogs of games easier.
That's all I'm going to say about that.
I'll just show you this just because it's relevant context for the sort of the discussion or the sort of points that we think are problematic. This just shows you just, it's a very, very simple API now in terms of what you see in the game. Extremely, like there's no logic in the game, it's just essentially a set of callbacks to sequence the behavior of ads within the game. All the ads logic exists outside of the game on the page. And by design, there's nothing here that ties you to Google, there's nothing here that ties you to AdSense, this API works across AdMob and AdSense.
The idea is that you can make these sorts of configuration changes in your advertising without having to go back and recode the games because often you're not going to be able to recode the games.
I'm going to move straight onto the questions because I probably lost you a little bit there, but I'll now try and contextualize that stuff a little bit.
And you can see this slide I wrote today because it's just a wall of text. I think what we're seeing - I can barely read that myself! - What we're seeing is like these H5 games are being distributed on a whole bunch of different surfaces, on different devices, different operating systems. It's very, very fragmented landscape. And we're seeing that this sort of streaming and syndication model is increasingly going to be big catalogs of games that kind of are pulled together on different surfaces, either with explicit distribution deals like Facebook Instant, MSN, GameSnacks, on the one hand, this is sort of the distribution side, or with some future versions of search where you can actually go out and search for these games. And again, you're bringing together search results and big collections of games.
In general, there's a sort of as a rule with this content type, there's always gonna be multiple entities involved. Somebody provides the game. Somebody's distributing the game. Both parties care about like optimizing the performance of that game. Both parties care about getting paid for that game. You end up with multiple stakeholders sort of concerned about the distribution of games.
And W3C... The core issue in all of this is just the lack of standardization across this content format. There's no app store entity that's forcing standards across these games. What you see is just this fragmentation in terms of distribution, in terms of the APIs, in terms of all the work that distributors and game developers have to do to plug their games together. And it's extremely painful.
At Google, we've already N+1 the standards. we have in our attempts to make the world better, I think we've added at least two more standards. But it's just the nature of the beast. Like GameSnacks has an SDK. Our good friends at YouTube have have a different SDK. We've already N+1 the problem twice.
Nut it is a systemic issue. Which is, unlike apps and app stores which have very powerful, entities like Google and Apple and etc. driving standardization and creating that hub. Where you code to a play standard, you code to an high-end standard and then like it fans out. We don't have that in the H5 space.
The consequences, every distribution platform has their own API. Every developer must release tests to maintain different versions of their games. Maybe they're building out APIs that are effectively drivers for all of these distributors. And it's just an M by N kind of complexity in this space. So that ecosystem is never going to grow easily if that's the kind of dynamic around content creation and sort of content distribution.
One of the questions we could think think about is, what would it take for distribution of discovery of games to be as easy as video? And I put quotes around easy because that's not easy at all. But it is actually easier than games in some key ways.
Video is linear. We know how to crawl it. We know how to crawl all of it. We know how to see all of a video. Like it's probably a halting problem or something to do the same thing for a game.
It standardized, it is transcodable. You can turn it into a smaller or larger format, etc. And it's editable. So you can split it apart and you can put ads and things in it. And you can do, and there's a lot of standards around doing that.
All of those things are actually quite hard for games and don't necessarily have a solution. So that's one of the questions we might ponder.
And then there's the second, some other things to think about, which is what I was touching on when I was talking about the design kind of our APIs and some of the problems. Which is, how do we untangle sort of the concerns of these different stakeholders around this content type?
You've got a game, why can't I just take your game and distribute it? Why can't I just inject ads into it without you having to worry about it? Why can't I set up a different rev share on this surface versus another surface? These are distribution concerns, they're not game developer concerns. Game developers just want to be able to put their content out there, get a good monetary return on it, support their business model. They don't want to have to go through all of these deep integrations. In the same way, like a YouTube content creator, it doesn't have to worry about the details of transcoding and ad insertion for their videos, they're just concerned about making content.
That's another question that we think about, which is how can we disentangle these concerns?
It may be impossible, and I mean, part of the feedback and discussion here might be just to say this is impossible. But maybe there is value and focusing on sort of key, the key business use cases. At the end of the day, different platforms will have different features, and there's probably some amount of customization that may be required.
But in those key areas where you've got money flowing around and stuff that impacts businesses. That's sort of the way that we're thinking is how do you make these things a bit more standardized?
And not standardized by just using Google Ads. I mean, that's a standard, but that's not what we're implying. How can you take a game and just inject Facebook ads into it or Google Ads, or whatever ads that you want to put into that game without sort of screwing with the game.
Same thing with payments. The point of these games is you can distribute them on different platforms. They can be on Android, they can be iOS, they can be all over the place. But how do you solve for payments and how do you solve for rev share of payments? Because again, there's usually multiple stakeholders around the money, how do you split the money up in a way that's scalable?
Analytics is another example. Different platforms have different analytics, developers and distributors care about the behavior and performance of games, people use different analytic systems. At the end of the day, the game should just be saying, hey, something happened and not worrying about the mechanics of the analytics system that's capturing it, for example.
As I say, rev share is just an aspect of this that exists right across the industry because it's sort of a post -app store ecosystem. And there's multiple stakeholders. And this is how the business model works.
How am I going for time? I'll be done in five minutes. So I'll just let me finish the next slide because maybe this just gives you the context and then then we can let Tom MC.
The other side of this, that's distribution and this is pulling together big catalogs of content typically where there's a publishing or distribution and a payments arrangement, but of course the other side, we're Google, is just discovery.
In some ways this could be an even bigger opportunity, Which is just, how do we make games on the open web more crawlable? As I say, how can we make the 20s look more like the 90s, where you've actually got a proper internet? I think, so this is just personal opinion, don't quote me on it, but the web is generally the open web. The internet root health, internet's great, it's taken over the world. But the web in general is not in a super happy place right now. Like the content ecosystem itself is probably suffering. And I think there's a lot of scope for it to be reinvented with content like gaming.
If you look at, say, Google's Play ecosystem, half of that organization is devoted to games. But is half of the web devoted to games? Is half of Google search devoted to games? Is half of Bing devoted to games?
You have this massive content distribution ecosystem, what remains of the web. It's sort of fallen on hard times, but why isn't this massive distribution ecosystem being used more for games?
I think there is a lot of opportunity around doing a better job of crawling and surfacing games. But obviously that is the core problem. Games are extremely difficult to crawl. We don't yet have solutions for this. Step one is something that Tom can talk about because this is all sort of inspired by Tom's thinking on this matter.
This is around maybe doing a better job of layering metadata on top of games. But what would be the next step? And I think we now are in a place where it's actually not inconceivable to say we're gonna have gigantic AI's that can interact with games and do a much better job of crawling them.
But the end state would just be, you go to your search engine, you go to these sorts of services and they can bring together games and they can do a good job of directing traffic out to people in the ecosystem.
Anyway, that's all I had to say. A little bit of background on GameSnacks, a little bit of background on our ads. Distribution is complicated and messy and discovery is hard.