Gamesparks and Playfab are ground zero for what will be one of the most interesting battlegrounds for cloud services. Though there are other options in LiveOps (HeroicLabs, ilogos, and Unity) these are limited to one platform or they are so small they won’t be able to compete with the 2 competitors backed by cloud providers. Google hasn’t got a dog in this fight. . . yet, but I would expect that to change in the months to come. They are looking at game streaming so they’ll probably realize this is another area they should get into and I’d expect them to snatch up one of the other cloud-based Live Ops companies to get there. Investigating these two LiveOps offerings and the cloud offerings of the companies can tell us a lot about the future of games and the role multilingual development will play in the global success of these companies.
So what is LiveOps? Think of live ops as a way to bolster game play for an established game or franchise. Games (mobile, console, or online) don’t become a success unless they retain players and create enough buzz to grow their user base. In the past this meant sequels and expansion packs, but now this all falls to LiveOps. And LiveOps is becoming one of the most complex and interesting areas within technology. Games generate more profits than movies, as much data as social networks, and handle as many real and virtual transactions as small fintech companies. They also need massive communication networks, robust analytics tooling for live streamed data, and storage for constantly updating content. On top of all of this they need to be able to handle the full game lifecycle from launch to deprecation. And most interesting for me they need to manage all of this across the world and stay in compliance with local laws while appealing to regional interests.
The bigger game companies have their own in-house teams and proprietary tools used for LiveOps. However Amazon, Microsoft and soon Google will enter LiveOps as a cloud service betting that smaller players will want to offload the tooling and storage necessary for LiveOps to a 3rd party provider that can address the server and database needs of the activities.
LiveOps encompasses leaderboards, in app currency, holiday skins, virtual purchases, matchmaking, chat, loot boxes, and many other components.
The LiveOps also affords opportunities to examine virtual currencies, micropayments, and freemium game models. All subjects I’d be interested in investigating at a later date. But for now my focus will be on how localization will be integrated into LiveOps cloud-based tooling.
LiveOps has compressed timelines since sales and changes are often unscheduled changes to the game necessitated by the need to increase playing time and user base. And LiveOps tooling will changing in the years to come to support game streaming needs as Stadia, XCloud, and whatever Amazon’s game streaming competitor will be called whenever it is announced.
LiveOps in large game companies
LiveOps is essential to large casual and MMORPG gaming companies. They have all built the infrastructure and tooling to run their own operations in-house. They may be loathe to give up those capabilities, but over time even the largest game companies may begin to view LiveOps tooling as less essential if the cloud offerings are secure and private.
Epic posted $3B in profit in 2018 largely due to its sales in-game facilitated by LiveOps. And even the silliest of casual games can extend the life of the game and realize serious profits if they keep players coming back with limited offers, holiday events and sales, and unique content or missions.
LiveOps in Social and MMORPG games
LiveOps in these 2 different types of games may have slightly different functionality, but the reason for the mechanisms are the same. Games need to create virtual communities, stay fresh, and continue to bring players back and bring new players to the game. To do this they not only depend on social aspects like chat, marketing, and friend-to-friend play, but they also depend on game theory, marketing physical products, and selling virtual currencies that can often be bought with fiat currencies. Each of these activities requires analytics, segmentation, targeted marketing, sales, finance mechanism and other infrastructure that needs to be developed, leveraged from open-source, or supported by cloud-based LiveOps tools. Each could be a topic of its own post, but for now lets accept the cost of these platform tools will increase exponentially as the complexity of LiveOps increases.
Localization is essential for LiveOps
As can be seen in the chart below localization will be essential to the cloud platform-as-a-service offering of LiveOps. Gamemakers and gamers are worldwide and speak every language. So not only will all of the LiveOps tools need to be usable in any locale, they will need to manage and support content for every locale. Microsoft and Amazon have started this long journey by supporting messages and other content in non-English languages, but the support is far from complete and there will be many missteps along the way that are common to any development team shifting to an internationalized code-base and supporting localized content. Though large gaming companies are very good at handling localization, I think we will see many issues as the two largest cloud players realize that localization is essential for the growth and sustenance of LiveOps as a platform service.
One area I’m watching is the documentation and training for their tooling. It will be an interesting challenge because both companies are balancing migration to corporate tooling, new feature development, documentation for complex features, training, and marketing to drive adoption.
Localization-as-a-service in LiveOps
Over time both Amazon and Microsoft will realize that localization-as-a-service will be essential to the growth of the LiveOps business. I can imagine a lot of different pay models and support mechanisms springing up to grow this service and sustain the demand for continuous localization. Most likely they will settle not on a per word cost, but rather a small percentage add-on based on the cost/time/quality triad that drives localized content production.
The models will require XLIFF and TMX support. Though they will not be full-fledged TMS products I can imagine shared termbases and translation memories for standard buttons, boards, and mechanisms within games. Or a terminology lingua franca used in video game design across locales. And of course besides the common localization standards they will also support CSV, JSON, and other export methods in case teams want to work offline or leverage the LiveOps tools to create their own GraphQL or run-time mechanisms.
Flow of content for localization
I imagine that the localization workflows adapted by LiveOps platforms will be supported with connectors for a myriad of TMS and MT products. The data can live within the LiveOps system CMS and there will be support for source-languages, pivot languages, and customized content , media, and images per locale.
Below I’ve drawn a workflow that depends on the TMS and MT systems to route content for TMs, TBs, and human and machine translations (statistical or neural). I’m sure the workflows will over time be much more complex, but the pricing will be built around MT vs. HT cost mechanisms. Customized engines, post-editing, and human translation will necessarily cost more than Microsoft MT or Amazon translate. And over time video game companies may see audio voiced by Polly and Cortana as cost and time saving mechanisms for prototyping and sim-shipping.
Mobile Networks, data centers, and partnerships
I grew up learning about landline phone networks, and cellular networks as they came online. Packet switching, CDMA, and GSM networks, and bandwidth sales were a part of my childhood. While other kids learned checkers I learned about the challenges of counting millions of dimes from telephone booths across the country. So I have an informed opinion about the role data centers, cellular networks and partnerships with carriers will play in the global expansion of LiveOps and game streaming and the success of Google, Amazon, and Microsoft. The race for game streaming boils down to latency, depth of catalogue, cost, and the realization that the medium changes the message.
My money is on Microsoft over Google and Amazon for a few reasons. First, they have a forward-thinking attitude towards their data center placement. They’ve invested heavier in India, China, Japan, Korea, and have a few data centers in Africa. And I suspect there will be more investment from Azure in South America and Africa at a faster rate than AWS specifically because of game streaming. They also have a long-history of partnering with games makers, running their own games studiso, and as the only competitor with a full-fledged console they are light years ahead of the competition. This gives them an edge in licensing, and international copyright and regulatory issues that will plague the industry as it expands worldwide. Their approach to roll outs is similar to what they’ve done for their legacy games. They are working to address the challenges of playing on a phone (they’ve got patents on add-ons to adapt the phone into a controller and play directly on the phone without a controller). And their investment in Playfab as an Azure service and the scale of Playfab they’ve reached to support Sea of Thieves shows they are thinking very differently about the space. They are looking to create an ecosystem that supports MMORPG and casual games deployed globally via Azure which handles latencies by games being assigned to the closest data centers. Their challenge at scale will be how to lower cost of the Xbox blades and bandwidth requirements and how best to pair and track gameplayer locations while preserving privacy. They will face more latency for game players as they move to the developing world’s less robust cellular networks supported by less data centers.
Google has the fewest data centers and no console experience. Though they captured first mover status with Stadia’s almost 40 games, and their streaming beta of Assassin’s Creed, they lack both the depth in data center locations and the partnerships with game makers to expand quickly into blockbuster games. They don’t have a ready-made LiveOps offering either. Instead they expect customers to piecemeal their system from the marketing, data, localization, and AI offerings that can be glommed together from their narrower service offerings.
Amazon at least recognizes the importance of LiveOps and have snatched up GameSparks to address the space. They have also taken the open source CryEngine and invested heavily in it to create Lumberyard. In the years to come we should see a much tighter integration between GameSparks and Lumberyard, but I know they are a long way from figuring out partnerships and robust console on a server that will be necessary to catch Microsoft or Google in the game streaming space.
Game streaming will not be won by a game company. Sony, Steam, and others have tried and failed. Epic and EA will be able to manage their own games but I think they lack the breadth of skills necessary to support a robust platform offering for third parties. And over time I would expect these game companies to evaluate their data center spend. Is the data and control more important than the bottom line? If they are private companies they can make that decision on their own, but if they are publicly traded companies I’d expect that to be a discussion prompted by shareholders in the future.
The future of gaming and game localization
Game streaming promises to not only disrupt the game deployment model but also the game development, localization, metrics, payment, and testing models. There will be an explosion of customized games that will come as the game mechanics and the operational burdens are separated from the the content of games, and as games become more regionalized for the data privacy, laws, and preferences of gaming communities I’d expect globalization, internationalization, and localization to be a much larger concern for the game longevity. I’ll return to this topic later after I’ve had more time to think about it and to test Microsoft’s XCloud service.