From Transactions to Relationships
The last decade of games has seen the rise of Games as a Service (GaaS) and the inglorious falls[1] of games attempting this business strategy. However, service-based games remain the dominant business model: they are almost exclusive on mobile and the major titles on consoles have adopted many features of service-based models. Yet, this does not mean all games should be service-based[2]. In this blog post I’ll dig into the main reason why this transition happened and why we have seen so many games fail. In short, the whole reason for live games is that there’s more value in building relationships and not just selling a game in a one-off transaction. However, this approach also has higher costs and these are not apparent from the beginning. Many underestimated the cost and complexity involved.
As I argued in the The Rise of Games as a Service already back in 2013, the F2P phenomenon was a feature of a bigger transition from boxed products to online, live games. It was just monetization strategy in this new frontier, and a very successful at that. What service-based promised publishers (and, more importantly, their stockholders) was recurring revenue, a stable revenue stream in a portfolio-led model where revenue could fluctuate with hits and flops. The other angle was the early promise of the internet, the loss of middlemen and a direct, closer relationship with players. However, this relationship would require them to operate infrastructure, support, and content pipelines. As it turned out, few games built durable businesses from this, but those who did, made insane amounts of money, both on consoles and mobile. Service-based games turned out to be very successful, but only to a select few who could acquire and retain their audience profitably - leading to games that have been live for well over a decade.
The pitch
A long time ago, the traditional product was simple: make a game, launch it, sell it to stores and move on. Use the revenue from the previous game to fund your next one. But this was risky, especially in a hit-driven market. Every game is a new clean slate. Sure, you might have IP recognition, either as a studio or through the game brand, so you didn’t have to figure out your potential market and audience from scratch each time. Everything is very cyclical. As a publisher, you could try to manage this risk by having a portfolio of games which would spread out your risk over time and product-market fit.
Your game might have some post-launch plans in the shape of DLCs or have multiplayer mode to keep players engaged after they had completed the main campaign[3]. Multiplayer being a feature of service-based games enabled by the adoption of high-speed internet connections. I mention multiplayer here so to emphasise that microtransactions or even being free-to-play should not be seen as the defining feature of a service-based game, or a GaaS title. These are just monetization choices, not definitions of a service.
However, with product-based games, it is such a waste to do these one-off transactions with customers, whose attention and wallets you need to capture again and again each cycle. Multiplayer modes would ensure that your game and IP would at least keep some share of mind in your audience while you were busy developing the next installment. Regular updates might bring players back, and a large, active community could make a game more appealing to new players. However, you’d still have to move that audience to your next title to keep the revenue coming in.
Once the building blocks were there, it is hard not to see the appeal of offering one’s games as a service: recurring spending and longer customer lifetimes. Maybe even sidestepping traditional gatekeepers and middlemen, and so keeping higher share of the revenue for yourself. I went deeper into the distinctions between products, services and platforms back in my post What I talk about when I talk about Games as a Service.
The even further promise of “games as platforms” hasn’t moved too much from what I wrote in 2018, except that streaming and eSports turned not to be the keys to this[4]. Fortnite and Roblox, and maybe Minecraft, remain the few successful examples.
The promise
Digital distribution and always online had other appeals to publishers as well: ongoing feedback. Going direct to customer allowed publishers and developers to communicate with players, observe how they play and upsell to them without depending as much on retail. For players, this very often manifested as creating yet another online account[5]. However, even on the digital platforms - especially on the consoles - the platform holders still controlled a lot of the customer surface: storefronts, authentication, discovery, payments, and so on. Even though the publishers started to own more of the customer relationship, they were far from owning the entire path to the customer. On consoles this would be impossible, on mobile it was a bit more flexible, but even on PC it wasn’t without major investments: see the rise and fall of many publisher launchers, that eventually folded back to Steam. Another example is the uphill battle that Epic has had with its Epic Games Store: a game’s audience does not automatically become a platform’s audience[6].
What wasn’t necessarily clear was how all of this would affect the competitive environment. How would online, F2P, platforms and other factors change the bargaining power of players, the threat of new entrants or substitutes? To be fair, it would take a long time for all of this to become apparent. For a while, it seemed there was still room in the ocean for fishes of all sizes.
The iceberg
However, a game does not become a “service” or “live” with occasional updates. That’s just a DLC model[7]. Just because your game is “online” doesn’t mean it’s a service. This distinction might sound obvious now, but the machinery and costs are easy to miss when the business case is just a line going up from recurring purchases - just because.
Live games need a steady stream of live ops: content, events and fixes. The game needs to monitoring before, during, and after a release, and be on call to respond to incidents, and keep doing this while also working on the next update. Suddenly there are multiple tracks to run parallel for a single game. Not that products didn’t need post launch support, or the occasional patch. The difference is that a service makes maintenance, operations, and support part of the ongoing promise to customers. This machinery is expected, and it is somewhat removed from the core of the actual game, the fun experience.
And that is just the player-facing parts, the visible part of the iceberg. To run a service, there is infrastructure to build and operate: accounts, matchmaking, databases, payments, analytics, … There are servers to scale and maintain. There is security work for moderation, fraud and other anti-abuse. There is customer support for lost accounts, purchases, bans, outages, and all the other ways the game might fail. Handling all these customer touch points requires real organizational change: new teams, new competences, new tools, new ways of working.
The treadmill
Operating a service changes the economics. The revenue is recurring, but so are many of the costs. A game can have a large audience and still be expensive to run. Players can leave faster than new content brings them back. The game teams doing a service-based game need to be funded for the long haul, not just through launch. If lucky, the game might be “live” profitable soon but the operational costs eat into the payback period for reaching break-even for the project as a whole.
As I mentioned in Transition From Product-Based Game Development to Service-Based Games, for a traditional title, the development costs go essentially to zero so essentially all cashflows are positive after launch. A GaaS title on the other hand has significant operating costs and for this reason a downward trend of engagement is unsustainable.
The service-based game is like a startup: it’s a growth business and the launch is where the iteration starts. A normal product sees a huge spike of sales at launch assisted with full court marketing campaigns, with seasonal spikes that ultimately turn into a long tail of “back catalog” sales. A service-based game on the other hand aims to grow year-on-year.
Maintaining existing audience and attracting new players is difficult and expensive. This has a drastic effect on what happens if a service-based game does not seem to be sustainable. For a traditional product, you’d just cancel post launch plans and try to recoup as much you can through discounts. However, a service has recurring costs so continuing running the game can easily be just throwing good money after bad. The sunk cost fallacy can be very real with service-based games, especially with a free to play model.
The winners
That trade has worked spectacularly for some games, just look at Fortnite. Its community, recognizable brand, and frequent spectacular updates give players reasons to return and spend over time. But the visible successes can make the model look easier to reproduce than it is. Fortnite is not just a shooter with a content calendar. It is dangerously easy to fall to the cargo cult to just repeat the aesthetics of a successful game but miss all the parts that make the original successful[8].
The transition from products into services has been challenging to many previously successful studios. Nothing exemplifies this better than Destiny. The game and its developer have had a challenging journey, and yet it is still one of the more successful transitions - at least in my opinion. The product offering of the original Destiny was very traditional: immersive campaign, multiplayer, expansions and DLCs - and a promise of a decade of Destiny. They fell victim to the content treadmill. It attempted to be a service, a MMO, bringing in the staples of F2P: microtransactions, loot boxes and season passes. And worst of all, they were working on a sequel as a follow-up[9].
The irony of it all is that after announcing the end of service for Destiny 2, I think they realized the market realities and risks of creating anything new in the current state of … well, everything … that it’s better to keep on working on what you have and try to make it work than starting from scratch. I’m eager to see what the next chapter of Destiny will be, even though I dropped off quickly after Destiny 2's main campaign.
The ocean of service-based games is a deep red one. They are black holes that have an iron grip on the free time of their audience. GaaS did not become the only model in town[10], and the annual franchises are still here - even if they adopted many features of service models. And then there are those who can have it both ways, like Grand Theft Auto as AAA product and its GTA Online as its service-based follow-up.
Conclusion
Games as a service transformed the one-off transaction into a relationship - and made the publisher responsible for maintaining that relationship. The monkey paw delivered: more time with players, more ways to serve them, and more chances to earn their trust and money. But the infrastructure, customer support, and new competences required behind that opportunity were integral part to make it all work. The mirage of “recurring revenue” turned out to be an iceberg.
Footnotes
It is unfair that most people probably think of Concord's short lifecycle here. Untold number of games died in the dark before they were ever seen by the wider public, as it became clear that they could not hit their business KPIs. The real question is, did they think Concord stood a chance, or did they just want the world to see what they had worked on for so many years - even if it ended in mockery? ↩︎
See Using Steam’s Basic Usage Statistics to Measure Staying Power of Games for analysis which games on Steam seem to have good dynamics for being service-based, and how different games/franchises manage to retain their audience. ↩︎
This would go to such lengths that very single-player games like Mass Effect or Uncharted would have multiplayer modes “just because” that’s what the (stock) market demanded. I don’t mean the multiplayer modes were often bad, but usually very unconnected to the main game and might serve an entirely different part of the game’s audience. The benchmark was probably set by the AAA FPS shooters like Call of Duty and Halo that had their main campaign, from which people then migrated to the shooter’s multiplayer mode for years to come. ↩︎
And, fortunately, neither did the NFT-fueled hype of “metaverses”. ↩︎
A distinction needs to be made between platform accounts and publisher accounts. Steam, Xbox, and PSN being the former, and Uplay, Battle.net, and EA/Origin being examples of latter. The former would not allow a third-party developer to build much relationship directly with its players. ↩︎
This is also a challenge Roblox is facing. ↩︎
Which is not necessarily a bad thing, this is very much Paradox’s or even Euro Truck Simulator 2's model. ↩︎
The worst part is that not even the original game’s team have a reproducible cookbook. Having a great game is just the table stakes, and reaching even that is very, very difficult to pull off repeatably. ↩︎
Where are the sequels to Fortnite, World of Warcraft, Clash of Clans, Dota 2? The whole point is that if you’re successful, you do not have to make a sequel and you do not have to re-acquire your audience. It’s also very hard to not only maintain a live game and work on another at the same time. Should you focus on what you have now or do you need to start again from zero? ↩︎
Although I do not want to underplay the effect it had on other business strategies, but it was hardly the only one affecting the landscape. ↩︎