Next Gen Development: Beware the long tail

The games industry will undergo a transformation during the development of the next generation of AAA games. The early phases of the development cycle will closely mirror what is currently done today with almost identical concept, prototyping, and vertical slice phases of development. Since these phases are at the beginning of the project and last for many months, it will give the impression that the development cycle for large AAA games has not changed significantly.  Logically, this implies that development costs will be similar to the current generation of games.  However, that is only half the story. Getting a project off the ground is very difficult and getting it to market as a quality product is another huge problem entirely. This is where the increasing costs of development will balloon in the next generation of games; not at the start of the development cycle, but in the widening and fattening tail of the development cycle.  This is where features and assets need to come together to flush out an immersive gaming experience.

Controlling the message

Depending on who you listen to, next generation development cost will either be relatively flat with a small bump in costs or double the cost of current titles. EA CFO Blake Jorgensen believes that development costs will be largely flat, with an estimate of a ten percent rise in costs over this generation.  This statement, at face value, strikes me as shockingly optimistic, especially when you consider that Epic Game’s CTO Tim Sweeney has spoken about the exponential increase in development costs.  These increased development costs are estimated based on Epic Game’s own investigation of what it takes to produce a next gen demo using an industry leading tool. Activision as well has said that costs will rise, as they have in every previous generation transition, but give no additional specific details. Crytek founder Cevat Yeril has been vocal about getting ready for the next generation of games and said their last title, Crysis 3, was triple the cost of the original Crysis, costing an astonishing $66 million. Perhaps even more surprising than Blake Jordensen’s comments are those of Take-Two Interactive CEO Strauss Zelnick.  He stunningly believes that development budgets will remain flat, and also thinks that top quality titles will not see a meaningful change in development costs as Take-Two, and presumably other publishers, will become more cost conscious.

Early indication

When a publisher says that next generation costs will be flat, I tend to hear “Don’t spook the shareholders; they are a nervous bunch”. Given that next gen titles will rise to $70 there are numerous reasons why a publisher would not want to reveal an increase in the cost of development. If development costs are largely flat, why would publishers raise the price of a next gen title by almost 17%? Clearly more revenue is desirable from a sales perspective, yet in the highly competitive gaming market such an increase could easily price you out of the sales range of many gamer’s dollars.  Rising development costs is likely the primary reason explaining raising prices. Also, if rising development costs were not the cause, and next gen development costs are indeed flat, why would publishers feel they have to raise prices? 

Unlike publishers’ claims to the contrary, when a developer says that next gen costs will double, I tend to believe them. Why? There is certainly motivation for developers, especially those with engines and toolsets to market, to proclaim that their tools keep development costs down. It is in the best interests of the Epics and Cryteks of the world to produce tools that are powerful, intuitive, and as production ready as possible in order to keep costs down and appeal to developers. To market their products they need to make games that showcase their potential to the community. Without exception, the best way to do this is to put out a great game using ones tools. If the toolset does not live up to the developer’s hype, internal costs will rise and quality will suffer during development. This does not make for a convincing case to adopt ones tools, especially when contracts are often signed for a series of games on multiple platforms.  This development cycle will last years and involves an ongoing relationship between the tools provider and developer. Clearly some pains have already been felt by both Epic and Crytek as they push the limits of development to prepare for the next generation of AAA games. As a result, they make a very strong case for an increase in the cost of development for next generation games. What is not mentioned are which areas of development will drive these cost increases?

Army of artists

While the technology and tools for the next generation of games have already been developed, it is the assets for next gen games that will drive the bulk of the cost increase. More memory per console means more assets will be required to fill that memory. Next gen consoles are coming with a tremendous increase in memory over current consoles. This means that every environment and scene in a game will be required to be more densely populated with assets to take advantage of the additional memory and computing power provided by new hardware. Not only will there be more assets per scene, but the quality per asset will be higher. This means that to produce individual assets an additional percentage of time will be required over the current generation assets. When this is scaled out over tens of thousands of assets, the increased amount of time required for asset development quickly balloons, as do the associated costs. The art team will have to grow to deal with these assets, adding numbers to an already large group at most studios. 

A possible mitigation strategy is to increase the use of outsourcing for asset development.  Developers may try to take advantage of a small army of external artists to help speed up asset development. However, this strategy comes with its own risks.  Most notably are quality concerns and the turnaround time required for feedback and changes. What looks like a great way to keep costs down and get a quick turnaround on assets can quickly turn into a nightmare for developers. When quality falls below expectations, iteration and feedback become next to impossible and timelines are consistently missed. Also, let’s not forget that even with a quality outsourcing partner, assets will take longer to make regardless of where they are made or who makes them. This means developers will be faced with an increasing number of assets to produce, a longer timeline to produce each of those assets, and a greater difficulty iterating on the game as a whole. All of these factors will mean that the bulk of the team has to stay on longer during the development cycle reworking assets, and trying to piece together a game experience while dealing with huge amounts of assets and delayed deliveries. At some point as well, development is going to need to engage in large scale quality assurance to deal with bugs and prepare for platform certification and gold master. Of course, all of the issues encountered during the production phase of the project will now spill over into the alpha/beta phases of the project.

Quality suffers

Once the alpha/beta phases of a project begin bugs will begin to flood in. With an increase in assets comes an increase in asset related bugs. Tracking, coordinating, and ensuring that assets are fixed will require tremendous amounts of time from existing art staff. This means that internal artists will be unable to move to other projects and that the current project must absorb the costs of retaining art staff longer just to fix bugs. Asset development won’t be the only area where bug numbers increase. With the rise of social and multiplayer focus, bug numbers will jump up considerably in terms of adding additional devices and extra features. Larger feature sets consisting of never-before-implemented social media integration will require extensive testing, which will involve a large team of testers due to the nature of multiplayer testing. And of course, let’s not forget the costs associated with making changes later in the development cycle. With bug fixes come the inevitable changes to game mechanics and features that aren’t quite working as intended. Even worse is potentially reworking a feature or system that does not live up to expectations at all. 

Bracing for Pain 

The functionality, features, and considerable increase in assets in the next generation of games will require considerable expanded testing efforts.  This will put an increasingly heavy load on QA and push out the alpha/beta phases of the development cycle considerably. Given the vast amount of investment in fewer titles, there will be mounting pressure to hold back a game until its quality is up to a certain standard required for release. This means not just a longer tail when compared to the current development cycle, but also a fatter tail as more members from the development team are held back to deal with last minute changes and bugs. Though some think the development costs for next gen might look flat now, they should wait until AAA titles start getting closer to the end of their development cycle. Beware the long tail.