TL;DR - I thought the environment around making movies would be very different from making a game and my research caused me to rethink things. But crucial differences do exist.
Intro
In a recent series of articles we delved into all the things that game developers have to do to ship a game. That research was driven by a spreadsheet that I saw several years ago. A spreadsheet that captured all the essential tasks necessary to create a movie, including budgets, timelines, etc.
I was floored. Having a programmatic set of rules for production must provide serious advantages in predicting completion of the project. After 13 years in the game development industry I've never seen an equivalent, complete framework for games. Why can't we have nice things, too?
As part of my research into why this might be, I started comparing the game and movie development processes. This article shares some of what I have learned.
Non-spoiler alert: designing, building, launching, and running a live service game that results in a positive ROI is very, very hard.
Context
I've shipped a lot of games and have run live games services, so I have an inside baseball view of game development. On the other hand, I am not an actor and I've never made a movie, so no direct experience. On the gripping hand, I am very comfortable researching and learning new things. So please accept my preemptive apology for getting any movie facts wrong.
I am not sharing all the research details here, so my short-hand descriptions might not match up to industry lingo. My apologies for any inaccuracies.
When I began this project I started with a preconception that there would be many large differences. I discovered that there are fewer than I expected, but there are a handful that truly impact predictability for game development.
To keep this short, I'll provide lists. 🙂
Note: the term game platform below refers to the bespoke, walled garden infrastructures provided by companies like Sony, Microsoft, Nintendo, Valve, and Epic.
Similarities
Let's start with all the things that are alike, or at least close. I've tried to be lenient in my comparison rules, so I may have stretched a bit further here and there than you might (like a TV series comparing to a live service). However, I don't see the differences as huge.
✔️ Games and movies can be made by a single person, small teams, or large teams.
✔️ Games and movies can be made on a shoestring budget.
✔️ Games and movies can cost $400M or more to make.
✔️ Games and movies can be crowdfunded.
✔️ Games and movies can both have cost overruns.
✔️ Game platforms serve the same function as movie theater chains.
✔️ Games and movies can leverage existing artifact (physical and digital movies; game clients) distribution systems. Movies are delivered to theaters, game clients are delivered to your hardware.
✔️ Movie theater concession stands are equivalent to game DLC stores.
✔️ Games and movies can both take advantage of project management software.
✔️ Games and movies can be 'live services' (classifying a TV series as a live service).
✔️ Games and movies can be of variable length in experience.
✔️ Games and movies can provide a linear experience (single player, story driven games).
✔️ Games and movies can be replayed or rewatched.
✔️ Games and movies can be delivered both via streaming and physical media (albeit with differing success and quality).
✔️ Games and movies can be obtained through subscriptions.
✔️ Games and movies can both have personal, in-home experiences.
✔️ Games and movies that are sequels to hits have a better chance of success.
Differences
Now let's enumerate the things that are different. These differences in and of themselves don't explain why we don't have beautiful spreadsheets that capture budgets within reasonable error bars, though.
❌ We have over 125 years of experience making movies while live, multiplayer games are only around 30 years old. Single (or coop) player games are older, call it 50ish years.
❌ Movies start with a script that is the essential core of the movie. Games generally start with much, much less.
❌ Movies are only experienced linearly.
❌ AAA movies are generally much shorter compared to AAA games.
❌ AAA movies with a length of 4 hours are considered too long and a AAA games that last just 4 hours are considered too short.
❌ Movies have been getting longer (143 minutes average in North America in 2023) while some shooter campaigns have been getting shorter in favor of live, multiplayer modes.
❌ The streaming movie services are ironed out but cloud games, games that just stream video to your device, are terrible (some services have already died).
❌ ESports - enough said, unless there is a reality TV show about watching movies. :)
❌ The average per-hour cost of a AAA movie is significantly more than for a AAA game. A movie at prime time ($15) is from $5-$10 per hour (1.5-3 hr). A AAA game at full price ($70) is from $0.70-$11.67 per hour (6-100 hr).
❌ The value of a game can depend on the player -- completionists can spend 100's or 1000's of hours.
❌ Free-to-play (AAA or otherwise) games can cost nothing to play.
❌ There are no free-to-play AAA movies.
❌ Movies don't crash after selling you a ticket or get DDoSed.
❌ Hollywood has a freaking formula for shaping the viewers emotional experience.
Critical Differences
There are, however, a few very important differences that contribute to the lack of predictability in game development. These are they.
😩 Multiplayer games (clients and servers) are large-scale software systems with complex interactions.
😩 Games often build their own backend and game server infrastructure. These are also large scale software systems that manage logins, monetary payments, player trades, and many other things. Doing this is comparable to designing and building your own theater chain.
😩 The cost of building large scale software systems is nonlinear w.r.t. size (lines of code, number of interacting systems) and getting those working right is notoriously difficult.
😩 Game teams often don't know what game will actually be built when work begins. There is no script that captures the complete game, and there's no equivalent to the Hollywood formula.
😩 Games can be nonlinear, and/or open ended sandboxes, and/or multiplayer.
😩 There are people with insufficient experience who are certain they can make a game. When the environment is right those folks can get funding from investors, hire a team, and start working.
Outro
Making a AAA game can be much more complex than making a AAA movie. Large scale software systems are notoriously difficult to build correctly because complexity scales nonlinearly with system size.
It is not necessary to reinvent all the wheels, but that means delivering on one or more third party game platforms. If the team wants more systems without building then that can mean partnering with third party services. If those work, then great.
But most teams want to explore new and different things. Passion to do something new leads teams to try things beyond their level of expertise. Things that other games have not done yet. They don't want to be restricted by existing systems, even if those systems work.
In fact, if your game is only slightly different from a previously published game, then you get dinged pretty hard. So, wheels get reinvented in the hope of hitting something that has never been seen before. And costs get multiplied when design teams decide to go back through a door that was supposed to be one-way.
All of this makes sense when we remember the #1 rule of making a successful game:
🙂 The game MUST be fun.
Because if it isn't, the rest doesn't matter.