Breaking The Wheel

A picture of a bottleneck, because auteurs bottleneck stuff

Bottlenecks and Hindsight: Why Auteurs Make Horrible Economists

This post is about an empirical issue: the economic cost of being an auteur. When I originally posted this entry on Gamasutra back in 2014 it was not without its detractors. David Jaffe even dropped a line on it, saying he thought it was neat, while simultaneously implying that I was full of shit. Nonetheless, in retrospect, I still feel this idea is worth considering in an industry like ours, one that consists of both public personas and massive-team-based endeavors.


By Reading This Post, You’ll Learn

  • Why more than your artistic reputation is on the line
  • What marginal personal, marginal external, and marginal social costs are
  • What an externality is
  • The definition of survivor bias
  • Why having all of your creative eggs in one basket is a massive risk

Stories about auteur-led projects are rife with anecdotes about decision bottlenecks and wasted work. But so what? If the end result is great, who cares how inefficient the production was? Why should we restrict the creative process of game designers with decidedly non-artistic concepts like budgets or ROI? Well, if your studio is just you then you shouldn’t. Make the art that feels right to you and hold onto it until it is absolutely the best thing you can make.

But, if you are going to spend someone else’s money on development – on a contractual basis, as an employee, or even when using Kickstarter – you have an obligation to be responsible with that money. And if you are going to hire someone to help you make a game/movie/product, you are responsible for that person’s livelihood and, in many cases, the well-being of his or her family. And in those scenarios, budgets and ROI are supremely important.

If you are going to spend someone else’s money on game production, you have an obligation to be responsible with that money. And if you are going to hire someone to help you make a product, you are responsible for that person’s livelihood.

Bottlenecks Destroy Value: A Simplistic Economic Example

Imagine a bridge that sits on a major commuter route. The amount of time it takes each car to cross the bridge is equal to the number of cars on the bridge. So if 20 cars are trying to cross the bridge at once, it takes each of them 20 minutes.

Now, imagine that you and your coworker both commute over that bridge everyday. You can carpool or you can drive separately. If there are already 38 cars on the bridge, it doesn’t really matter to you whether you carpool or drive separately because the difference in your commute time would only be a minute. The marginal personal cost (MPC) you incur by driving yourself is only 1 minute greater than if you were to carpool  (40 minutes vs. 39). And, really, who cares about +/- 1 minute?

The other 38 drivers, that’s who. In addition to your MPC, you incur a marginal external cost (MEC) on them. If you and your friend carpool, you both increase the commute time of the 38 other drivers by 1 minute. In other words, your action consumes 38 minutes of someone else’s time. Further, by deciding to drive separately, you consume an additional 39 minutes of someone else’s time.

In mathematical terms, every car that someone adds to the bridge incurs an MPC of n and an MEC of n-1, for a total marginal social cost (MSC) of 2n-1.  And that assumes that every car only carries a single occupant. If the other cars have additional passengers, the MEC and MSC go up.

Free Markets Aren’t Free: Meet The Externality

This is what economists call an externality: the economic impact your actions have on those around you. Externalities can be positive (you repaint your house and the value of your neighbor’s house increases as a result) or negative (you paint your house black with pink polka dots and the value of your neighbor’s house decreases).

And when you incur an economic cost against someone else, you are destroying value. And if you incur X cost against Y people, the total value you destroy is X*Y.

That concept is ludicrously simple. You might even feel insulted by the fact that I bothered to spell out an equation that basic. But I want to drive home the point that a short delay across several people can add up to a massive waste of time. And the real world often overlooks this dynamic.

Decision Bottlenecks Are Just As Effective At Destroying Value

Let’s apply this same principle to a game production setting. You are an auteur and you want to make sure everything that goes into the game meets your full approval before it’s actually integrated into the build. In other words, you are the canal through which all decisions flow. And you’re a real stickler for details: you want to go through everything with a fine tooth comb.

So, let’s assume that, on average, it takes you 30 minutes to review a potential submission for a feature or an art asset. And you can only effectively review one thing at a time. Five people need you to review their work before they can submit and move on or start addressing your feedback. No big deal. It’s going to take you 2.5 hours to get through it, but that’s your job as the creative lead. You’re being productive and things are totally awesome.

Except they aren’t. The marginal personal cost for each task you review is only 30 minutes. But you’re exacting a massive external cost on your team. Remember: these guys can’t move on to something else until you’ve approved their work. So the first person in your queue loses 30 minutes of productivity. The second person loses an hour (waiting to meet with you and then meeting with you). So on and so forth, to a grand total of 7.5 hours of time someone has spent waiting for or sitting in a meeting with you.

Your bottle-necking just consumed almost a full day of productivity across those five people (or around half a day if you’re crunching).

Now, Scale That Externality

What if your typical daily queue is more like 10 people? For every day of work, you’re now losing 27.5 hours of productivity. What if your queue is 5 people, but each of those folks has an average of 3 other team members waiting for direction from them? In that case, you’re eating up 30 hours. Every day, you are destroying more than a day’s worth of productivity.

Now, let’s extrapolate:

Productivity Destroyed by Decision Bottlenecks (given in 24-hour Days)

So, if you’re bottlenecking like a maniac, and your project goes on for 3-years (the low end for many of the highest profile auteurs in the industry), and you have a five-person queue at any given time, you destroy 234.38 days of productivity. Not working days – 24-hour days. You’ve wasted an aggregate total of 234.38 FULL DAYS of someone’s life.

Certainly, this example is simplistic. Professionals will find ways to be productive whenever humanly possible, and people generally have more than one thing to work on. I’d be surprised if even the most controlling of control-freak auteurs really bottlenecked every decision that badly.

But, if you want to be an effective manager and leader, these are the sorts of death-by-a-thousand-cuts time sinks you need to be aware of. You don’t get centralized decision-making for free.

Down with Auteurs? Of Course Not

I’d be ignorant to say that auteurism is only downside. Clearly there is value in a having a single vision drive a project. Some of the best characters and series have emerged from auteurs. My point isn’t that auteurs are terrible, destructive people. But, if you are going to take or endorse the auteur-route, you need to be sure you understand the trade-off you are making. The more control one person maintains, the greater the marginal external cost to the rest of the team.

And let’s be fair: the opposite risks are true to distributed decision making. People move faster but it can be harder to maintain project coherence. Any development strategy carriers trade-offs in time, resources, or quality. You can’t have your cake and eat it too. My point is not to say that auteurism is invalid, but to point out its risks when pursued recklessly.

People often site Miyamoto’s famous quote, “A delayed game is eventually good, a bad game is bad forever.” There’s a logical fallacy in that statement. Yes, on an infinite timeline a delayed game will eventually be good, much as those monkeys will eventually type Shakespeare. But, on a finite timeline, the sunk costs of a delayed game can become financially irredeemable, especially in the very finite window after launch that publishers care about. And that’s where the auteur theory breaks down. 

The Unified Musk Field Theory of Elon Musk1

One need not look very far to find the examples of wildly successful control freaks in history. Steve Jobs is an obvious one. Elon Musk another. Surely, their success is a feather in the cap of the auteur theory.

Not so.

There are a few things to keep in mind when it comes to looking at the these success stories. First off, SpaceX and Tesla are TERRIBLE examples to use for small businesses. Why? Because Musk was already a god-damned multi-millionaire when he started the former and took over the latter. In other words, he could financially muscle through the extravagances and operational issues that his control-freak, perfectionist nature created. And even then, he and both companies almost went broke doing it.

So, if your Jonathan Blow and can bankroll your own projects, fine. Otherwise, don’t compare yourself to Musk.

Survivor Bias

But he makes amazing products, so he got results, right?  Yes, HE got results. He’s also an eidetic genius who only needs to sleep 6 hours a night, is consistently willing to bet his entire fortune on his own abilities, and can afford to hire an army of assistants and nannies to help manage his personal life.

And even then, Musk almost lost SpaceX and Tesla in the same month. Both companies narrowly survived by getting a round of funding days before running out of cash. He even had a handshake deal with Larry Page for Google to buy Telsa. If that eleventh-hour funding had not come through – if the gambles and bluffs he made to secure that funding had failed – he would have been exactly what the haters at Valleywag thought he was: a modern day PT Barnum.

In other words, if it wasn’t for luck, the world wouldn’t revere Musk the way it does.

Now, I’m not saying that you aren’t a genius, just that you can’t base your managerial model on one hyper-successful unicorn. You also need to make your analysis based on the people who went down in flames.

If you base your calculus on the Musks, Jobs’s, and Kojimas of the world then you’re only seeing the the people who actually survived the brutal filter of commerce. You’re falling for something called survivor bias.

Abraham Wald and Bomber Analysis

During World War II, the Navy did an analysis of the most common places returning aircraft had been hit by enemy fire. The analysts, then made the logical conclusion: add more armor to those locations. Fortunately for the bomber crews, a man named Abraham Wald suggested something counter-intuitive: put heavier armor everywhere else.

Why would he put armor in the areas where the bombers weren’t damaged? Because Wald could see the forest for the trees. If all the planes that returned had bullet holes in the same places, then it stood to reason that the bombers could sustain damage in those areas without crashing. On the other hand if few or no bombers returned with damage in the other areas, then it also stood to reason that damage to those areas was catastrophic.

In other words, if you only look at the surviving plans – if you cherry pick the data –  then the (completely wrong) conclusion is that planes only get hit in certain areas. On the other hand, if you take the entire data set into account (including loses) you have a more accurate picture of where the points of failure are. Planes take damage everywhere, some areas are simply more critical than others.

The Successor Problem

The other trick with the auteur theory, or the Steve Jobs-style central idea man method of leadership, is that it creates a problem when the idea-man or woman exits the stage. If one person is the sole-source of ideas, if one brain is the fountainhead of ideas, then what happens when that person leaves?

Has Apple done anything exciting since Tim Cook took over? By his own admission, SpaceX would not survive is Musk left the picture. If Chris Roberts decides he’s done with Star Citizen, how many people will still pay upwards of $1000 for limited edition space ships?

In short, if all of your eggs are in one, very-smart basket, what happens when that basket leaves? Or, alternatively, what happens when that idea-person stops putting out top-tier work? (See: George Lucas)

Everybody Loves a Bad Idea When It Works

Here’s the trick with genius: it’s only apparent with the benefit of hindsight.

Peter Molyneux famously took 10 Amiga computers from Commodore International when its representatives confused his company, Taurus, for a networking software company called TORUS. That legendary story of video game entrepreneurship helped launch an industry luminary.

But it’s only a great story, and Molyneux only looks like a daring genius, because it worked. If he had been caught deliberately misrepresenting himself and his company, or even been punished in a civil or criminal court, he would have looked like an asshole.

Auteurs Are Only As Good As Their Last Game

The same is true of auteurism: you are only as much of a creative genius as your last game was a critical and commercial success. If your game fails, you can go from a genius with exacting standard to an out-of-touch, high-maintenance, pretentious artíst just as fast as Polygon or Destructoid can post an exposé about it.

As much as this industry – professionals, journalists, and fans alike – loves and adores its heroes, it loves schadenfreude even more. We revel in the bloodshed of a fall from grace like Lisa Bonet in Angelheart. If you’re the creative figurehead for a project, you are also the avatar of its failures. Nobody takes any of the claims Molyneux makes about his upcoming games seriously anymore. Denis Dyack has become a pariah. Ken Levine’s stock has taken a serious dip.

The Slow Burn Of Genius

The time wasted by decision bottlenecks is expensive and the opportunity costs begin to skyrocket.2 Pair that with the reputation for exacting standards and scrapping/replacing existing content that is common for auteurs, and it’s not a surprise that many of them regularly take 4 or 5 years between games. That magnitude of sunk cost is hard to recover. Making games is already a high-risk endeavor, and indulging such a large need for creative control is gambling the fortunes of the publisher, studio, and employees on the convictions of one person.

And that cost carries disastrous consequences when auteur projects fail. To reiterate my previous statement, as a studio head, manager, or lead you have a responsibility to the people who invest in your project. If you don’t deliver on that responsibility, your creative prowess will only carry you so far. Silicon Knights is gone. Junction Point Studios is gone. Irrational Games is gone.


Further Reading If You Enjoyed This Post

Perfection is the Enemy of Productivity, Or: Why You Should Be Like Stan and Jack

If You Want To Lead, Know Your Values (Externally Hosted Content)

5 Vital Business Principles in the Form of Pithy Quotes


Careers Go Down The Toilet With Failed Games

Putting aside the damage done to the reputations of people like Levine, Dyack, and Warren Specter, their teams were negatively impacted as well. Employees bear their own opportunity costs. They forwent other employment opportunities that might have provided more stability. The employment options they might have after a studio closure may not be as profitable as those they turned down to work at your studio in the first place.

And to reiterate my previous statement again, as a studio-head, manager, or lead you have a responsibility to avoid disrupting your team’s lives and those of their families. Your vision is important, certainly. But is it MORE important than someone’s family?

Business is business. Some amount of failure is inevitable. Every employee in every company bears risk. But, if you want to be an auteur, keep the economic costs you incur in mind. The career you impact might not be your own.


Key Takeaways

  • Bottlenecking decision-making to avail one person of creative control creates significant, compounding costs in terms of time
  • “Survivor bias” is the logical fallacy of basing your analysis only on individuals that succeeded, and ignoring those that failed
  • Having one central idea person creates massive problems for companies when that person departs or loses his/her edge
  • High-risk moves are only genius in hindsight; nobody looks smart when they do something risky an fail
  • If you are the auteur, you will own the failures as much as the successes; make sure you appreciate that trade-off

1With all apologies to Ashlee Vance for stealing a chapter title from his biography of Elon Musk
2An opportunity cost is the return available from the best alternative use of your resources. This is the key difference between accounting profit and economic profit. Your accounting profit is your cash-in minus your cash-out. Your economic profit is your accounting profit minus your opportunity cost/s. For example, a venture capitalist can invest in you or in a bond that yields a 10% return. If she invests in you, and you provide a 7% return, you have destroyed value for her even though you turned a profit: she would have made more money with the bond.
In the auteur example above, if your resources are employees who are sitting around waiting for you to give them feedback, the best alternative use for those resources would be to have them actively produce code/content/etc. Wether the value destroyed by inefficient use of resources is offset by the value created by a potentially superior product that sells better is probably too abstract a question to answer definitively. The important takeaway is that it’s not just accounting profit that you should worry about.

If You Enjoyed This Post, Please Share It!

Leave a Reply

%d bloggers like this: