Skip

The Beer Game
April 5, 2012 10:17 AM   Subscribe

Mark Sweep describes an interesting game from his childhood which helps explain why companies move and keep their manufacturing in China.
posted by gilrain (34 comments total) 13 users marked this as a favorite

 
Delay in the supply chain causes amplified downstream problems. The problem wasn’t that we were kids running beer supply, the problem was the structure of the chain itself. Small changes at the front end lead to massive mistakes down the line.

And that's because you defined hysteresis as a bad thing right up front: You lose money for keeping inventory and for unfulfilled backorders.

If you took a "loss" on storing some inventory you would be loosely coupled to your supplier. Loose coupling is what prevents resonance.

You cannot have a system that is optimized both for profit and robustness. Which do you want?
posted by DU at 10:28 AM on April 5, 2012 [9 favorites]


Ah, the bullwhip effect. I learned all about it one day I was searching for new bullwhip cracks to learn. Pretty interesting stuff but not as cool as an actual bullwhip.
posted by fuq at 10:28 AM on April 5, 2012


This dad sounds awesome, though and I'd really like to try that game.
posted by DU at 10:29 AM on April 5, 2012 [3 favorites]


Apple needs to have factories in China because the supply chain is there.

This is about as useful an explanation as Turtles All the Way Down. So ask the next question: "Why is the supply chain there?" Ahhh. Cheap labor. Poor working conditions. Little concern for pollution and the environment. Total lack of foresight for our grandchildren, who will pay the hidden costs.
posted by weapons-grade pandemonium at 10:33 AM on April 5, 2012 [24 favorites]


You could use PID control instead of hysteresis, but that depends on your input being a continuous function. That fails under many conditions too, although probably not beer sales.

I really don't get the absolute loathing finance people have for stocking inventory. "Time value of money" blah blah blah. There's value in being prepared for catastrophes too. And I don't mean "we don't have enough beer" catastrophes. I mean "the solar flare knocked out the power grid and we only have 1/100th the parts we need to rebuild it" catastrophes.
posted by DU at 10:44 AM on April 5, 2012 [1 favorite]


And that's because you defined hysteresis as a bad thing right up front: You lose money for keeping inventory and for unfulfilled backorders.

In the context of the game it was defined to be that way, but in the real world it's that way almost immutably. Keeping inventory costs money because warehouse space costs money. Unfulfilled backorders cost money because contracts and competitors exist. One cannot merely define those facts away.

Loose coupling is what prevents resonance.

Apple loosely couples itself to suppliers in other ways, typically by having multiple suppliers and switching between them as needed. It sources display panels from multiple companies, for example. Foxconn is a major supplier of final assembly (over 90% of Apple's products), but Apple does use a few other companies.

Interestingly, extremely high import tariffs in Brazil have led Apple to begin assembling iPhones and iPads in Brazil. So at least in some cases Apple is willing to bring manufacturing into a country given the right incentives. Of course, most economists would say that such protectionism is ultimately a bad thing for everybody.

You cannot have a system that is optimized both for profit and robustness. Which do you want?

By moving its supply chain to China, Apple has minimized the risk of the kind of problem described in the article. It may not be "robust" by your definition, but it is safe. And also very profitable.

But moreover, a business must necessarily favor profit over robustness because a business that loses money will necessarily fold, whereas a business that isn't robust can stay in the game indefinitely, especially if it saves profits in order to cover the occasional loss.

This dad sounds awesome, though and I'd really like to try that game.

According to Wikipedia it came out of MIT's Sloan School of Management in the 1960s. There are some online implementations; links are on the Wikipedia page.
posted by jedicus at 10:45 AM on April 5, 2012 [2 favorites]


It's also very silly, because what producer wouldn't be talking to his retailers about what actual demand they were looking at?
posted by Malor at 10:49 AM on April 5, 2012


I mean "the solar flare knocked out the power grid and we only have 1/100th the parts we need to rebuild it" catastrophes.

Let's assume you're talking about a company that makes electrical components and not the utility itself, since arguably a utility is a public good and shouldn't be so focused on profitability.

So this company decides to keep a lot of extra inventory around. First that means producing extra inventory, so there's an immediate loss. Then they have to buy the warehouse space to keep it, so that's an additional ongoing cost. If the company doesn't have the capital to do all of this, then it has to take out a loan, so there's interest. And then there's the opportunity cost of investing in all of this extra inventory as opposed to R&D, marketing, or whatever.

Their competitor, on the other hand, has decided not to follow this strategy. They don't need to build extra production capacity, hire more workers, and buy more input materials to produce extra inventory. They don't need to buy extra warehouse space, pay property tax on it, etc. They can invest their profits in expanding the business or R&D.

In a competitive market, the margin on these components is going to be fairly thin. The losses could be enough to make the first company unprofitable or at least upset the balance of competition to the point that the second company runs away with all of the marketshare.

So it's all well and good if a solar flare happens and the company wins a contract to rebuild the grid, but that's a bit of a gamble and in the mean time the company may go bankrupt (whereupon the competitor may buy up that inventory for pennies on the dollar). It doesn't take much risk adversity to choose not to gamble on producing extra inventory in the hopes of cleaning up after a catastrophe.
posted by jedicus at 10:58 AM on April 5, 2012 [4 favorites]


When President Obama asked Steve Jobs what it would take to make iPhones in the United States, the late Apple co-founder supposedly quipped: “Those jobs aren’t coming back.”

They'd come back if the cost to transport the devices from overseas outweighed the savings in labor from importing them. That could be done using tariffs. Although it would probably harm relations with China, it's not the case that we need to help China give jobs to their labor force at the expense of our own. Is it?
posted by JHarris at 11:00 AM on April 5, 2012 [5 favorites]


Of course, most economists would say that such protectionism is ultimately a bad thing for everybody.

Yes, this is true, in the very long run. In the meantime there are all kinds of problems with the world in the short run that could be helped with it, and in the long run we're all dead.

And really, we're only interested in the well-being of this particular aspect of economic climate because China has such huge piles of cheap labor to supply, as there are plenty of other unhealthy problems with the system as it stands that we don't seem especially serious about solving. We care about cheap foreign labor more than we care about domestic unemployment.
posted by JHarris at 11:07 AM on April 5, 2012


Their competitor, on the other hand, has decided not to follow this strategy. They don't need to build extra production capacity, hire more workers, and buy more input materials to produce extra inventory. They don't need to buy extra warehouse space, pay property tax on it, etc. They can invest their profits in expanding the business or R&D.

Well yes. They decline to invest in the future because it would detract from immediate profits. I think that was my point.
posted by DU at 11:10 AM on April 5, 2012


They'd come back if the cost to transport the devices from overseas outweighed the savings in labor from importing them. That could be done using tariffs.

As I mentioned above, that worked for Brazil. But a trade war with China could potentially be very bad. China has a lot over leverage over the US. For starters, it could decide not to buy US debt. Or it could impose tariffs of its own. The US exports about $100 billion worth of goods to China every year (compared to ~$365 billion in imports from China), so that would be a big hit to the US economy.

Negotiating labor improvements, environmental reforms, and IP enforcement that ultimately make US labor more competitive and US exports more valuable is probably the way to go.

We care about cheap foreign labor more than we care about domestic unemployment.

I would argue that there are other, better solutions that are less likely to spark a trade war. For example, a mandatory 35 or 40 hour work week (or alternatively a mandatory two or three weeks of paid vacation) would force employers to hire more people rather than pile on the unpaid overtime. An increase in the earned income credit would enable more families to live on a single income. Alternatively, a more progressive tax code. Single payer, universal healthcare would do the same. And of course a guaranteed minimum income is the straightforward solution.
posted by jedicus at 11:13 AM on April 5, 2012 [1 favorite]


Well yes. They decline to invest in the future because it would detract from immediate profits. I think that was my point.

Investing in marketing and R&D are not investments in the future how, exactly?

But even if the competitor simply paid all of its profits out to shareholders and didn't reinvest it, the point still stands. Piling up extra inventory in hopes of making a killing later is a significant gamble, and one that does you no good if you go bankrupt in the meanwhile.
posted by jedicus at 11:16 AM on April 5, 2012


Delay in the supply chain causes amplified downstream problems ... Apple needs to have factories in China because the supply chain is there.
This seems like an inadequate reason/explanation/excuse. The supply chain in China didn't materialize in an instant. Manufacturing didn't pack up and move there overnight.

If delay in the supply chain is really the only important variable, then how did the supply chain ever migrate its way over there, piecemeal, in the first place? A company building widgets in the US with some widget parts and sub-assemblies made in China would have been dealing with this bullwhip effect big-time. Outsourcing some of the supply chain to China would have been a disaster.

But that's not the way things seem to have worked out.
posted by Western Infidels at 11:19 AM on April 5, 2012


Piling up extra inventory in hopes of making a killing later is a significant gamble, and one that does you no good if you go bankrupt in the meanwhile.

You say this as though the answer is so clear. But in any context other than business the answer is clearly the other way. "We decided not to give our kid braces because what if he died before he was old enough to mate?"

The problem here is not that I don't understand the tradeoffs. The problem is that businesses are set up to maximize one thing and one thing only.
posted by DU at 11:23 AM on April 5, 2012


Oh and one more rule: no communicating between parties - your only communication was through order slips.
This is not at all like the real world, and seems designed explicitly to create a disaster - why include this rule otherwise?

Real beer distribution does, in fact, work, delays and all, after all.
posted by Western Infidels at 11:24 AM on April 5, 2012


According to Wikipedia it came out of MIT's Sloan School of Management in the 1960s. There are some online implementations; links are on the Wikipedia page.
posted by jedicus at 10:45 AM on April 5 [+] [!]

yes, it was originally created for management grad students and business executives. It is still one of the most well known and widely practiced business school games.
I wish the blog post had pointed this out rather than just stick to the "something my dad made for us kids when we were little" folksiness angle, but maybe the blogger just didn't look it up
posted by Bwithh at 11:31 AM on April 5, 2012


Obama was right in bailing out the auto industry when GM and Chrysler were going bankrupt. Because it wasn’t just GM and Chrysler that were going to fail, it was an entire ripple-effect of suppliers, tool manufacturers, raw material suppliers, trucking companies and dealers that were going to go away.

I believed this but didn't understand why. Now I do. And there is no way in hell that any of the conservative types that I know will understand or agree to it. But I will try. And suffer.
posted by Splunge at 11:35 AM on April 5, 2012


The other thing about locating your manufacturing facilities in China is that there are a whole lot of people there. Apple will probably be selling more products in China than in the US in a few years, if they aren't already.
posted by ghharr at 11:38 AM on April 5, 2012 [1 favorite]


This is not at all like the real world, and seems designed explicitly to create a disaster - why include this rule otherwise?

I believe the point is to show how misperceptions are magnified down a supply chain, not to be a beer distribution sim.
posted by ChurchHatesTucker at 11:42 AM on April 5, 2012 [1 favorite]


Another reason not to keep extra inventory, is that if the beer is skunked, by the time someone tastes it, you've already got a big pile of bad beer.
posted by StickyCarpet at 11:43 AM on April 5, 2012 [1 favorite]


This is pretty much how places become known as "The Glass City", "Hog butcher to the world" or the "Motor City". Not only do you get the large company that produces glass, you get all the other supporting businesses that supply that company in order to keep the supply chain short. And then more, similar companies move in to take advantage of that supply chain...

Pretty simple stuff, really. China originally made cheap bits that went into electronics and eventually it became more economical to move the whole supply chain over there. To get a meaningful number of jobs back, you can't just move some factories, you need to move the whole industry. It is actually better to first be the people producing the parts before becoming the places for assembly.
posted by charred husk at 11:48 AM on April 5, 2012


I really don't get the absolute loathing finance people have for stocking inventory.

I understand it but it drives me batty. I hear the phrase "it's not in stock but we can get it for you over night" at least half a dozen times per week.

Worse though is the computer systems of the wholeslaers are set up such that there is no way to signal demand. If I've got four wholesalers who can supply a widget but two are prefered, the third is OK and the fourth is a last resort but the perfered vendors don't stock my desired widget and the third or fourth does I end up dealing with a non prefered vendor and the prefered vendors have no way of fixing it because the staff are limited by their computer systems. Stock is based on orders and the prefered vendors never get an order because they don't have the widget in stock.

The worst is when I need something so esoteric that no one but the manufacturer stocks it and it'll take four weeks to get it. There is no way something should take four weeks to get from point A to point B in North America.
posted by Mitheral at 12:13 PM on April 5, 2012 [2 favorites]


There is no way something should take four weeks to get from point A to point B in North America.

You're right, of course.

I'm prepared to bet that a big chunk (not all, but a hefty portion) of that delay is due to two things: 1) there being no formal procedure for acquiring the item you want (atypical requests are handled atypically), and 2) because they want to "punish" you for being an irritant (to teach you to only purchase those things which they want to sell you, for which they have neat & tidy processes already in place).

Never underestimate the power of well-worn procedures in corporations -- they determine what is easy, and what's nearly impossible. Procedures are frequently more important than profit, oddly enough. Or, at least, that's my personal explanation for why some companies seem to be unable to pull out of a nosedive, even when everyone involved sees what's happening; it's too hard to deviate from their procedures, so they just sorta drive the company into the ground hoping the external conditions will change before they are forced to change internally. When that doesn't happen, the company dies (cf., potentially, RIM).
posted by aramaic at 12:35 PM on April 5, 2012 [4 favorites]


I used to be a purchaser. I managed several million dollars worth of cabinet-building supplies for a wholesaler. Before that, I was in sales - so seeing the exact-opposite-world of purchasing was an eye opener.

There are several legitimate reasons for keeping surplus inventory to a minimum; namely, surplus is essentially dead money that could be used for other purposes and standing inventory is a taxable asset.

That said, modern management (as usual) has gone way overboard in this area. Just-in-time inventory is the vogue because it aims to squeeze every last nickel out of inventory management possible. The problem is it's very difficult to do correctly (see the bullwhip effect above) and, done badly, it simply hoses the customer.

The closer modern management can get to something-for-nothing, the happier they are. Fuck the workers and the customers.
posted by Benny Andajetz at 12:39 PM on April 5, 2012


Apple loosely couples itself to suppliers in other ways, typically by having multiple suppliers and switching between them as needed.

For example, with the 2011 Mac Air 11" 256GB SSD, there were two different suppliers of displays and of SSDs, so four possible unique combinations of sources just for those elements.
posted by zippy at 12:54 PM on April 5, 2012


You say this as though the answer is so clear. But in any context other than business the answer is clearly the other way. "We decided not to give our kid braces because what if he died before he was old enough to mate?"

That context isn't really comparable. You're comparing stockpiling parts in case of a solar storm (very hard to predict when it will happen and statistically unlikely) versus a child reaching adulthood (easy to predict when it will happen and statistically very likely).

The point of the game is to show that forecasting errors lead to massive problems. When forecasting is easy (i.e. most people survive to adulthood, most people attract partners or at least want to), it's much easier to invest in the long-term. And the braces example doesn't really have the multi-stage structure and iterative nature of the manufacturer->distributer->wholesaler->retailer relationship that greatly magnifies forecasting errors.

Your example also suffers from the first world perspective. If you consider a family in a developing country with high infant and child mortality, they might well decide not to invest very much in any given child because, for example, they might die of an unpredictable disease before they're old enough to work. I don't think many people would fault such a family for not throwing all their eggs in one basket. So even in personal relationships people discount the future in favor of the sure thing today, and often that sure thing is making money.

But in general it's a good thing not to treat business relationships like personal ones. If you treat business relationships like personal relationships you get things like cronyism, irrational revenge-seeking, and the old boys' club.
posted by jedicus at 1:29 PM on April 5, 2012 [1 favorite]


This seems like an inadequate reason/explanation/excuse. The supply chain in China didn't materialize in an instant. Manufacturing didn't pack up and move there overnight.

Foxconn was founded, and is still headquartered, in Taiwan. They opened their facility in Shenzen in the late 1980s. The supply chain has been put together piecemeal over the last 30 years.

As someone mentioned above, the capital and the manufacturing plants followed the cheap labor.
posted by notyou at 1:54 PM on April 5, 2012


Interesting article. Although I have to question the use of somehow bringing back iPad manufacturing jobs to the US. China makes $8 on every iPad made. The US makes $150 per unit (to Apple) plus whatever profit the retailers make (source.)
posted by Triplanetary at 9:52 PM on April 5, 2012


There was an Interesting video in the NYT recently talking about these issues. They call it "the iPhone economy" but really has nothing to do with the iPhone. It's about the fact that sending low-wage jobs overseas pulls higher wage jobs as well, because it's easier for a manager to manage where he lives, it's easier for an engineer to make sure things are up and running correctly if he's nearby.

The result of shipping all these jobs overseas is that now the good jobs are following them.

Oh, and of course there's the fact that China offers tax breaks to companies to bring good, high skilled jobs over as well.
The US makes $150 per unit (to Apple) plus whatever profit the retailers make
The problem, though is that that $150 isn't distributed evenly to every American. Only a few rich people get it. The $8 gets divided up among low skilled laborers and their managers. The $150 goes mostly to wallstreet investors, especially now that Apple is disgorging it's $100 billion cash hoard as dividends, rather then new investments that create jobs. most of that $8 probably goes to the "99%"

How much of the money stays in the pockets of investors? Does it go back into new investments? Are those new investments in the US? Do they provide many jobs here, or are most of the jobs overseas? How much of it just goes to Bugatti Veyrons and Versace other luxury goods made overseas?

That's why you have a huge problem when you only analyze this from an "America" vs "China" vs "Other random country" framework. On paper, free trade benefits the economy of both countries. In practice, the gains are not distributed evenly.
posted by delmoi at 10:05 PM on April 5, 2012 [3 favorites]


I really don't get the absolute loathing finance people have for stocking inventory. "Time value of money" blah blah blah. There's value in being prepared for catastrophes too. And I don't mean "we don't have enough beer" catastrophes. I mean "the solar flare knocked out the power grid and we only have 1/100th the parts we need to rebuild it" catastrophes. -- DU
You can calculate the probability of various catastrophes, and figure out what the cost * probability would be in various circumstances. You can also take a gamble. If the probability is low for a given quarter, just ignore the risk, pay yourself well, and when it happens? Oh well, too bad for the shareholders, I guess I'll go cry about it on my yacht, while looking for another overpriced CEO gig from one of my friends. (Just look at TEPCO, for example -- all the money that all the employees made over 50 years they get to keep, despite the fact the cleanup costs are going to be higher)
By moving its supply chain to China, Apple has minimized the risk of the kind of problem described in the article. It may not be "robust" by your definition, but it is safe. And also very profitable. -- jedicus
Right, you have everything in one place, it works great. The top designers and engineers at apple can just hop on a jet to china when they need to, and if you keep everything else there.

If there is a catastrophe, hey, it's not even their problem, it's foxconns problem. And if one of foxconn's competitors figures out a better, cheaper way to make the phones they can move.

That's the real cost savings there, you have multiple different companies who can provide the same commodity service of constructing the phones. So if you have a problem with one, you just use another. If apple built their own factories, then they'd be taking on all the risk DU is talking about. By outsourcing it, there's no problem.
This seems like an inadequate reason/explanation/excuse. The supply chain in China didn't materialize in an instant. Manufacturing didn't pack up and move there overnight. -- Western Infidels
I think we should be clear on one thing here. The supply chain didn't slowly come up over time because western imperialists thought would be a great cheap source of labor. It came about because the Chinese government wanted it. They built the infrastructure needed. The entire city of Shenzhen, for example.

The Auto bailouts are an example of the US government doing the same thing, and short sighted republicans are bashing it. The Chinese government doesn't have to worry about republicans. That's the difference.

On the other hand, for decades they were fucked over by bad economic policy from Mao. So it's not like you can say democracy is a bad thing, and we should be run by technocrats. Competent, unaccountable technocrats can do great things, incompetent technocrats can fuck things up by an enormous amount and no one can stop them.

Democracy is a more stable system. The question is whether or not the Chinese government will be able to hold itself accountable if they start screwing up. It won't be easy.
You say this as though the answer is so clear. But in any context other than business the answer is clearly the other way. "We decided not to give our kid braces because what if he died before he was old enough to mate?" -- DU
The mistake here is assuming that corporate titans have the same relationship to the American people as a parent to their children. They don't care, and why should they? Why should a corporate CEO care more about an American job then one in China? A good job in a 3rd world country does far more social good then a good job in a western country.

And finally, you have to actually do the math. You can't just say "bad things might happen!" You have to calculate the actual probability of said bad thing happening (or at least an upper bound), then you calculate the actual cost.
posted by delmoi at 10:25 PM on April 5, 2012 [3 favorites]


The author says it's not about jobs we don't want, but rather delays in the suppy chain. Doesn't he realize that minimizing delays in the supply chain makes them jobs that we don't want?
posted by benbenson at 4:40 AM on April 6, 2012


The author says it's not about jobs we don't want, but rather delays in the suppy chain. Doesn't he realize that minimizing delays in the supply chain makes them jobs that we don't want?

Not really sure what you're talking about. I think most people would rather work in a factory then at McD's. In general, the greater the inefficiency, the better the job of the person who's job it is to mitigate.

So for example, remember the early 90s, when anyone who knew photoshop and HTML could charge thousands of dollars for PS work and a few HTML files?

That happened because of market inefficiencies.

Getting pages on the web was massively inefficient for people who hadn't already learned the tech on their own (they would have to buy PS, teach themselves that and HTML, learn how to post things to servers, etc, etc, etc in order to have a web page). So even though it might only take the skilled designer a few hours of work, they save the business owner hundreds of hours of effort it would take to get up to speed (if it was even possible)

On the other hand, nowadays there are lots of templates. Small Businesses can just setup facebook/G+ accounts and actually end up with something that's much more engaging and interactive then a site from some HTML coder from '97. So, probably people can't get jobs that easy anymore, and web designers have to work a lot harder, or have much better skills in order to make that kind of money.
posted by delmoi at 6:16 AM on April 7, 2012


Interesting article. Although I have to question the use of somehow bringing back iPad manufacturing jobs to the US. China makes $8 on every iPad made. The US makes $150 per unit (to Apple) plus whatever profit the retailers make (source.)

Read one more source up the line. Google Translated:
SAN FRANCISCO, February 14 morning news, according to the Korea Central Daily News reported that Apple every sale iPads a profit of $ 150, equivalent to 30% of the price of Chinese workers can only gain $ 8 income, equivalent to 1.6% of the selling price.

That's Chinese workers. You can bet that Foxconn and the Chinese government, not to mention transport costs, claim their own shares of the pie.
posted by JHarris at 12:24 PM on April 7, 2012


« Older "This is not the atmosphere I grew up with."   |   "I am very real" Newer »


This thread has been archived and is closed to new comments



Post