The Beer Game was first developed (by Jay Forrester) at the Massachusetts Institute of Technology in the nineteen sixties. It is a simulation I have used in workshops to show that by making rational decisions, we can become victims of the system we are trapped in.
The beer game works along similar ideas to chaos theory where a butterfly flapping its wings may cause a tornado somewhere else:
"The flapping of a single butterfly's wing today produces a tiny change in the state of the atmosphere. Over a period of time, what the atmosphere actually does diverges from what it would have done. So, in a month's time, a tornado that would have devastated the Indonesian coast doesn't happen. Or maybe one that wasn't going to happen, does." (Ian Stewart, Does God Play Dice? The Mathematics of Chaos, pg. 141)
Playing the beer game in a workshop environment is fun and participants are amazed at the results.
We can't play it together but I have found a computer simulation of the beer game prepared by major consultancy firm AT Kearney and Matthew Forrester on the MIT website (Beer Game Simulation). If you are curious and have some time or if you manage a business with stock I recommend that you give it a go.
There is a full description of the Beer Game in Peter Senge's seminal book "The Fifth Discipline: The Art & Practice of the Learning Organisation" which was my first introduction to both the beer game and systems thinking about fifteen years ago.
You can buy the book from Amazon Amazon.co.uk or Amazon.com (affiliate links)
I must do a full review although the book has been revised since my battered, much thumbed and read copy was printed.
A Simple Explanation Of Beer Game
Imagine a three stage supply chain, retailer, wholesaler and factory.
Each holds stock (inventory) to cope with minor fluctuations in short term demand and any unexpected delays in new supply coming in.
There is a time delay between ordering new beer and receiving it and this is a key assumption of the beer game but is commonly true in stock situations.
Each business wants to maximise profit and in particular wants to meet customer demand when it arises.
The retailer starts seeing an increase in demand from customers so naturally orders more from the wholesaler. The retailer doesn't know but the beer is mentioned in a new rock song.
The wholesaler sees extra demand from all the retailers in their area and is delighted to supply the extra from stock and orders more from the factory.
The retailer continues to sell more, so decides to not just order at the replenishment level but to increase their stock, just in case they sell out.
The wholesaler sees this surge in demand and again supplies from stock which is very low and will run out soon. The wholesaler orders more from the factory.
At this stage the beer game is accelerating demand because both retailer and wholesaler want to hold more stock as well as replenish to their previous stock levels.
The retailer continues to sell more at the new level but doesn't receive the expected supply from the wholesaler who has now run out. This worries the retailer who decides that "to be on the safe side" they will increase their order some more.
This goes on through some more cycles, where first the supply chain is stripped of all the beer but then, as supplies come through, the retailer is overwhelmed. Everyone suddenly has far too much.
Now the beer game goes into reverse.
The retailer stops ordering but continues to sell, the wholesaler stops selling and ordering and with massive amounts of cash tied up in stock/inventory, starts worrying about their survival.
The factory sees no orders coming through from the wholesalers and when it makes enquiries, it discovers that there won't be any orders at current rates for months. Layoffs and plant closures begin.
The story of the beer game works best in simulations when you are making the decisions but the description in "The Fifth Discipline" describes the logic much better than I have.
The trigger in the beer game was a doubling of demand because a pop song named the beer but the system should have been easily able to cope.
The system couldn't because it was flawed because of the time delays in the process and the mis-information that was passed backwards along the supply chain.
The beer game exaggerated the change in demand, players couldn't see what was really happening to true demand and because of the time delay had to make decisions in a vacuum.
An Example From My Own Past
I have told the story a few times of when I was Finance Director of a manufacturing business in the West Midlands in the early nineties and we hit the recession which particularly affected the building industry which was the sector we served.
I have never seen big monthly profits (of over £100k) turn into masses of red ink so quickly. We were hit on three sides:
- Our sales volumes dropped sharply
- We found ourselves caught in a price war on nearly all our products and our prices were cut and cut again. Our competitors say we caused it, we said they did.
- Our costs increased as our production volumes shrank much faster than our sales.
It's that last factor that I want to explain here because we had our own beer game happening.
Our stock policy at that time was to hold say 4 weeks worth of stock based on an average weekly sales rate calculated over the last three calendar months.
So imagine a product which had weekly sales of 1,000 units based on selling 13,000 over the previous 13 weeks and therefore had a stock target of 4,000 units.
We had previously thought that this self adjusting stock policy was a clever and easy way for us to adjust our stocks in line with recent demand.
Month 1
Then one month we sold an average of 800 per week but we made 1,000 each week.
So at the end of the month we had stock of 4,800 (= 4,000+ 4,000 -3,200)
Our sales over the last 13 weeks reduced to 938 (= 13000 - 4000 + 3200)
Our new stock target was 3,752 (4 x 938)
So our new production target for the month was only 2704 (based on expecting to sell 3,752 in the next 4 weeks and reducing our stock from the actual 4,800 to the new target of 3,752).
Our production was therefore reduced from 4,000 to 2,704, a 32.4% reduction when sales had only reduced by 20%.
Month 2
Sales carried on at 800 per week (3,200 in total in the four week period) which unfortunately was lower than the production forecast had assumed of 3752.
Stock was now 5,352 (4,800+3,752-3,200)
Sales over the last 13 weeks were now 11,400 (13,000 - 4,000 - 4,000 + 3,200 + 3,200), our expected average weekly sales used for planning was 877 (based on 11,400 / 13 weeks).
This meant that our new stock target was 4 weeks at 877 = 3,508
Our new production target was the expected sales less the stock reduction = 3,508 -(5,352-3,508) = 1,664.
In two months, our expected production of this product (and all the others) was down from 4,000 to 2,704 to 1,664 - a drop of 58% from normal levels while sales were down 20%
Can you see the exaggerated effect that the Beer Game was having and we thought we were behaving rationally!
It turned out that sales didn't drop by 20%, our customers, the electrical wholesalers were doing just what we were doing. They were reacting to reduced orders from their customers by cutting back on the stock they carried.
But we didn't know that at the time.
We had expected a reduction and official building industry forecasts indicated that new builds were going to be down by around 10% that year.
If the industry was down by 10% but we were down by 20%, it must be because of those ****** we were competing against who had slashed prices and were stealing our customers.
We had to protect what we had so we reduced our prices.
When our customers had finished their de-stocking, they needed to re-order to meet ongoing demand but our stocks were then so low we couldn't cope. Our service suffered.
Our production planning system responded but guess what, it over-responded and we went from short time working and lay-offs to extended shifts and high overtime premiums.
Our employees must have thought the management were idiots when the production planning and stock system had seemed so clever.
If Only We Had Understood The Beer Game
In 1991 I hadn't read "The Fifth Discipline". I had never heard of the beer game.
But if I had, it could have saved a lot of money and a lot of unpleasant times as we struggled and failed to cope with the extreme fluctuations we were seeing in our production planning and stock control system.
The upside was that the feast or famine problem became the central issue for the company and responsibility was taken away from the Operations Director and passed to me.
Since then, I have been fascinated by systems thinking and stock control issues and it was this fascination which introduced me to systems thinking and the beer game.
The automatic updating of stock control minimum and maximums was replaced with fixed limits and a sense check based on thirteen week trends. The stock limits were changed when we thought it was appropriate based on our knowledge of the beer game.
Stories From My Personal Life
Have you noticed what happens if you hear on the News on the radio or television that crops have failed and there could be a shortage?
These are all examples of the beer game and systems thinking exaggerating the effect of a small change.
Everyone wants to buy, shoots off to the supermarkets and buys three extra units. Demand shoots through the roof and creates a genuine shortage.
I remember in my teens, there was a sugar shortage and the supermarket shelves were cleared. We didn't have any and I've never taken sugar in my tea or coffee since.
A few years ago, we had a petrol shortage in the UK. Petrol station after petrol station was dry and those that weren't had long queues of cars with motorists prepared to wait two or three hours to make sure that their fuel tanks were full.
The authorities said that there was no shortage.
The motorists knew different because they created the shortage. Every tank was kept filled to the brim and if you drove 50 miles, it was natural to want to fill up again.
Lessons To Learn From The Beer Game
- Understand the way your "system" reacts to fluctuations and beware of automating too far. Once you understand the dynamics of the beer game and see the biases in your system, you have to use judgement rather than just calculations.
- Shorten the lead times wherever possible. One of the main causes of extreme fluctuations in the beer game is the time lag between action and reaction.
- Improve your market intelligence and share that intelligence with other parts of the chain. In the Beer Game explained in "The Fifth Discipline", market demand doubled one week because of a pop song and then stayed constant at that new level. If the players had known the true facts, rather than reacting emotionally to the shortage, their decisions would have been different and brought the system back into a steady state much faster.
It Is How You Adapt That Matters
"It's not the strongest nor most intelligent that survive. It is the one most adaptable to change." (Charles Darwin)
The beer game shows that what are apparently sensible decisions taken in isolation spread chaos throughout the system.