A little history
How we learned to make a lot, 1901–1913
A hundred and more years ago a factory was a workshop: a worker built a product from start to finish, as best he could. The American engineer Frederick Taylor proposed timing every movement of a worker with a stopwatch, finding the best way and teaching it to everybody; in 1911 he published a book about it, «The Principles of Scientific Management». Work began to be split into simple operations.
The first to do this with cars was Ransom Olds. In 1901, at his plant in Detroit, cars were put on carts and pushed from post to post: at each post the workers did one and the same operation. Output of the little Oldsmobile grew in a year from 425 to 2,500 cars — more than anyone else in America.
Henry Ford took the next step. His engineers had watched the overhead lines of the Chicago meat-packers, where carcasses rode past the workers, and in 1913, at the Highland Park plant, the Model T chassis started to ride past the workers on a rope by itself — the moving assembly line was born. Everyone did one operation, and the car came to him. Assembling a chassis went from about twelve hours to an hour and a half, and the car itself became affordable to the worker who built it.
An assembly line is a line where every section depends on the one before. While everyone works to the same beat, it works wonders. But let one section stumble — and all of them stop.
Toyota: make only what is needed, 1950s–1970s
After the war Japan's Toyota had neither Ford's money nor his huge market. Engineer Taiichi Ohno, building the Toyota Production System, bet on something else: not to push out as much as possible, but to make exactly what is needed, exactly when it is needed — «just in time». The next section takes from the previous one as many parts as it used up, and the signal is a kanban card. The stocks between sections became small.
Ohno used to say that stock is like water in a river: while there is a lot of it, you cannot see the rocks on the bottom. Take the water away — and the rocks, the real problems, are in plain sight. When in the 1980s Japanese cars began to push out American ones, the whole world rushed to study Toyota. In 1988–1990 its approach got the word lean — «lean production».
Goldratt and «The Goal», 1984
Eliyahu Goldratt (1947–2011) was an Israeli physicist. In the late 1970s he wrote a program for a friend's plant that made up the schedule of the machines — OPT. The program worked so well that Goldratt took up plants seriously and noticed: almost everywhere he went, one or two spots slowed the whole plant down, while everything else was being improved.
Explaining this to managers in formulas was hopeless, so in 1984, together with Jeff Cox, he wrote a novel — «The Goal». Its hero, plant manager Alex Rogo, has three months to save his plant from closing, and the hints come from his former physics teacher. A book about machines and stockrooms became a bestseller, millions of people read it, and its ideas got the name theory of constraints: every system has a weakest link, and that link sets the output of the whole system.
The match game
One of the key scenes of «The Goal» is the hike. Alex leads a troop of scouts that includes his son. You cannot overtake on the trail, the boys walk at the same speed on average — and still the column keeps stretching out, and the whole troop reaches the rest stop when the slowest one, Herbie, gets there. At the stop Alex makes up a game to understand what is going on.
The rules. The players sit in a row. In front of each one, except the first, there is an empty bowl. The first has a box of matches. The turn goes in order, from the first to the last:
- the player rolls a die;
- moves from his bowl into the next player's bowl as many matches as the die shows — but no more than he has. The first one's box is full, he always has enough;
- the last one moves matches «to the warehouse» — those are finished products.
A die gives 3.5 on average. So each player on average can pass 3.5 matches a round, and it seems that 3.5 should reach the warehouse every round.
A game for three. Ann (A) takes from the box, Bob (B) and Vera (V) from their bowls, Vera puts into the warehouse:
| Round | Dice A B V | A → B | B → V | V → store | Bowl B | Bowl V | Warehouse |
|---|---|---|---|---|---|---|---|
| 1 | 6 1 5 | 6 | 1 | 1 | 5 | 0 | 1 |
| 2 | 1 5 2 | 1 | 5 | 2 | 1 | 3 | 3 |
| 3 | 4 3 5 | 4 | 3 | 5 | 2 | 1 | 8 |
| 4 | 3 5 2 | 3 | 5 | 2 | 0 | 4 | 10 |
| 5 | 5 3 1 | 5 | 3 | 1 | 2 | 6 | 11 |
Look at what happened:
- round 1: Bob rolled a one — he passed Vera one match. Vera rolled a five but could pass only one: the five was lost, because you can only move what lies in your bowl;
- round 5: Vera has seven matches in her bowl but rolled a one — one went to the warehouse, six stayed in the bowl;
- over five rounds the dice of all three gave about 3.4 on average. By the average about 17 matches should have reached the warehouse, but 11 did. Another 8 got stuck in Bob's and Vera's bowls.
Nobody was lazy, the dice are fair, and yet the system made noticeably less than it «should». In this kit you will play this game on the canvas: the bowls are the machines' buffers, the matches are parts, and every machine rolls the die before every part.