[The Pull Principle] On Supermarkets and Balance Weights

This entry is part 1 of 5 in the series Autonomy at Work: The Pull Principle

Piggly Wiggly store 1916, the first supermarket in the world

When Clarence Saunders built the very first supermarket, he did not know that he invented a principle that changed the world. It was 1916, when he was annoyed of the fact, that grocery stores – think of small corner shops and the like – only gave one person at a time the chance to order their products. An assistant would then go into the storage room and fetch the goods. This was a very slow way of shopping, and it was also very labour-intensive because of the assistant’s payroll costs.

Saunders thought there must be a better way. In his stores he called supermarkets, several customers could select products from the shelves at the same time. Instead of a grocery assistant there was a cashier where customers could pay their selected products. That approach was much faster for the customer. Also, product prizes decreased, because there was much less need of labour to serve the same amount of customers. Needless to say, supermarkets were and still are a huge financial success.

Almost 40 years later, a young Japanese manager called Taichi Ohno was very impressed by the concept of a supermarket. He was an executive at the Toyota Motor Corporation, a small automobile production company in Japan these days. The principle behind supermarkets solved a problem for him, which was equally simple and difficult: the balance weight problem.


A modern-day balance weight

You might know balance weights from your own car. After your car gets new tires fit, the mechanic has to balance the wheel afterwards. Without balanced wheels, the faster you’d drive the more the wheels would clatter and eventually would brake your wheel suspension. To prevent this, the mechanic attaches as many little pieces of lead or zinc to the wheel, i.e. the balance weights, as it takes to have the wheel balanced, only a few grams in total.

Now the problem is, you can’t know in advance how many balance weights you’ll need to balance a wheel. At Toyota, they used 5 different kinds of weights, and they were never able to predict, if they would use, say, more 1 gram weights than 2 gram weights on a given day. They were confronted with a complex system, where you can’t predict the future and where only in retrospect pattern emerges and eventually makes sense. Complex systems are unpredictable.

In the early 1950s, Toyota built up to 40,000 cars per year, so they had to use quite a few balancing weights. But still, no biggy, right? Worst case would be that only one balancing weight would be used (very unlikely, but just to make the point here). In this case, the mechanics would have needed the maximum amount of balancing weights in every category, because he was never able to predict today’s kind of balancing weight. So, to prevent the worst case, all Toyota had to do was to produce a large amount of balancing weights, so that the mechanics would never run out of them.

Of course, this comes with a huge downside: For large amounts of balancing weights you need a large inventory, which produces costs not only for building and maintaining the inventory, but also for the costs of having valuable raw material lying around, useless at the moment: money tied up in stocks.

When Ohno saw the supermarket, something clicked in his mind:

“A supermarket is where a customer can get (1) what is needed, (2) at the time needed, (3) in the amount needed. […] In principle, however, the supermarket is a place where we buy according to need. […] From the supermarket we got the idea of viewing the earlier process in a production line as a kind of store. The later process (customer) goes to the earlier process (supermarket) to acquire the required parts (commodities) at the time and in the quantity needed. The earlier process immediately produces the quantity just taken (restocking the shelves).” — Taichi Ohno in Toyota Production System

This was revolutionary. From now on, the mechanic who balances wheels had 10 boxes where he kept the 5 balancing weights: 2 boxes full of 1 gram weights, 2 boxes full of 2 gram weights, and so on. Whenever he used up all weights of one of the boxes, he carried that box to the earlier process, in this case the place where they made the balancing weights. Attached to the box was a card, indicating the kind of balancing weights and the amount needed for the box.

The arrival of this box was the trigger for the balancing weight maker to produce just enough weights to fill up this box. While he was working on the balancing weights, the mechanic could use the second box to continue his work. When the balancing weight maker filled up the first box, he delivered it to the mechanic, and the cycle closed.

Ohno called this system kanban (lowercase k), after the Japanese name for “signal card”, like the card attached to the boxes.

“Kanban was introduced to manage the balance weight problem, one of the most difficult processes in automobile production. […] As a result, inventories of the five weights were kept constant and, eventually, reduced drastically.” — Taichi Ohno in Toyota Production System

This was the beginning of the pull principle at Toyota – and it changed everything. In fact, it was worked so well for the balance weight problem, that Ohno installed it almost everwhere in his company’s automobile production. For Ohno, the Toyota Production System and kanban were the same, and both were a pull system.

“Let’s look at this production flow in reverse: a later process goes to an earlier process to pick up only the right part in the quantity needed at the exact time needed. In this case, wouldn’t it be logical for the earlier process to make only the number of parts withdrawn? As far as communication between the many processes is concerned, wouldn’t it be sufficient to clearly indicate what and how many are needed?
We will call this means of indication kanban … and circulate it between each of the processes to control the amount of production – that is, the amount needed. This was the beginning of the idea.” — Taichi Ohno in Toyota Production System

The system was called a pull system, because the earlier process didn’t push its goods to the later process anymore, but the later process pulled the goods from the earlier process only if there was a need for it.

Pull systems are actually markets, compared to push systems with their hierarchies. The huge benefit of pull systems is their strength when it comes to reallocation. Push systems just suck at that:

“Hierarchies are very good at applying resources – laying out plans, sequencing activities, and meeting deadlines – but they’re lousy at allocating resources – or, more specifically, at reallocating resources from old strategies to new strategies.” — Gary Hamel in The Future of Management

An impressive example of the pull system can be seen in every city. The bigger the city, the more impressive the example of pull:

“Markets are capable of solving highly complex allocational problems. At any point in time, New York City, a metropolis of more than 8 million inhabitants, has a scant three-day supply of food. In a sense, the city is always close to famine.” — Gary Hamel in The Future of Management

But NYC never had a famine. There’s no mayor with the authority to direct the supplies of food to every citizen. The market takes care of that, without central control.

For Toyota, adopting the pull principle from supermarkets became quite successful. Today, Toyota is the biggest car manufacturer in the world, with a production capacity of almost 9 million cars per year.

Series Navigation[The Pull Principle] People Signing Up for Work >>
This entry was posted in General and tagged , . Bookmark the permalink.

About Bernd Schiffer

Bernd Schiffer is consultant, trainer and coach for Agile Software Development in Melbourne, Australia. Learn more about him on his personal homepage, have a look at his company Bold Mover, or contact him on Twitter, Google+, Facebook, XING or LinkedIn.

Leave a Reply

Your email address will not be published. Required fields are marked *