Gemba Keiei Chapter 10: The Misconception that Mass Production is Cheaper

“Another example of a misconception that becomes common sense,” Taiichi Ohno begins “is the idea that mass production is cheaper so that by the same logic low volume production must cost more. I have seen many factories but very few examples where mass production reduced the cost. In most cases increasing production volumes resulted in higher costs.”

Very interesting. I like to think that I ‘get it’ when it comes to TPS and kaizen, but I struggled with that Ohno is saying here. But let’s continue.

Ohno doesn’t argue with the fact that if your capacity is 10,000 per day then the piece price will be less on a day when you have orders for 10,000 compared to a day when you have orders for 8,000. Taiichi Ohno argues that when you attempt to go beyond 10,000 per day the cost actually gets higher.

The maximum capacity of a workplace is finite. Taking the example of a press, the lowest cost to produce may be at 1,000 units per hour. Increasing production to 1,200 per hour with these assets may not result in a lower cost. A 20% increase doesn’t justify adding equipment, so you work overtime. Overtime premiums can be 30% to 40% (says Ohno). The union will complain if you work too much overtime, so you need to buy another machine, which is initially run at less than full capacity until you have more orders, increasing cost.

The comment about the union needs some explanation. In Japan very few companies run a graveyard shift. Most run on one or two eight hour shifts. Part of this is because in the cities the trains and public transportation don’t run 24 hours so you can’t get people to and from the factories.

Another reason is that many factories employ housewives on a seasonal or temporary basis. They expect to be able to go home and take care of their school age children in the evenings. Too much overtime and it starts to affect the quality of life, and the labor union is there to keep that from happening.

Back to the misconception. This is the case when your volumes are increasing, says Ohno, but what about when your volume are declining? You can do SMED to kaizen your changeover times from 60 minutes to 10 minutes. This allows you to do multiple changeovers per hour and run smaller quantities of a wider variety of products. If you can produce 50 pieces per hour of product A (the example he gives for monthly demand), then change quickly and produce B, C, D, E, F, etc. then you can produce a mix of products at a low cost. Although he doesn’t name it, he is talking about Heijunka.

Ohno argues that since you did kaizen to cut changeover times, the 50 minutes of production time (60 minute changeover before kaizen – 10 minute changeover after kaizen = 50 minutes press time) actually costs nothing. This is an accounting headache, says Ohno, but just don’t use this time to run a larger lot of parts in the false belief that this lowers cost.

Ohno concludes: “As long as there is a perception in the market that a low volume product should cost more and you can charge a premium for this. If you can produce it cheaply through SMED, take advantage of this profitably cost accounting may lead you to believe that a certain product is profitable or that it is not. Products that could be sold profitably are not, and others are sold at a loss due to this misconception. This is a widespread phenomenon that is present no only in the automotive industry but in others.”