Sunday, March 23, 2003

Sunday, March 16, 2003

Saturday, March 01, 2003

Book: Eliyahu Goldratt "The Goal - A Process of Ongoing Improvement"

I'm a bit undecided if this is the work of a genius or of a dazzler. The book is about common problems in production plants and how to solve them. Many results can be applied not only to production plants, but to project management in general. The style of a novel makes if fun and entertaining to read: "A survey of the reading habits of managers found that though they buy books by the likes of Tom Peters for display purposes, the one management book they have actually read from cover to cover is The Goal." (The Economist).

One of the basic principles the book promotes is "if you really want to help someone, don't try to answer his questions or solve his problems, but challenge his potentially wrong belief system and mental models by asking questions" (deductive learning, Sokratic method). This assumes that most of us believe many many things just because we've grown up with other people who believe the same and we've learned these things. We accepted most things as if we were unable to think. We don't question these convictions because this would be hard mental work and would mean insecurity. To most people this is really frightening, and this was one of the reasons why Sokrates was put to death.

The first question the book raises is "what is the goal of a production plant (or generally any company)?". Is it maximizing efficiency of ressources, i.e. machines and workers? Is it reducing unit cost? Cheaper replenishment? Is it power, marketshare, creating safe jobs, quality, customer satisfaction, technical advancements, turnover? Well, the ultimate goal is earning money. That sounds pretty obvious, but Goldratt argues that many common controlling instruments and management efforts in companies focus on all those other things, and this actually often drives a company away from its ultimate goal of earning money. E.g. local maximization of each production unit efficiency of the plant will drive up stock of intermediate products, which will have negative effects on earnings of the whole plant. Therefore the plant may actually be more productive if certain workers or machines are sometimes idle... The ultimate goal of "earning money" is then expressed by three indicators: net return, return on investment and cash flow.

Next question: how are these three monetary indicators related to the "real" things going on in a production plant? Goldratt's answer: the three key indicators for the productivity of a plant are throughput, operating expenses and stocks. Goldratt's definitions of these three terms are a bit special:
-> Throughput is the amount of money per time unit which is earned by system through sales
-> Stocks is the amout of money which has been invested into the system for the purpose of being sold
-> Operating Expenses is the amount of money used by the system to convert stocks into throughput
According to Goldratt, these are the key factors that determine net return, ROI and cash flow.

In a balanced production plant the capacity of each single production unit is equal to the market demand of the produced goods. Goldratt argues the more balanced a plant becomes, the more it approaches bankruptcy. The reason for this are interdependent events in combination with statistical fluctuations.

The effect of interdependency is that the weakest link in a chain determines the strength of the whole chain. The bottleneck of the plant determines its whole throughput. Statistical fluctuations make this even worse, i.e. the chain throughput is even less than the throughput of the weakest link. Lets assume each production unit is operating with the same throughput as the weakest unit in the chain. If a unit in the chain now has temporarily less throughput than the rest, this immediately slows down the whole chain. On the other hand, if a unit in the chain has temporarily more throughput, this does not help, except when all subsequent units in the chain have "spare capacity". Because of interdependency, statistical fluctuations don't compensate, they accumulate, except if all units are ordered by capacity and the unit with the most capacity is the last unit of the chain. Unfortunately it's generally not possible to reorder units in production chains.

... eventually to be continued ...

Final question: what makes a good leader?
Answer: leaders should be able to answer these three simple questions: what should we change, how should the replacement look like, and how do we get the change started.

Simple Design is Safer

"The Space Shuttle design is too brittle and too complex to allow for the kind of incremental improvements that are the cornerstone of safety. Boeing can continuously improve and rejuvenate the 747, originally built in 1969, because the underlying design is simple enough to understand and tweak. By contrast, the Shuttle is almost too complex to touch. Hell, it's almost too complex to fly."