Reflections On Eliyahu Goldratt's 'The Goal'

This blog entry summarizes and provides comments on the book The Goal, by Eliyahu Goldratt. The Goal tells the story of Alex, a fictitious manufacturing plant president. Alex is on the verge of losing his job for having an unproductive plant. Alex is constantly reacting to emergencies, he has a large backlog of customer orders, long lead times and significant amounts of work-in-process inventory. Alex's old friend, Jonah - a former physics teacher turned process consultant - asks Alex a series of probing questions that lead him to an epiphany which helps turn the plant around.

Alex wants to improve his plant's productivity. Jonah responds by posing a question to Alex: What is his manufacturing plant's goal? After determining that productivity in the abstract is not the goal and sifting through a series of cost accounting metrics, Alex finally realizes that the goal of his plant and the overall company is simple: to make money. Jonah posits that each and every activity that occurs in the plant should be measured by one metric: does that activity advance the company toward its goal. If an activity does advance the company toward its goal, then it is productive (even if it does not appear to be productive in the abstract). For example, having an employee sit idle by a machine could be productive if that machine is a bottle-neck operation for the company and redeployment to another task at a non-bottleck operation would result in a small loss of set-up time for the bottleneck machine when the current process is complete.

The Goal describes a "theory of constraints," which focuses on identifying bottleneck operations in a manufacturing process, increasing throughput and identifying statistical fluctuations and dependent events in the process. Statistical fluctuations and dependent events can have a subtle impact on the productivity of a manufacturing process.

For example, assume that Step A is a dependent event with respect to Step B (i.e. Step A must be completed before Step B can begin). Assume the average time to complete Step A is 10 minutes per batch. If there is statistical fluctuation with respect to Step A, then the actual time to complete Step A would vary by batch. For example, the actual time to complete Step A may vary from 7 minutes to 13 minutes. The combination of these two factors will create a constraint on Step B. Assume...