Balanced Production Operations - How Desirable?

Imagine you are the production manager of a multimillion dollar factory. To avoid waste, all the machines that process materials from raw state to finished state, are designed such that their capacities are balanced. They have sufficient capacity to meet the demand placed on the factory by the market: no more. As a result of this equality of capacities, balanced by market demand, all the machines run at full capacity all the time. Your utilisation, productivity and efficiency charts are any production manager's dream - at a constant 100%.

What is wrong with this picture? The only thing wrong with this picture is that it ignores the implication two vital characteristics of real operations. Any production facility (indeed any organisation at all) has inherent dependencies. This means that what happens in one area or machine or function affects what happens in another and when. Two sources of dependency exist in production. These are structural or sequential dependency (as a result of the fact that one process must occur before another) and resource dependency (one resource is scheduled to do more than one task).

The second fact ignored by the balanced operations design is the reality and pervasiveness of Murphy's Law: uncertainty exists. Negative uncertainty manifests itself in the form of cancelled customer orders, machines downtimes, setups taking longer than scheduled, employee unavailability (due to illness or plain truancy) etc.

What is the impact of the combination of these two realities? Consider a balanced production line capable of producing either of two products, A or B. Let the line consist of four machines w, x, y, z. Assume we are currently producing A and are scheduled to change over to B when the planned quantity of A is done. If machine x develops a problem and it takes one hour to fix, then after resumption of production,
i. some wip would have accumulated in front of x
ii. y and z will initially have nothing to process
iii. this means there is a gap before the required quantity of A is completed
iv. and finally production of B cannot begin on schedule
The disruption caused by temporary unavailability of a resource inevitably leads to irrecoverable delays.
The resources propagate disruptions from one product to the next (unavailability of x caused delays in completing production of A and B). Conversely too, products also propagate disruptions from one resource to the other (assume A is input for producing some other item on some other machines, those machines cannot begin working until A is available).

Because of the two phenomena inherent in operations, balanced capacities are impossible. To operate smoothly, one or more resources will be forced to operate below capacity in order to have some protection against uncertainty.

The lesson here is that to focus on balancing capacities is to guarantee disruptions in delivery of products to your customers. Accumulation of WIP throughout the system is also guaranteed. By Little's Law, this translates to long production lead times.

Attention is better expended on maintaining undisrupted flow to the customer. For this, you must deliberately unbalance your resource capacities.