Casco Development > ShopVue Blog > Bottlenecks in the Factory: An Overview
Bottlenecks in the Factory: An Overview of the Problem 

By Morgan Witthoft, Senior Consultant, Casco Development

Bottlenecks in manufacturing hurt your business in several ways.

Bottlenecks increase your lead time. From the date the customer order is taken, you have to allow extra time for the WIP to sit around at the bottleneck operation.

Bottlenecks increase the amount of WIP on the shop floor, leading to several problems. Expensive WIP ties up your money. It gobbles up storage space and creates clutter and confusion. Create a big enough clutter and it becomes difficult to find any particular order. Operators who can't find the desired order will perform jobs in the wrong sequence, working on a non-urgent job while there is a past due order somewhere in that heap of stuff waiting at the work center.

Convincing, eh? So let's get rid of our bottlenecks.

Before you get rid of your bottlenecks, you have to find them. To find them, you have to define measurable criteria; then you can start measuring. There are several ways you can measure, and each one will deliver a list of suspects. A workpoint that shows up on the suspect list over and over is likely to be your bottleneck. 

To start, we must define bottlenecks.

APICS weighs in with:

A facility, function, department, or resource whose capacity is less than the demand placed upon it. For example, a bottleneck machine or work center exists where jobs are processed at a slower rate than they are demanded.

For shop floor execution purposes we'll substitute "workpoint" for "facility…resource". Still, this definition is unsatisfying. My inner mathematician rebels at the last phrase

… jobs are processed at a slower rate than they are demanded.

Oh no! That means, as time passes, the backlog at your bottleneck will approach infinite size!


Demand = 10 pcs/hr

Processing rate at bottleneck = 8 pcs/hr

Material piled up awaiting processing at bottleneck after 1 hr = 2 pcs

after 2 hrs = 4 pcs

after 24 hrs = 48 pcs

after a week = 336 pcs

after a year = 17532 pcs (that's an astronomical year, not a fiscal year)

More likely the demand backs off periodically, rather than stack your WIP up to the roof – but now the definition collapses, because the bottleneck is now processing jobs just as fast as they are demanded. You could try to patch it up by saying

… jobs are processed at a slower rate than you wish you could demand them

but now you're defining your math in terms of your hopes.

Much better, from Wikipedia:

A bottleneck is a phenomenon where the performance or capacity of an entire system is limited by a single or limited number of components or resources.

You want to make 10,000 widgets this month? Somewhere in the factory there's a workpoint that can only make 9,000; so 9,000 is what you're going to get. Your challenge is to pick out which workpoint that was. The next articles describe different computations that can identify possible bottlenecks.