Processes & Number of Steps

[TacitBlue]

Is there a rule of thumb of how many steps a process should not exceed?

I know its dependent on the process, but ideally is 20 steps, for example, far too many?

We are running into issues where process maps are becoming a nightmare, and I am thinking that we need to seriously rethink our process creation and approach.

Thanks,

 

[Miner]

There is no rule. The number of process steps is whatever is required for the process to correctly function. If, however, you mean how many process steps one operator can handle, that is a different question.

Regarding your process maps, you may be breaking them down into too much detail. Consider the following structure:

• Process Step - A collection of related tasks combined for one or more reasons (e.g., performed on same machine or within same work cell). List this level on your process map.
• Operation - A single task performed within a process step. Do not include this in your process map but do break this out in your PFMEA when considering potential failure modes and causes.

 

[John Predmore]

If @TacitBlue is asking how many steps a worker can reliably remember, I remembered an 'unwritten' rule about human memory being able to recall 7 different things. Maybe that is not the question you asked, but let me share the interesting trivia I discovered in a search.

It turns out the foundation of that written rule was a 1956 scholarly paper, saying the ideal number is 7. Wikipedia identifies this as Miller's Law. I am not a biologist (tongue-in-cheek), but I recognize that human comprehension and ability to recall diminishes as complexity increases. It appears to me the OP wants to advocate 'simpler is better', and perhaps this citation provides evidentiary weight to the argument for reduced complexity. I believe a quality engineer has a valid professional interest in preventing mistakes arising from memory failings.

 

[Bev D]

Miner is of course correct - there is no rule.
There is an approach in Toyota production system (aka lean) that organizes work into time intervals that match takt time (except batch processes which are treated differently. In this case there are typically only enough individual steps to last about 1 minute. The whole idea of short steps that the operator can memorize adn instill in muscle memory short intervals substantially reducing errors

 

[Miner]

If @TacitBlue is asking how many steps a worker can reliably remember, I remembered an 'unwritten' rule about human memory being able to recall 7 different things. Maybe that is not the question you asked, but let me share the interesting trivia I discovered in a search.

It turns out the foundation of that written rule was a 1956 scholarly paper, saying the ideal number is 7. Wikipedia identifies this as Miller's Law. I am not a biologist (tongue-in-cheek), but I recognize that human comprehension and ability to recall diminishes as complexity increases. It appears to me the OP wants to advocate 'simpler is better', and perhaps this citation provides evidentiary weight to the argument for reduced complexity. I believe a quality engineer has a valid professional interest in preventing mistakes arising from memory failings.

You are thinking of "Working memory", which Miller believed had a capacity of 7 "chunks" of data. However, for some people, 5 chunks of data is a more realistic estimate. Also, working memory is diminished when trying to absorb new information versus trying to memorize static data.