Inspecting one at a time is quicker?

[johnnybegood]

I am introducing lean concept and pull system to the production line. On the production floor I observed Inspectors would pick up 5 or 10 pieces and lining them up on the table and inspecting them all at once. According to them it is faster. Beside that they claim that by lining 5 pieces, the 'odd' one can be easily identified. How can I convince them that inspecting one at a time is quicker and took less work than inspecting in batch?

[Tim Folkerts]

How about a trial to compare the two approaches? If you have evidence, it makes your arguments much more powerful.

Have a few inspectors try each method and measure how long it takes. While you are at it, check the effectiveness of the two approaches. Include a few bad parts to see if they are better at catching them if they inspect in batches or individually.

Without knowing any more, I think it is quite possible that the inspectors are correct. It is often easier to spot a bad item when it is compared to a good item, and it is often quicker to do a process several times in a row rather than one at a time.


Of course, any test that involves judgement has the possibility of being biased. If the inspectors THINK (or hope) they are more effective when they check several at once, it is quite possible that they will actually BE more effective because they try harder.

Tim F

[johnnybegood]

I read on an article on Lean...according to Taiichi Ohno that inspecting one at a time is quicker and took less work than inspecting in batch. Taiichi Ohno is one of the gurus in TPS. I am very sure he has all the prove.

[Wes Bucey]

I presume from the wording of the original post we are talking of a visual "attribute" inspection versus an instrument or gage based "characteristic" inspection.

If so, the inspectors are looking for anomalies. Usually, such a process provides a "gold standard" for comparison. Perhaps the inspectors ARE correct in being able to spot anomalies faster among a group versus one-at-a-time against the gold standard.

I would favor an "experiment" to confirm either hypothesis (one versus many), but I would wonder about removing conscious or unconscious bias from the test.

[Claes Gefvenberg]

Originally Posted by johnnybegood

I am introducing lean concept and pull system to the production line.

That sounds interesting. Will you provide us with information about the progress as you go along? I for one would love to hear the tale.

Originally Posted by johnnybegood

On the production floor I observed Inspectors would pick up 5 or 10 pieces and lining them up on the table and inspecting them all at once. According to them it is faster.

For all I know they could be right, but there are other aspects. For instance:

If you inspect the parts one by one, you will probably find a defect sooner and be able to correct it at once, as opposed to producing those four or nine extra parts, possibly with the same defect.

Besides, if we are talking about lean: Could you let the operators do the inspection?

/Claes

[Bev D]

So it all depends on the actual situation.
For certain types of characteristics that are visually apparent, the soldier method IS faster in detecting the odd man out than inspecting each pice individually. However, there are many characteristics that must be measured individually and so I've typically used solder inspection as only ONE of the inspections that are performed. (and this is typically with higher volume production AND it's a FINAL inspection, not in process)

using poke yoke inspection devices that 'test' each device as it's made (single piece flow is faster and more effective overall and can and should be done in process - real time feedback and limitation of defective product created)

each approach is viable, but which one is best depends on the situation

[Wes Bucey]

Depending on how the product is made (by automated process or hand work), most lean enthusiasts would consider these solutions (ALL BY OPERATOR WHO MADE PIECE):

• automated process - pick 5 in a row every 100 pieces (or any other number which makes sense) and chart critical characteristics to determine process is in control.
• hand work - worker checks EACH piece upon completion against a "gold standard" to assure it meets requirements BEFORE going back to make next piece.

No one of my acquaintances would consider independent 3rd party inspections as part of a "lean" operation because of the lag time between production and inspection and lack of serialization (i.e. "When" did process go awry?)