One Piece Flow versus Mass Production

Great video, no nonsense approach, Methodology made simple… Looking forward to more presentations.

awesome work

Ron/Eric,

One thing I still don’t get….how does the average time to fold, stuff and seal go down in a lean one-peice flow as compared to a batch processing? Beats me !!

Still don’t get it. Where are the extra* picking,puting etc?

Wow! Gotcha! That’s neat..didn’t strike me..Thanks a lot.

Am back. You really made me think. What if I throw in 3 operators as someone suggested? The picking up & sitting down of envelopes would still happen in a OPF..won’t it?

Ron,

That was terrific:-)) Best presentation so far I have seen or experienced.

For most people this will look unreal and as long as they don’t experience it themselves they probably won’t change.

STABILITY is the main issue -as you said- and that is definitely true to my experience.

Keep up the great work.

Cheers,

Ralf

this is wrong
it’s not faster there’s no simultaneous processing. the only 2 obvious reasons that i can think of why it’s faster doing it with the one piece flow method:
1. they’re not identical in terms of positioning the envelopes, doing the actual operations of glueing, etc.
2. learning curve. the first method allowed you to learn how to do it and by the time you did the second one you’re already better at it.

Ron,
if you have one person doing the folding, one person doing the glueing, and so forth then yes you can benefit from one piece flow (with all the experiment conditions being the same in both methods and no learning as a result of doing the first method). but not in this case.
please explain why one piece flow is better in your opinion because i’m trying to understand all this from the industrial engineering perspective.

Ron,

Very very great thought-provoking video. It was a very good introduction to “one piece flow” processes for the beginner that I am !

Here are my 0.02 on the 3 operators line : Let’s break the process in
Folding (f) +
Transfer (t1) +
Stuffing (su) +
Transfer (t2) +
Sealing (sl) +
Transfer (t3) +
Stamping (st).
Ok, this makes 4 operators.

Let’s assume that the processing times (f, su, sl, st) don’t change between the two methods.

So what about our ten envelopes ?
With a lot size of 10, we have the batch process time : 10f + t1 + 10su+ t2 + 10sl + t3 + 10st

And as for the one piece flow process (if we assume from Eric’s datas that st<sl<su> f) ? Whereas the batch time formula doesn’t change, the one piece flow becomes f + 10t1 + su + t2 + sl + t3 + st, to be considered if the guy in charge of stuffing lives oversea, or at least, next door 🙂

From another p.o.v. transferring 10 envelopes sounds as painful as transferring 1 which is, if we count total number of operations (and not overall time) we have
batch : 10f + 10su + 10sl + 10st + t1 + t2 + t3
opf : 10f + 10su + 10sl + 10st + 10t1 + 10t2 + 10t3
Wouldn’t we be saving some energy (employees, machines, whatever) in the batch process ?

Last question from my side, how long is the customer waiting ? If your customer is the postman, will he come to you every time you are producing an envelope, even if you are working in a one piece flow fashion ?

Of course, WIP, defects, …, don’t stand the confrontation 🙂
Looking forward to reading your opinion
Matthieu

Let me apologize for stammering and add some lines I just forgot to write in my first post.

————————

These lines :
“So what about our ten envelopes ?
With a lot size of 10, we have the batch process time : 10f + t1 + 10su+ t2 + 10sl + t3 + 10st

And as for the one piece flow process (if we assume from Eric’s datas that st<sl f) ?
Whereas the batch time formula doesn’t change, the one piece flow becomes… ”

—————-

Should have been
“So what about our ten envelopes ?
With a lot size of 10, we have the overall batch process time : 10f + t1 + 10su+ t2 + 10sl + t3 + 10st
With one piece flow, we have (if we assume from Eric’s datas that st < sl < su f) ?
Whereas the batch time formula doesn’t change, the one piece flow becomes…”

—————–

again sorry. Ron, if you could edit these, I would be grateful.
Matthieu

Hi Ron, nice video despite the constraints you mention.

Although I passionately sell the idea of one-piece flow to my clients let me play devil’s advocate here. The logical end to this process would be to get the letters to the post office and suppose (for arguments sake) the post office was 10 minutes away. Therefore, delivering 10 letters on the basis of one-piece flow would take 200 minutes. By carrying 10 letters in one trip this can be done in 20 minutes … saving 180 minutes and a lot of gasoline. That’s significantly more than the 1 minute gained in actual production. I know some folks would say we have a balancing problem but there is no way you can cut down the travel time unless you move house next door to the post office and that we all know is not a feasible solution. Alternatively, you could convince the postal department to come to your door and pick up each letter the moment it is ready!! That would be something!!

Regards
Arup

OPF will have the advantage if it reduces the amount of product handling. In the presented scenario, where there was only one person doing the work, the handling was reduced, because the OPF approach eliminated the wasteful steps of adding a folded letter to a WIP pile and then picking it up again to put in an envelop, and then adding the stuffed envelop to a WIP pile only to then pick it up again to seal it, etc.

However, if each step of the process was performed by a different person, I don’t think this handling waste would actually be reduced; it would just be spread across 3 people. Piles of WIP would not accumulate, and lead time would be reduced, but the total effort to produce a finished part would be no less than for the 1-man batch process. Total effort might, in fact, be larger, because instead of moving a pile of WIP once, the “pile” (of 1 part) is moved for each and every part. If increased throughput is needed, might a better scenario be to have each of the 3 people doing all of the steps in parallel rather than each doing one step in series?

Also, I’m curious how you respond to the challenge that there are no economies of scale to be had with OPF? OPF processes are readily scalable via mass parallelization, but only linearly, meaning it takes X times as much effort, time and material (i.e., cost) to produce X times as much product. Done right, batch processes can do X times as much product (at the operation in question, at least) for less than X times the cost. One possible answer to my own question is that there is an often significant investment of time and effort required to develop (and validate) a scaled up batch process.

Finally, the shorter time until the first product is produced is only an advantage if you can release the product as soon as it is produced. If you need to wait until all product is made to release the batch, then this offers no advantage.

Thank you very much for posting this video. When I first watched it, I was incredulous. I said he’s not doing it right, simulating the sealing and stamping is easier/faster than doing the actual job. I say this with confidence as a person who has to do the exact fold/stuff/seal/stamp routine on 350 pieces a month for my job. So today, I decided to recreate your experiment and timed myself doing 10 pieces in batch operations (the way I normally do this task) and 10 pieces using one piece flow. The result: batching took 4:42, OPF took 3:48.

Hi ,I am intersted in comments in apparel and other factory on one piece flow and mass production .
Also please send me email about your comments on “Unit Production System”

I was just thinking.. So say I want a solar panel on my roof, is the one-piece processing better for that? I’d place the order with a fitting company, they would then order it from a supplier. The supplier would then build then construct the solar panel, build it, test it and deliver it. Id’ then have it fitted.

Using mass production, would the item be in stock and already tested? Wouldn’t the customer see the item in days rather than the entire time it would take to produce the whole thing from scatch?

In the example you showed, it relies on the fact that a single operator is present. What if for each envelope you folded, you had to wait for the tree to be cut down for the paper by a single person? Batch processing a tree for all the paper is surelt more efficient in this case?

I like this process and I think this is a very clear way of presenting the Lean methodology …

Excellent Video to explain the mall batch size. It would have been time saving if it was time lapsed. Thanks for the content.