Make-to-order or Make-to-stock
May 17, 2006 at 5:07 pm #43425
I hope this isn’t too far from this forum’s subject range, so here goes: I’m working in my company on setting up a process that will supply parts to final assembly stations at 3 different lines. There are roughly 300 different parts between all 3 lines. I’m trying to decide whether I should use make-to-order or make-to-stock. Another person in my team has taken the idea of using a 2-bin kanban system and run with it. He has determined this is best because the capacity of the machine in the process doesn’t have the capacity to sustain a make-to-order system. He says, “Because we’ll have the ability to make more parts than necessary on days when demand is below average, that should mean we’ll have enough extra inventory to supply the lines on days when the demand is above average.” I think this has the potential to fail for two reasons: 1) We’re not sure we’ll be making the correct parts when we’re making to stock for future demand, 2) On days when demand is unusually high for one particular part, the kanban amount won’t be sufficient. On the other hand, I don’t see how make-to-order will be any better.
I’m open to any suggestions for the line or suggestions on reading material.
Thanks, Andy0May 17, 2006 at 6:49 pm #137730
ElbrinParticipant@Elbrin Include @Elbrin in your post and this person will
be notified via email.
This is a classic Theory of Constriaints exercise. Without addtional information it is difficult to offer any meaningful advise, however, if the machine is the constraint to your process, this is where you need to concentrate your efforts. Exploit the constraint, Subordinate everything to the constriant, Find ways to Elevate the constraint. Set up Drum Buffer Rope. The Drum beat is the pace at which your machine can produce parts. The Buffer an amount of WIP infront of the machine to ensure you don’t starve it, losing production. The rope is the time necessary to complete the process up to the machine, and timing of resource / material release to be sure your Buffer does not grow out of control or disipate and starve your contraint. The only place you should have signifcant WIP is infront of your constraint, as your buffer. To elevate your constraint think about set-up reduction, mistake proofing, visual order, possibly outsourcing or even an additional machine. Measure throughput, throughput and throughput.0May 17, 2006 at 6:50 pm #137731
I don’t think that this is too far from the forum’s subject range.
I think that a Kanban system could work – but – it has to be the right system and the right prep work has to be done.
Kanban is rarely the first tool used or applied in a business when making the jump to Lean. Hopefully the proper steps of VSM, waste elimination, etc. have been looked at on the process first.
If this is the case, then I would recommend looking at a mixed model kanban system. It is lengthy to explain so I will just recommend some reading material which you can get from pretty much anywhere (easiest at Productivity Press) and it is not too expensive. Try the book “Implementing a Mixed Model Kanban System” by James C. Vatalaro and Robert E. Taylor (isbn #1-56327-286-5) – it is basically a how to guide for implementing this type of system.
Another type of calculation that you can look at for production rates is EPEI (Every Part Every Interval). It provides the minimum producible batch size, given the demonstrated capabilities of your manufacturing resource(s). This can help you from the scheduling side and let you know when you should be making which part (which seems to be one of your key questions).
Hopefully this helps. There is reference material available on line and through books that can help you with both of these concepts.
D0May 17, 2006 at 10:02 pm #137735
ramblinwreckParticipant@ramblinwreck Include @ramblinwreck in your post and this person will
be notified via email.
I agree that a Mixed Model 2 Bin System might very well be the answer, BUT I still do not think there is enough information to make sure we all understand the question.
Are the 300 parts relativly inexpensive or somewhat costly?
Is it set up time or process time that leads you to believe that make to order will not work?
What has been done to reduce both in your system?
You say it is over 300 parts, but do you have demand history on these parts?
If so and the parts are costly you might want to Kan Ban the low runners and make to order the high runners, assuming process and set up times allow for this.
In the end it is just a math equasion delaing with time, cost and the allocation of lablor, floor space and money.
As you business changes so will the optimal answer. So once you have set up a system make sure you have operational parameters in place to monitor the output and demands on the system.
More information will only help
0May 17, 2006 at 11:13 pm #137736
Adam L BowdenParticipant@Adam-L-Bowden Include @Adam-L-Bowden in your post and this person will
be notified via email.
a 2 bin Kanban may work BUT a 3 bin version along with demand / statistical analysis wll enable you to calculate appropriate levels per Kanban and allow you to determine the risk you want to live with with regard to stock outs – is it 5%, 1% or 0.1%.
A kanban will not be appropriate in all situations – you need to set up processes that will typically cover all demand types:
new – no history of demand, fast moving – kanban may be appropriate, slow moving – kanban or make to order, sunset – make to order, obsolete – re-use, modify or scrap etc
Setting up processes using classical lean methods will work ok but applying statistical techniques will enable optimization and risk tollerence to be built in – it’s not too hard either.
Hope this hasn’t confused you – seems pretty logical in my head :-)
Adam0May 19, 2006 at 12:15 pm #137811
You’re right, I didn’t give enough information. The process is new to the plant and although it has been producing parts, it hasn’t been assigned any demand as of yet. Currently we’re just making parts to see what kind of issues we might face, looking at quality, etc.; we’re also stocking the finished parts.
The process input is raw material, so no need to worry about WIP in front of it. The capacity of the process IS limited by the machine. We can’t improve the machine cycle time, but we have ways to improve the unloading cycle time. Would that be a form of set-up reduction? E.g. we’re improving the time between batches.
Thanks, Andy0May 19, 2006 at 12:23 pm #137812
“I think that a Kanban system could work – but – it has to be the right system and the right prep work has to be done.”
Thats exactly what I’m trying to figure out…which system is the best for this process? This process isn’t incorporated into the overall plant yet, I’m still looking at different ways to set it up. Although I’ve focused on how we’re storing the finished goods from the process, I think I really need to look at the bigger picture here. For instance, I could theoretically run this based on MRP, JIT, push, pull, make-to-order, make-to-stock, etc. The make-to-order versus make-to-stock jumped out first and foremost so thats why I initially went in that direction.
You also mentioned VSM, but would it be helpful to do a current state map when in all reality there is no current state? I do have estimations though.
Thanks, Andy0May 19, 2006 at 12:28 pm #137814
I’d be happy to share more info.
First and foremost, this is a new process to the factory, so some of my numbers(especially demand across the different parts) is shaky.
“Are the 300 parts relativly inexpensive or somewhat costly?” I’m looking into it, but thats a great question!
“Is it set up time or process time that leads you to believe that make to order will not work?” Process time, the machine isn’t capable of handling the estimated demand. But I have the option of running the machine on the second shift.
“What has been done to reduce both in your system?” Very little process improvement has been done because the process hasn’t been in use yet.
“You say it is over 300 parts, but do you have demand history on these parts?” I have an overall estimation of the demand for all parts, but not specific to each part. For the most part, the machine will require the same cycle time for any part.
Thanks, Andy0May 19, 2006 at 1:46 pm #137818
I was just thinking about this some more and the real question I should’ve asked in the first place was this:
Because I know my daily capacity won’t match my average daily demand, am I forced to shy away from make-to-order?
Thats in a nutshell, but you get the point.
Thanks, Andy0May 19, 2006 at 2:01 pm #137819
Ovidiu ContrasParticipant@Ovidiu-Contras Include @Ovidiu-Contras in your post and this person will
be notified via email.
The driver for MTO or MTS is given by the demand for the part:
if the demand is 1 per month, the part is MTO
if the demand is 300 per day, the part is MTS
0May 19, 2006 at 2:58 pm #137821
BritParticipant@Brit Include @Brit in your post and this person will
be notified via email.
Go back to the constraint message from Elbin – that should be your focus if you cannot meet demand from the capacity of the system. Seems like until that is improved, the only way to manage the inventory efficiently is make to order + holding stock based on the lead time of your process and your demand. This holding stock should be reduced as you increase the capacity of your process to met the average demand.
By the way – didn’t see this, but does the process feed another process or produce a finished part? If it feeds another process, then the holding issue isn’t as critical in the short term – you can test a bit. If it produces a finished part, then along with my suggestion, I would also work with your sales people to determine the real customer demand in terms of time and competition. They may be over-promising on you when it is not necessary. Just a check. I would rather promise and deliver than promise and be late.0May 19, 2006 at 4:10 pm #137828
In spite of you use make-to-order or make-to-stock system.you must insure process under control & customer request.If you unilateral use make-to-order or make-to-stock system is not fulfill produce request,you might united use make-to-order and make-to-stock.
0May 19, 2006 at 4:23 pm #137830
Yes, but what if the capacity of the process was 600 per day??? Would the same law still apply? I would think not, but I wonder where I should draw the line when looking at the ratio of average demand to capacity.
Andy0May 19, 2006 at 7:31 pm #137840
RomelMember@Romel Include @Romel in your post and this person will
be notified via email.
Before identifying if you intend to do make to stock or make to order, I would suggest that you validate first your resources —- machine capacity, changeover/setup rate, line layout, OEE, among others — because this will determine your run rate basing on the items that you intend to feed. Second, check your days of inventory, the rate the demand is fluctuating, cost per unit, others and see how your inventory system will fit. Third, check the capability of your suppliers to bring in the raw items timely and in good quality. Once everything’s done, decide which fits best if it’s make to stock or make to order. Kanban looks good as well as the one-piece flow as well.0
The forum ‘General’ is closed to new topics and replies.