Thursday, February 28, 2008

Value Stream Mapping - A Frank Discussion

"An emperor who cares too much about clothes hires two swindlers who promise him the finest suit of clothes from the most beautiful cloth.  This cloth, they tell him, is invisible to anyone who was either stupid or not fit for his position.  The Emperor is nervous about being able to see the cloth himself so he sends his ministers to view it.  They see nothing yet praise the cloth.  When the swindlers report a suit of clothes has been fashioned, the Emperor allows himself to be dressed in their creation for a procession through town.  During the course of the procession, a small child cries out, "But he has nothing on!"  The crowd realizes the child is telling the truth and begins laughing.  The Emperor, however, holds his head high and continues the procession."

http://en.wikipedia.org/wiki/The_Emperor's_New_Clothes

 


Dear Friends of Atzari,

 

We find some of our clients asking about Value Stream Mapping.  On the surface, the term  conjures a very healthy approach to the elimination of waste in a process.  In fact, some of you are asking for that as the standard tool through which to streamline a business or manufacturing process.  We feel it is time to take a deeper dive to understand the strengths - and limitations - of this tool.

 

VALUE

First, let us take the term itself.  In Lean, the word "value" is typically used in the context of separating "value-added" from "non-value-added" steps.  This is a very powerful concept and one which is easy to understand.  However, like many concepts, it is based on a philosophical assumption that appears to be without reproach - until we see that it has a fundamental fallacy.  Unfortunately, both the concept - and the fallacy - find their way into the tool.

 

The concept is very easy to understand.  Let's assume we are making a widget, and it will take ten (10) steps to make that widget.  The first step is to take a piece of bar stock and to cut it to the correct length.  Let's say the next steps are necessary until after step 10, we have a finished widget, ready for sale to the end customer.

 

If we are creating the widget, and we cut the bar stock to the correct length, we are adding value.  In other words, we have moved it one step closer to the final product.  On the other hand, if the widget travels 50 feet across the shop floor, we have not added any value.  We are no closer to the final product, since we have not changed it in any matter other than its location on the shop floor.

 

If we are providing a service, we first take the customer's order, then make a seat, room, or time available.  However, until the customer actually utilizes the seat, room, or time to enjoy the service, all the preparations were simply operating expense.

 

So where is the fallacy?  Did we not add value by cutting the widget to length or preparing for the service? Yes and no.  Yes, we did take it one step closer to its finished state.  No, the end customer will not pay for a piece of raw material that has been cut to length but is otherwise useless without the remaining 9 manufacturing steps being completed.  In fact, one could argue that that piece now becomes a nuisance because it has to be labeled, stored, and handled to ensure that it is not confused with other pieces of raw materials cut to a similar length for a different end-product.  In the service scenario, its just another empty room, seat, or time slot.  And here is the grain of the matter.  End customers pay for end products which are usable.  A part, whether it has traveled 1/10th of its supply chain or 9/10ths of its supply chain, is nothing more than work-in-process.    A patient sitting in an empty doctor's examining room waiting for the doctor is the same thing - no value until the doctor arrives.

 

Is this important? Yes.  And here is why.   Suppose that the widget will "cost" $10.00 to make.  The cut-to-length step "absorbed" $1.00 of labor and overhead.  For lack of other metrics, we tend to use this as a measure of "value."  Has the piece really increased its value to the end-customer by $1.00?  Let's now go through the next three steps.  We now say it has a "value" of $4.00.  If we are honest with ourselves, we know this not to be true.   How do we know this?  Let's say we have a defect in step 5.  One piece "worth" $5.00 is scrapped, while another proceeds to step 9, and then is scrapped.  Examining our trash can, we would tend to say, we have a $5.00 piece and a $9.00 piece.  In fact, all we have a bar stock missing some material.  When we discarded the $5.00 piece, did we do any of the following:

 

(1) Send any line operator home without pay?

(2) Pay less electricity, gas, water, or wireless internet?

(3) Pay any less rent for the building?

 

We know that all of these expenses remained the same whether or not we produced anything.  These expenses are something else altogether: operating expenses.  Most of these are not variable, even though we are told otherwise.  We had to pass these expenses on to the end customer in order to cover our losses.  The customer paid for this only because of the value he or she ascribed to the widget, and because he or she did not have - or pursue - other options.

 

In fact, the only time the widget is of any value is when we sell it to the end customer for say $50.00, making a tidy profit of $40.00.  If we look at the operating expense and the raw material, the only true variable price is the raw material.   Notice I did not say cost - I said price.  Price is what one charges for goods and services.   There is no such thing as an inherent cost.  There is only the price we chose to pay for the bar stock - until we found someone willing to sell it to us for a lower price. 

 

So, in fact, our widget is somewhat worthless until we have a complete widget that we sell for $50.00.  Everything else that happened prior to that was operating expense plus the price we paid for the raw material.

 

However, we instinctively do not want to throw out work-in-process because, in doing so, we throw out the opportunity to make the finished widget.  One could argue that both the $5.00 piece and the $9.00 piece each represents a lost $50.00 opportunity to sell a finished widget.  Then the true loss is $100.00, not $14.00.  Once we have scrapped that piece, we also scrapped the opportunity at each downstream operation to make something that we could sell - and therefore incurred the "cost" of having those operations idle.


We could just as easily talk in terms of uptime and downtime rather than dollars.  However, when any delay occurs, we incur the downtime of all downstream operations.

 

And that is where we arrive at the true value.  If we are wasting time moving bar stock around the production floor, or letting it sit on a shelf, we are keeping downstream operations idle from the opportunity to produce finished widgets which we can sell for $50.00 each.    However, the only operation that matters is the rate-limiting one.  The one that determines how many widgets we can make.  If we rapidly make excess cut-to-length bar stock, yet we have a slow de-burring operation at the end, then it is the de-burring operation that sets the rate of selling widgets.  This is our correct throughput.   That is what occurs when we produce something that our end-customer is willing to consume. 

 

Value stream mapping by itself would not normally make that distinction.  It would simply divide operations into value-added and non-value added.  The accumulation of excess inventory could be considered to have "value" when in fact it does not.

 

STREAM

Stream is another term that conveys the thought of streamlining - or making something more efficient.  In fact, the efficiency only matters at the rate-limiting operation.   Beyond that, efficiency is simply the accumulation of work-in-process inventory.  Below that, the rate-limiting operation is idle and accumulating lost opportunities to provide finished widgets to our end customer.  The word stream incorrectly implies that we can have a balanced line.  In fact, there is no such thing as a balanced line unless we achieve the statistical perfection of no delays, no scrap, and perfectly timed events in sequence.    How often do we see the focus on the goal of a perfectly timed assembly line - only to see the product sit on a shelf waiting to be sold or to be moved to the true rate-limiting step?  I actually saw a line kept at a speed of 600 pieces per hour - only to have all the parts sent to a room where it took an hour to inspect 100 pieces!  The reason?  They could not afford the loss of "efficiency" in slowing down the line to be able to inspect parts as they went on the conveyor belt!

 

The real stream is to ensure that the rate-limiting operation is never allowed to be idle.  And, yes, we may need some excess inventory in front of it to prevent that possibility.  Taking value stream mapping to its extreme, we could try to balance the line and starve the rate-limiting operation.  

 

We must be very careful when balancing a line to make sure that we have enough buffer to prevent the bottleneck from becoming starved, yet not so much that we simply accumulate work that cannot be processed.  That is the true balance of a line.

 

MAPPING

Mapping conveys the idea of having true knowledge of a process.  We link steps in sequence  by identifying inputs and outputs.    This is valuable in order to make sure that a piece at one step is ready to proceed to the next step.  However, this must have been done from the very beginning if we have made a successful widget.  The real purpose is to isolate those inputs and outputs that have meaning to the next steps from those that become orphans - wasted motions and delays that do not contribute to this effort.  

 

Mapping is an excellent way to  eliminate waste - as long as we know which waste matters.  Keeping an operator idle to  be available, watch over, and ensure that the rate-limiting step is never idle is not a wasted resource.   On the other hand,  keeping an operator busy on making excess WIP is a wasted resource to the extent that it does not help us to make additional finished widgets for sale. 

 

A more valuable map will show the constraint, the buffer, and the line's correct takt time.  Other inputs such as temperature, time, pressure, and similar parameters have their place when designing the process or when trying to reduce scrap due to low yields.  When these are placed in a value-stream-map, we could easily confuse the two.   It is possible to populate the value stream map with inputs that have nothing to do with eliminating wasted motions, thereby masking the true opportunities.  Excess information does not equal more knowledge.   In fact, it may mask our ability to find useful knowledge.

 

Value Stream Mapping, when used correctly, can yield many benefits.   For this to happen, we must first brush aside the mental cobwebs of local efficiencies, local costs, and that of assigning an arbitrary value to WIP pieces.


--
Thanks and regards,


Jose I. Mora, Principal Consultant
Atzari Consulting, L.L.C.
www.atzari.com
Office: (973) 835-6313 Fax: (866) 223-5813
Mobile: (917) 566-0965
jmora@atzari.com, joseimora@gmail.com


No comments: