VSM Part #7 - Aftershock

I've had the good fortune to hear from some of my readers and noticed a few disturbing trends.  It appears that there may be one more post I need to write. Those of you who use Value Stream Maps as a matter of course, can ignore what follows, but those who hesitate to create one; or, who find yourselves all wound around the axel as you try to create or use it, you may want to read on.

I wrote this series because I find two recurring themes with VSMs:

Don't make the process too difficult

1. Practitioners make the process too difficult;
2. The wrong people try to use them.  

I hope that this series helps readers to change that.

One other trend I've noticed is that practitioners try to get too granular in the collection of data for their initial "Current State."  The result is that they make the process too complicated and get defeated before completing.  As an example: I've had a reader tell me that they try to average 30+ cycle times before they feel they have a good sample size for a process step.

Heck, until you've instituted Standard Work, that's way too much work for something that is, at best, a moving target.  Start with course measures.  Take two, maybe three samples to start, and get more detailed with each successive Kaizen.

At first, focus on the whole forest, not individual trees

Besides, initially, you're looking at the big picture: the health of the entire forest, not individual trees.  You want to know: How are we communicating inside & outside the organization?  How are we ordering material and how are we shipping finished goods?  How much inventory (WIP, Raw Materials & Finished Goods) do we have tied up?  What's our ratio of VA to NVA?  How many people are we employing to perform the work?  Are we pulling or pushing?  What's the state of maintenance, changeovers, quality, etc.?

Remember: the "Current State" VSM is just a baseline.  It will be the standard against which you'll judge future improvements.  Also recall the words of Taiichi Ohno, the putative Father of Lean:

That first VSM, the initial  "Current State," is just a quick "State of the Business" analysis.  Like a Hoshin Kanri, it's used to guide your transformation from there.


I hope this series has achieved two goals:

• You'll give the creation of a "Current State" VSM a shot.  Take your time.  Get it close, not precise.
• You'll ask yourself, what are we learning about our business and how will that help us transform into a better version of ourselves?

Feel free to contact me with your questions, comments or observations: robert@gettingtolean.com

Meanwhile, Get Lean, Stay Lean!

VSM Part #6B - Importance to Decision-Making ("If this, then that")

The VSM tells leaders* what is obstructing flow within the value stream.  Trust me, there’s always something, but suppose this is the first time they’re reviewing the VSM.  What are they looking for?

1. CT > TT.  The number one obstruction to flow is any Cycle Time that is greater than Takt Time.  You'll want to correct that post haste.
2. Large amounts of Inventory.  Inventory is one of the original seven wastes.  The more you have, the more of your money is tied up and the slower your cash flows.  Reduce inventory as much as possible and still be able to meet your customer's demand.  Notice: I didn't say reduce to zero.
3. High Defects.  Defects are also a waste.  Defects automatically engender additional cost.  Defects must be winnowed down to zero.  Yes, zero.
4. Low Uptime.  Low uptime speaks to undependable equipment and that will adversely affect your ability to consistently meet customer demand.  Equipment should be available 100% of the time it s needed.
5. Low Available Time (AT).  Recall that Available Time is the numerator in the calculation for TT.  The less AT you have, the faster you need to make products to meet customer demand.  Excessive speed can lead to defects; so, you want to increase AT whenever possible.
6. High Changeover time (CO).  As we all know, CO time has a direct and inverse impact on Inventory: i.e. the longer the CO time, the higher the Inventory.  You'll want to be ruthless in driving CO time as low as possible.
7. Large supplier deliveries, especially if made infrequently.   This is a signal that you aren’t using Kanbans or some other JIT methodology.  The result is excess inventory.  See #2 above.  Drive supplier delivery size and frequency to the best mix of small size and high frequency.
8. Large, infrequent deliveries to the customer.  This condition leads the customer to have excess money tied up in inventory (they have to warehouse what you ship until consumed), but it also means you have to warehouse it before you ship.  The consequence is that, unless you’ve made provisions to be paid for work in progress, you have your money tied up while you wait to ship.  Drive delivery size and frequency to the best mix of small size and high frequency (taking into account transportation costs).
9. Unrealistic forecasts are useless.  An annual, or rolling six month forecast has value as it conveys trends, but you want data on which you can hang your hat.  Work with suppliers and customers to develop reports that give "actionable data."
10. MRP & Scheduling need to be addressed.  MRP systems should be used, almost exclusively, to develop material forecasts and costs.  If used to schedule your operation, choose only a few "pacemaker" operations (one is preferable) to schedule.

Once you know where flow is obstructed and what those obstructions are, you can plan your Lean events.  Typically, these are Kaizen events.  What events will you plan?  Here's a list you'll want to review.

• For CT > TT you’ll generally choose a Standard Work event.  
• Since inventory is generally directly proportional to changeover time, you’ll want to conduct a SMED event to reduce CO time.  You may also want to implement Kanbans or supermarkets while you're at it.  
• High defects are frequently the result of variation (executing the process differently from time to time), so you’ll want to conduct either a Standard Work or Total Quality Control event.  Since poorly maintained equipment can also lead to defects, you'll need to use your judgement whether you'll need to conduct a Total Productive Maintenance event.
• Low uptime is generally indicative of poor maintenance practices and should lead to a Total Productive Maintenance event.
• Low available time can result from a number of concerns, but its remedy should begin with Standard work. 
  
  Once you've completed a Standard Work event, the sum of CTs divided by TT for the operation gives a great indicator of the number of workers needed.  REMEMBER: you can't have fractions of workers, but you can flex workers or hire part-time help.  You'll also want to reduce that fraction to zero in that operation's next Kaizen event.
• As discussed above, high Changeover Time is addressed by a SMED event.
• Large supplier deliveries point to a number of concerns:
• Inability of supplier to deliver on time
• Inability of supplier to deliver reliable quality
• Wild fluctuations in customer orders (Acme’s +/- 50% variation is a good example)
• Excessive transportation costs

The first two and last item need to be addressed by your procurement group.  The third needs to be addressed by the sales organization.

• Large customer deliveries are often the result of customers not being able to trust your delivery or quality (or their own forecasts).  You need to address these first, then work a new schedule with your customer.
• Unrealistic forecasts (on your part) probably need to be addressed using a Kaikaku.  This isn’t a week-long event.  More likely, it will take weeks or months.  It will have representatives from all the internal stakeholders and, if you’re smart, will include representatives from both your customer and supplier(s).  This event will take time to identify the root cause(s) of your inability to forecast more realistically and then develop reliable solutions.
• MRP & Scheduling should be de-coupled, as much as possible, from your day-to-day operations.  Use your MRP system to schedule and track incoming raw materials and customer orders.  Identify the "pacemaker" operation(s) that can "pull" product to them and, if you use your MRP system in production, schedule only those operations.  The "pacemaker" is where you'll indicate what version of your product you want, and in what frequency.  Once the "pacemaker" is scheduled, it cascades a "pull" signal back upstream causing the production of the indicated product.  This results in both "flow" and "pull."

I welcome your insights and feedback.  You can contact me at: robert@gettingtolean.com

* In my last post, I explained why only the most senior leader or the Value Stream Manager should use the VSM to make improvements.  

VSM Part #6A - Importance to Decision-Making (Who, What & How Often?)

To this point, we’ve examined the mechanics of the Value Stream Map (VSM).  We’ve looked at the the flow of the process, at the data boxes, at the flow of communication and the overall assessment of value adding vs. non-value adding.

All of this has been to reach the point that leaders can make informed decisions about how to use their limited Lean assets to improve flow.

There are two key points I’d like to make before addressing those decisions.

• First, the VSM is a leadership tool.  While everyone is encouraged to read and understand it, VSMs are not intended to be used by others in the organization.  Why?  Because only senior leaders (or the Value Stream Manager) control the Lean Assets (discussed below) needed to improve the value stream.
• Second, Value Stream Analysis (VSA) needs to be performed from the customer backwards.  Remember: the VSM is a tool used to identify obstructions to flow and then remove them.  Not all obstructions will impact the customer equally when removed.  So, the order in which we remove them is from the customer back to the supplier.

Removing Obstacles to Flow 

This last bullet (above) may not seem intuitive, so let me use an example:  Picture a value stream in which Takt Time (TT) = 22 seconds.  Cycle Time (CT) at operation #2 is 45 seconds; more than twice TT.  Meanwhile, CT on operation #5 also exceeds TT, but only by 15 seconds.  Some might want to attack the larger problem (operation #2) first, but let’s examine what would happen.

Hypothetically, we’d conduct a Kaizen event to get CT at operation #2 down to TT or less. But ask yourself, how much faster will product flow to the customer?  It will flow at the pace of the slowest operation closest to the customer, in this case, operation #5.  Why?  Because we didn’t remove that obstacle.  So, all the product we can now push through operation #2 will just jam up between operations 2 & 5.

If, however, we get the CT at operation #5 to TT first, it will start draining the lake of inventory in front of it.

Then, when we remove the obstruction at operation #2, product will flow at TT all the way to the customer.  So, with rare exception, correct problems closest to the customer first, then work your way back to the supplier, one operation at a time.

REMEMBER: The goal isn’t just to get Value Stream flow down to TT or less, but to do it in such a way that value is pulled from each successive operation.

As I said above, VSMs are a leadership tool.  In the event that your organization hasn’t been structured into individual value streams, with their own Value Stream Managers, the senior leaders at the site should be the ones reviewing the VSM.  This review is most often referred to as a Value Stream Analysis, or VSA.

Reviewers, which include the senior leader and their staff (or the VS Mgr and their staff), should analyze the VSM no less frequently than monthly.  Obviously, if there is a VS Mgr, he or she should be reviewing the VSM as a tool of constant feedback. 

What’s magical about a monthly review?  In a Lean organization, a month is a long period of time and during that period at least one Kaizen event should occur.  Every time you conduct a Kaizen event, you change one or more dynamics in the VSM; hence, the cause for review.

What are Lean Assets?  When I refer to “Lean Assets,” I’m referring to some or all of the following: 

• Your Continuous Improvement (CI) Manager* and any staff under his or her direction;
• Any process or equipment that will need to be taken offline so as to conduct an event;
• Any product you’ll need to build ahead so you can shut the process or machine down for the event, including any overtime you’ll have to approve to build that product;
• The team that will be assembled to conduct the event and perform any follow-on activities.

So, in this post we’ve examined Who uses the VSM, What they are looking for and How Often they will conduct their reviews.  In the next post, we’ll examine how leaders will use the VSM to allocate their Lean assets. 

*NOTE:  The title for the senior Lean person is immaterial.  What is important is that you have one and that they report directly to the senior person at the site.  Even when you have VS Managers, an overarching Lean leader (CI Manager) is critical and becomes a key resource to the VS Managers.

Value Stream Map - Part 5: Value Add vs. Non-Value Add

One of the critical features of a Value Stream Map (VSM) is the sawtooth chart at the bottom of the map.  Many overlook this chart, as it seems superfluous when you’re first constructing the VSM, but it’s has two key uses:

1. The sawtooth gives you an immediate sense of how much of your customer’s time is spent waiting, rather than performing things for which they’re paying you.
2. It becomes a standard against which to judge future progress, what we call a "baseline."  

Let me take a quick sidebar.  I hope it's clear that YOU pay for all Waste (Non-Value Adding, or NVA Activities).  You might say, "It's just wait time and I'm not really adding labor or anything to my work, so what's the cost?"  The answer is that you are prematurely paying for all material and labor already invested in this product.  Since you aren't prepared to work on the product, you have your $$$ sitting in queue.  You are also delaying your own reimbursement.  

The longer you take to make a product, the longer you have your capital tied up and the longer your cash flow.  And then there's the cost of capital, the difference between what you are making while your product/service is being worked on (nothing), and what you could make if you invested that money.  Finally, there's the opportunity cost: the cost of losing opportunities you could have undertaken had your process not been clogged with things you aren't even working on.

Where do the data come from?

Let’s examine the chart above in greater detail. An example of this detailed exam is on the left.

The value on top of each tooth is the amount of inventory ahead of that operation.  As discussed in a previous post, Inventory is measured in units of time.  This value is a rough gauge of how long the customer waits before you perform your next operation for them.

The value on the bottom of each tooth is the cycle time (CT) to perform that operation for one item.  See the example to the left.  It says that your customer waits nine days for you to perform 25 seconds of value adding on the first part in this lot.

NOTE:  Some would have you post the value of the cumulative CT to process the whole lot, and not just one piece.  That will create problems later, when we perform a CT vs. TT comparison.  That comparison is critical, since it informs us whether we can meet customer demand  It's also used in calculating the number of employees needed to meet customer demand.  As a result, I put individual CT in this space.

Finally, at the far right side of the sawtooth (pictured on your left) you can see a rectangle with two boxes within it.  These boxes contain the sum of all the Inventory and Cycle Times respectively.  This ratio gives us a rough idea of our Value Add vs. Non-Value Add for the value stream.  That ratio is important for the reasons stated above.

In the example at the top of the page, the sum of Inventory time is 13 days.  The sum of Cycle Time is 63 seconds.  From here on, this becomes a math problem.  First, we need to convert everything to common units of measure.  13 days x 24 hours/day x 3,600 seconds/hour = 1,123,200 seconds.  Since waiting is a form of waste, we can accurately say that this time is NVA.

63 seconds / 1,123,200 seconds = .00005608974….  We then multiply that value times 100, to calculate the percent of total wait time that is spent working on the client’s product.  The resultant value is .00560897%.  In short, this organization works on their client’s product less than one percent of the time they make that client wait.  The message?  We have lots of room for improvement!  

This value now becomes part of our data baseline.  Every time we improve the flow of this value stream we should also be diminishing the amount of wait time.  As a consequence, the percent of Value Adding time should constantly increase.  Because we baselined, we can tell if we're improving.

As Taiichi Ohno stated: "Where there is not standard, there can be no Kaizen."  If you don't know where you started, you can't tell if you're changing for the better.

Next up: the value of the VSM in decision-making.

Value Stream Map - Part 4: Material Flow

Remember, a Value Stream Map should be communicating to you and  to your organization.  So far we’ve examined how the flow chart and data boxes communicate.  Now we’re going to examine how the supply and delivery icons communicate.

On the top left are the supply icons.  They tell us who the major suppliers are.  In the case of Acme Fire Suppression, Inc., the two major suppliers supply the steel blanks for use in the Upset operation and the chemicals used in the Auto Fill operation.

We know from my earlier description of the process that there are other supplies used in the process, e.g. the paint, the inert gas, the spray nozzles, cardboard overpack and boxes.  None of those made it onto this VSM.  The reason is mostly because they’re commodities, and easily available.  We just want to look at the high $ items, or the items that could potentially shut us down if they weren’t delivered in a timely fashion.

So, looking at the steel blanks, what do we see?  Well, at the top we see that we’re getting two different types of blanks: the 5 pound and the 10 pound.  We can see they come from the same supplier (normally the name of the supplier would be in the factory symbol).  Finally, we can see that we get shipments of 14,000, 5# blanks and 8,500, 10# blanks.

Next we see that the steel blank supplier delivers twice a week.  And, we see that we keep a 20 day inventory of their blanks at all times.  One should logically ask why we’d inventory 20 days worth of these blanks when they have the potential to deliver twice weekly.  Good question.

Next let’s look at the chemical supplier.  They are shipping 35 super sacks of the chemicals a month (each super sack weighs 1,000 pounds).  The same chemical is used in both the 5# & 10# extinguishers.  Meanwhile, Acme keeps a 25 day inventory of the chemical.  

Now let’s look at the delivery side of the VSM.  From this diagram we see that Acme ships out of their warehouse once a month.  Each shipment contains 40,000 5# extinguishers and 24,000 10 # extinguishers.

Again, this is all just data at this point, but it’s also something we call Actionable Information.  By that we mean that we have enough information to actually make decisions and formulate action plans.

Case in point: why would Acme only ship once a month and keep a 30-day inventory of fire extinguishers?  Doesn’t that violate at least one of the 7 wastes?  Armed with that information, what actions might that lead Acme to make?

Value Stream Map - Part 3: Information Flow

Perhaps one of the least well understood or appreciated parts of the Value Stream Map is the information flow at the top.  That said, it’s an important part of the entire flow of a value stream.  There is much to be learned from it.  Let’s examine.

COMPANY (Production Control) BLOCK:  Let’s start at the Company block, top center.  Here we note the Name of the company being studied.  If there are multiple value streams in this company, you may also list the value stream's name.  In our case, the company's name is Acme Fire Suppression, Inc.  Also here we’ll indicate who is at the Hub of the Information flow: Acme’s Production Control office.  It's rare that production control does not control the flow of production-related information.

Finally we note whether our processes are all manually scheduled or computer scheduled, and what type of scheduling tool is bing used.  In our case, processes are being controlled by a Material Requirements Planning (MRP) software program.  That bit of information alone tells us volumes.

COMMUNICATION FROM CUSTOMER:  Recalling that, by convention, the Customer is always on the top right of the VSM, we see that Acme’s customer is providing them with three separate forecasts.  First, there is six month forecast of their needs.  

This forecast is “squishy,” in that it only gives us a guess of what our customer really intends to buy.  Sales can use this forecast and we can develop an estimated Takt Time from it, but we really shouldn’t spend money or commit resources based on this forecast.  The reason for that recommendation will become evident when you see the monthly order.

The semi-annual order is a little more firm, but still nothing you can sink make long-term plans around.  Even the monthly order represents a target value.  Implied in this order is the hope that this order quantity will stay the same from month to month; however, the contract with the customer has provisions in it that allows them to alter this standing order, up or down, by as much as 50% in any month.  

If you think about it, that’s a lot of potential fluctuation.  Only when the final, amended, order arrives can we really schedule the month’s production.  The amended order alone gives us "actionable" information.

COMMUNICATION TO SUPPLIERS:  Now let’s look at what Acme communicates to its Suppliers.  Again, by convention, supplier information is found on the top left side of the VSM.

Like their own customer, Acme gives their suppliers as much of a view of the future as they can; so, they send suppliers a six month forecast.  As I said earlier, this is just a guess for planning purposes, but nothing the supplier can hang their hat on.

Next, Acme sends suppliers a two week forecast.  This is getting firmer, meaning that Acme knows they’ll be close to actually buying these quantities.  Then comes the weekly shop order that gives suppliers specific quantities to produce and ship.  Finally comes the daily priority list that gives Acme the ability to tell suppliers the order in which they want materials made that day.  

One might well ask why the daily list when there are only two shipments a week, but this is Acme’s way of projecting consumption by their own factory.

Seem confusing?  At least it’s all documented!  From a well-documented process you can make corrections that will improve it.

This VSM is likely the only place in a company where all this information exists in a single place.  Think about the implications of that.  If no one has been looking at all this, on what have decisions been based?

In the absence of such cross-functional communication many organizations fall into silo behavior, where everyone does their own thing, ignoring its implications on the rest of the organization.  By contrast, imagine if all departments worked from a single sheet of music that was updated on a timely basis.

INTERNAL COMMUNICATION:  We have one more flow to follow, and that’s the flow within Acme.  This one is easy.  The VSM shows us that once a week, the MRP systems sends ("pushes") shop orders to each operation in the value stream.  Each shop order is an order to “build,” instructing that operation in what they are to make, in what quantity, in what order and when it’s needed.  

One thing common to all communication in this Current State map: it all is pushed.  There is no feedback loop or any sense of give and take to the flow of information.  Especially in the factory.

Keep in mind:  When every operation is being scheduled, even though some operate at very different cycle times, MRP-generated schedules almost always lead to products being “Pushed” and a lot of WIP (read: excess inventory).

I’ll end the discussion of the information flow here, but can you start to see how densely the VSM is packed with information?  As I said earlier, this is information rarely found in a single place anywhere else in an organization.  Once you’ve made the effort to document it, you’ll want to keep the VSM accurate by updating any changes.

Next, we’ll examine the flow of raw material and finished goods.

Value Stream Map - Part 2: The Data Box

In the picture at the top of the page, you can see both flow chart (discussed in my last post) and the data boxes below them.  I’ve included the flow chart in that picture, because the data boxes are linked to specific operations and paint a compelling story of that operation.

So, what is this story the data box tells?  

First what goes in a data box needs to be tailored to the value stream being mapped.  I’ve used five standard metrics, but they might not work for your business; or, your business may need to add another measure specific to your industry or process.  

CT - Cycle time is a critical measure of any operation.  It tells us how long it takes to complete one iteration of that process step.  It’s important, because, when compared to Takt Time (TT) we learn immediately whether we can meet customer demand.  Remember, CT must be less than or equal to TT if we’re to meet our delivery commitments on time.

C/O - Change Over time is the time between making the last good piece of a previous product and the first good piece of the new one.  While sometimes referred to as SMED (single minute exchange of die), many industries don’t employ dies.  Also, it’s not just about changing a die, but about taking whatever measures are necessary to make the first good part of the new product.

AT - Available Time is the time operators have to actually work each day.  It’s the time between when they arrive and when they leave, less any time spent on breaks, lunch, meetings, clean up, etc.  As you know, available time is a key component of calculating Takt Time.

Defects - This is pretty self-explanatory.  We’re trying to get an idea of how reliable the process at this operational step is.

Uptime - This is the measure of how reliable machines used in that process are.  Calculated by subtracting machine down time from 100%, Uptime gives us an idea of where we have problems with equipment.  There is a much more comprehensive measure called OEE, and I'll deal with it in a later post.

CONSIDERATION:  When you have multiple machines or multiple operators covered by a single data box, you need to make the decision how you’re going to calculate data values.  Some average the values.  I believe that creates misconceptions about the process.  I elect to post the worst data.  You’ll understand why when I later discuss how to use your VSM to determine where to deploy your lean assets.

Value Stream Map - Part 1: The High Level Flow Chart

There are some wonderful books on this subject.  I have no desire to compete with them, but I would like to break open this extremely important and underutilized tool.  In essence, let's look under the hood and see what this thing's all about.

Here are my goals:

1. Explain each segment of the Value Stream Map (VSM)  

   A. High Level Flow Chart

   B. Data Boxes

   C. Communication Flow

   D. Material Flow

   E. Value Add vs. Non-Value Add

   F. Importance to decision making

2. Explain the way to use this tool as an ongoing guide in allocating your Lean resources

Let’s start with the High Level Flow chart.

What you see above is a high level flow chart.  It’s high level because it only depicts the major operations that take place in the making of a household fire extinguisher.  

The flow starts on the left with “Upset” and flows to the right.  We know this by following the directional arrows between steps.  It's also the convention of Value Stream Map (VSM) construction

These directional arrows tell us one more thing.  Because they are straight and dashed, we know by VSM convention that material is being “Pushed” to, not “Pulled” by, the next operation.

At the “Upset” station, a steel billet, the size of a shuffleboard disk, is mashed into a shape resembling a bowl.  

Note: the “Upset” box contains a “3” at the bottom.  That “3” represents the number of workers required to perform the operation.  This convention is repeated throughout the flowchart.

Hundreds of the billets arrive in huge steel bins and are stored in front of the upset machine.  This  theme of large amounts of inventory in front of operations is repeated between almost all of the operations.  That trend will be discussed more below.

Once upset, the bowls are placed in other bins until full.  They’re then pushed to the “Form” operation where they are pressed into deep cylinders.  

Bins of the resultant cylinders are pushed to the “Weld” operation.

At the “Weld” operation, two halves of the cylinder are welded together, creating a single canister.  

Bins of canisters are pushed to “Pressure Test” where the canister is submerged in water, filled with air and tested to see if it holds pressure.  

Canisters that pass are inverted in racks to drip dry, then are pushed to “Drying.”

In “Drying,” the racks of canisters are placed in large ovens to bake off any residual moisture, before painting.

After being pushed to paint, canisters are hung on a conveyor and electrostatic paint is applied.  The conveyor then carries the canisters through a drying oven.

Dried canisters are removed, placed in racks, and pushed to “Auto-Fill.”

At auto-fill, canisters are removed and manually put into a machine that fills them.  The machine weighs the canister as it fills.  As soon as the proper weight is reached, the fill process stops.

Filled canisters are removed and placed on a conveyor.

The ”Conveyor Operation” consists of three contiguous operations: spray nozzle insertion, inner gas pressurization and test. 

Canisters that pass test are placed in a packing sleeve for shelf display.  Packing sleeves are inserted into shipping boxes.  Full boxes are sent to the “Warehouse.”

Pretty simple.  Nine steps.  

One last thing.  The symbols between each of the process steps give us additional important information.  Those triangles with an ‘I’ in them stand for “inventory.”  Below the triangle are the units of time it will take to consume that inventory by the downstream (receiving) operation.  Those units should be consistent throughout the VSM.

In my next post I’ll address the “Data Boxes.”