Let’s Create a Future State Value Stream Map!

Update (4/4/14): We recently released an excellent podcast with Karen Martin on the topic of Value Stream Mapping. You can listen to it, while reading this article, by clicking here.

Let’s continue with our value stream mapping (VSM) series this evening by discussing the future state VSM. If you are just joining us, please feel free to read the previous articles before jumping into this one.

I Have a Dream!

The process of creating a future state VSM can be quite exciting as creativity and excitement merge together. It’s time to dream a little as we work to create better, safer, more productive workplace. And, please be sure to have some fun during the process!

In the last article we documented the “current state” process of our fictitious peanut butter and jelly sandwich factory. We learned the following.

Obviously, we have room for improvement.  And while this is a completely fabricated example, I think you will find similar opportunities in your operation.

Let’s see what improvements we foresee.

ct-tt.JPGStep 1: Create a Cycle Time / Takt Time Graph. From the data collected and calculated during the creation of the current state VSM we are able to draw a cycle time / takt time graph. This graph simply compares the individual cycle times to the overall takt time of our process.

This is an important step as it will help us make decisions as to how and what to improve in future steps. Click the above picture to see what this looks like for our example.

As you can see, both the peanut butter and jelly application steps are under takt time. The packaging step, however, is over takt time and is cause for alarm. In its current state, we are not capable of meeting customer demand without the use of overtime.

fs1.JPGStep 2: Decide if you will build to stock or make to order. Next, we must decide what type of distribution model we will develop. Will we build to a finished goods supermarket or ship directly to the customer?

In our example, we only produce one product and customer demand is relatively stable. Therefore, it would make the most sense to develop a make to order model. This means we would only produce what the customer wanted, when they wanted it.

Now, since most companies produce more than one product and battle inaccurate sales forecasts, building to a finished goods supermarket often makes the most sense.

One normally asks, “Isn’t this just building up inventory… one of the 7 deadly wastes?” The answer is yes. But, the key difference is we are controlling the inventory levels instead of the inventory levels controlling us.

So, for sake of example, we will assume the folks at KB&R, Inc. decided to implement a finished good supermarket where they will hold 0.5 days’ supply of finished stock (a.k.a. PB&J sandwiches) before shipping as demonstrated in the picture.

fs2.JPGStep 3: Calculate Optimal Crew Size and Implement One Piece Flow. Currently, the operators at KB&R work in isolation. You could say they all work to their own drum beat. Whenever possible, our goal  as lean practitioners, should be to get all operations producing to the same beat – namely takt time.

As we learned in step 1, both the peanut butter and jelly application cycle times are under takt time while the packaging step is over. This needs to be addressed.

The solution is to rebalance the work and implement one piece flow. The team first determined the “optimal crew size” for the entire sandwich making process.

To do this, the team took the total cycle time to produce a sandwich (97 seconds) and divided it by the takt time (38.6 seconds). This resulted in an optimal crew size of 2.5 operators. Since we cannot have 0.5 persons they rounded up to 3.

If the team is ever able to reduce the total cycle time to less than 77 seconds the crew size could be trimmed to 2 persons, freeing up the additional resource for other value adding tasks.

Next, the team worked with their lean sensei and designed a u-shaped cell where all 3 operators would work in a one piece flow manner. Once the work was redistributed the team was able to produce a PB&J sandwich approximately every 32 seconds. The total time for the first sandwich to be complete is still 97 seconds.

fs3.JPG Step 4: Pull when One Piece Flow is not Possible. Anytime we build to supermarkets we must have a way of signaling when to produce and when not to produce. There are a variety of ways to accomplish this. The easiest way is to use kanban.

Without going into great details on how to size and implement kanban, let me explain it this way.

When a customer places an order for 10 PB&J sandwiches, which we will assume is the standard pack-out quantity, a “withdrawal” kanban for 10 sandwiches is sent to the finished goods supermarket. The shipping clerk pulls off a “pallet” of 10 PB&J sandwiches and ships them to the customer.

This now leaves a “hole” in the supermarket. Therefore, a production kanban (white on a VSM) for 10 sandwiches is sent back to the u-shaped cell. This production kanban basically says, “Hey, we just shipped 10 sandwiches and need to plug the hole! Make us 10 more please!”

Additionally, the supply chain team has worked with their supplier and negotiated frequent milk runs. This allows the team to only hold 0.5 days of “bread” in a WIP Supermarket before the u-shaped cell. Again, kanban and a kanban post (goal post symbol on VSM) are used to replenish this WIP supermarket.

Finally, the implementation of kanban has allowed the team to turn off the MRP portion of their ERP system… and leagues of lean angels sang “Hallelujah!”

Future State Value Stream Map Step 5: Improve Communication and Schedule a Pacemaker. So far we have gone a long way to improve the flow of material and manpower. Now, let’s zoom in on the information side of the equation.

Rather than scheduling each process “on an island” the lean approach is to schedule one process. We call this process the pacemaker process since it sets the pace for the entire system.

By definition, the pacemaker is typically defined as the process step closest to the customer whereby everything after it flows. In our example, this would be shipping since the supermarket after the u-shaped cell technically impedes flow.

So, we will aim to schedule production at shipping. Once shipping gets the instructions as to what to do they will alert the upstream processes via kanban.

Additionally, the team has decided to ratchet up the communication with its customers and suppliers. This two-way communication will help all parties prepare for things like demand fluctuations.

Summary

On paper we have made some sizeable improvements. Production Lead-time (PLT) has gone from 2.39 days to 1 day, and the process cycle efficiency (PCE) has gone from 0.15% to 0.358%.

Next up, we will discuss action planning and how to go about closing the gaps between the current state and future state.  As always, if you have any questions or comments please do share them with us by clicking on the comments link below.

Subscribe to Gemba Academy

If you enjoyed this article please consider subscribing to our full feed RSS. You can also subscribe by email and have new articles sent directly to your inbox.

Keep Reading

Read the last article in this series: Introducing the Kaizen Newspaper

Tagged in:,

8 Comments

  1. Charles Watson

    February 26, 2008 - 8:50 pm

    Can you clarify what you mean when you say, “Once the work was redistributed the team was able to produce a PB&J sandwich approximately every 32 seconds. The total time for the first sandwich to be complete is still 97 seconds.”

    I am not sure I follow this? Is the cycle time 32 seconds or 97 seconds? And do the number of operators matter here?

    Sorry if this is obvious. I am new to this and am trying to learn. Thank you.

  2. Ron Pereira

    February 26, 2008 - 9:03 pm

    Thanks for the great question Charles. In hindsight, I probably could have been more clear with this section. Here is the explanation.

    The total cycle time to make a sandwich is 97 seconds. However, when we work in one piece flow (no waiting or delay between steps) while using 3 operators we simply divide 97 by 3 which gets us around 32 seconds.

    Now, this assumes we can balance the work evenly… essentially giving each operator around 32ish seconds of work to do.

    In real life, it may be that operator 1 has 30 seconds of work, operator 2 has 35, and operator 3 has 32 seconds of work.

    Does this help?

  3. Charles Watson

    February 26, 2008 - 9:36 pm

    Thank you. Yes, it helps. I knew it was probably something easy like that.

  4. nick ong

    March 4, 2008 - 6:17 am

    i recently have been asked by my boss to do a future state map for despatch area, but as you know despatch area have no takt time or lead time. so where should i start from? any ideas?

  5. Robert

    February 13, 2009 - 11:50 pm

    One things that I note regarding the measurement phase. What would you advise about confirming measurement? A GR&R or a some sort of test such as t-test? On thing about the Takt time would be better to round down the take time to protect the customer?

  6. Pankaj

    August 28, 2014 - 6:06 am

    How to draw VSM is we have more than one m/c in one process and and more than one person in one process or one m/c.
    How does it affect cycle time or takt time and how to decide optimum no. of person ref. to takt time.