Value Stream Mapping for Fun and Profit

Guest post by Evan Durant, author of the Kaizen Notebook blog [The broken link was removed].

I tend to get a little preachy about the importance of value stream maps, but they really can be useful tools not only to plan an improvement effort but also to monitor your progress going forward. In particular they provide a way to quantify the impact of changes to your process. Here’s a real life example as a case in point.

For a particular value stream a team went to gemba, followed the flow of material and information, collected process cycle times, and counted inventory. When everything was mapped and all the data tallied, here was the current state that they came up with:

So what does all this mean? First of all the Total Lead Time represents the amount of time that a new piece of raw material would take to enter the value stream, be worked on, wait around with all the rest of the material in process, and then finally make its way to the customer. This number is usually driven higher by large amounts of in-process inventory caused by pushing between operations.

Second, the Process Lead Time is the amount of time it would take to process a single batch through the process, if it didn’t have to wait behind any other batches. Note that even though parts are processed one at a time through all of the manual operations, a certain amount of batching is required to overcome long machine cycle times in automatic operations. Also we do not ship parts to the customer one at a time, but rather in standard package sizes.

Third, we have the Process Time. This is the total amount of value added time, manual and automatic processing, that a part sees in the value stream.

Finally the Operator Cycle Time (also called manual time) is the amount of actual “touch” time required to make a part. The difference between the Process Time and the Operator Cycle Time is the Machine Cycle Time (also called automatic time). This is when a batch of parts is on a machine that does not require any operator intervention during a cycle. (We have a lot of machine cycle time in this value stream.)

Then the team applied the concepts of flow and pull to reduce overproduction and pace the value stream to the rate of customer demand. The results of the future state map were as follows:

A pretty decent result: a 46% lead time reduction without making any changes to the actual process or to the batch size. This is the magic of demand-based pull in a value stream.

But here’s where it gets interesting. During implementation the team was considering increasing batch size in order to smooth out some issues with the new kanban system. Rather than speculate as to the impact of such a change (as might have been the case in the past) the team returned to the future state map, changed the batch size, and recomputed the timeline. Here’s what they came up with:

Now the team could see that making the change would result in almost a day of added lead time. (Pretty cool, huh?) And since the calculation for kanban quantity is basically (Lead Time x Daily Demand), they could figure the exact cost of the extra inventory that would be required. So were the benefits of the change worth the added cost in inventory and lead time? Maybe, but at least we could make an informed decision.

So the point is that value stream mapping shouldn’t be a one-time activity after which the maps are rolled up and archived somewhere never to be looked at again. The real value is in continuing to use them, keeping them updated, and looking for opportunities for continued improvement. It’s a journey. You need a good map.

Previous guest posts include: Where to Start Improvement by Dave Nave – One factor at a time (OFAT) Versus Factorial Designs by Bradley Jones – The Achilles’ Heel of Agile by Jurgen Appelo