The common heijunka definition, production leveling, means transforming the typical peaks and valleys of customer demand into something flatter. That flatness, in turn, makes standardizing production processes easier.
If you like this reference guide, please help us spread the word about it!
Why does heijunka work? Customer orders, in the short term, tend to be highly variable, but as the window of time increases, there is more evenness to demand. For example, weekly demand over a 3 week period might be 200, 218, 197-fairly close. The Monday demand for those same three weeks could be 37, 61, 47-not quite as stable. And day-to-day demand might fluctuate even more wildly-Wednesdays might be the day that three of your biggest customers all place their orders (hey, not everyone is Lean, right?).
Your product mix also likely has similar variation. The breakdown of what people order every day probably fluctuates. Over a longer period, product mix is probably smooth, but the day to day change in product mix might be significant.
Heijunka is a workaround for uneven demand. It is the bridge between how customers order and how a Lean company operates.
Think about this: Lean promotes pull, and at the same time pushes for standard work. In truth, the two concepts don’t play very well together. Standard work promotes rhythm. Pull promotes responsiveness. If your customers don’t buy at an even pace…well, you can see the problem.
In Lean, Heijunka aims to find a way to level the demand on the factory to prevent having wild shifts in production. In effect, a heijunka system uses actual demand from customers over a long period to set production rates, and then stabilizes the short term demand based on those average rates. That way, you only have to shift production schedules to match long term changes in customer demand. You don’t chase the small, day-to-day fluctuation.
In the example above, the average demand is 205 units. If you set your production at 41 per day, you would be balanced with your demand, on average. In the example above, though, you would end up with 4 extra units the first Monday (overproduction), and then run 20 units behind the second Monday. But, at 41 per day, you’d end up flush on the third Friday-even though you might have only gotten the daily demand to match production a couple of times.
For the three week period, you were right on the money. Plus, you kept from having to wildly shift production around in your facility. Just remember–during those three weeks, you were either pulling from a finished goods inventory, or increasing your order backlog whenever your demand wasn’t identical to your production.