What Exactly Is AB Control at Toyota?

In lean, you may have heard the term AB control. This is often used at Toyota, but is also often confused in the West, especially with the terms one-piece flow and lot size one. AB control is related, but different. In fact, it is closer to a pull production than to lot size one and one-piece flow. But all three help enable a smooth flow of the product toward the customer without overproduction. Let’s first have a look at the other terms:

Lot Size One

Lot size one is probably the simplest of these terms. It simply means that you create products of one type in lot size one. In other words, every product in your production system could differ in its type from the next one. The big advantage of lot size one is that you are more flexible, need less buffer inventory to cover the time between lots, and overall can produce so much closer to the customer demand. See also my post series How to Determine Your Lot Size and Toyota’s and Denso’s Relentless Quest for Lot Size One.

 

One-Piece Flow

One-piece flow moves each product to the next stage as soon as it is completed at the previous stage in the value stream. The next stage should be ready to accept the part immediately. There is no accumulation of larger batches for moving material. The basic idea of one-piece flow is that you do not wait until you have a pallet/box/batch of parts before moving them to the next process, but that you move smaller quantities more frequently. Ideally, the transport quantity is one. The benefit of one-piece flow is also a reduction of fluctuations and hence a reduction of work in progress inventory. You don’t get a big box and then nothing; instead, you get the parts one by one, ideally just in time when you need them.

This helps to reduce transit time of the material through the value stream and reduces the lead time. The system can react faster and information can flow quicker. For example, if a subsequent station notices a systematic defect, a faster transit time means that there is less material in between that has to be reworked or scrapped. See also my posts What is One-Piece Flow?, What Are the Advantages of One-Piece Flow?, and Steps Toward One-Piece Flow.

Pull Production

Before I go into AB control, I need to briefly explain the fundamentals of pull production. There is a lot of confusion and misinformation on what pull actually is. It has nothing to do with make-to-stock and make-to-order (as both can be in a pull system), it is not MRP vs. Kanban (as MRP nowadays can easily do kanban), and it is not about the very vague direction of the information flow. Instead, a pull system is a way to limit yor inventory (or more precisely, your workload, of which the inventory is often a good approximation). For me, a pull system is defined as follows (as described in my award winning book All About Pull Production).

• A pull system must have an explicit target limit on your inventory or workload!
• A pull system must release a signal when an item or batch of items leaves the system. For batches of material, the signal can be with either the first or the last item in a batch.
• This signal must start replenishment for make-to-stock items or release the next job for make-to-order items. The replenishment or release must be the same quantity or the same workload as the items that left the system.
• A system missing any of these three requirements above is a push system.

For more see my blog post on The (True) Difference Between Push and Pull.

Why We Need AB Control

So, now let’s have a look at what AB control actually is. It is a way to support pull production. It is a way to limit the inventory in a production system. With a kanban-based pull system, you can have never more products in production than what you have in corresponding kanban cards. However, usually most kanban cards are in the supermarket, waiting for the consumption of the associated product (after which the kanban initiates a reproduction of the next such product).

However, the problem is that depending on the customer demand pattern, a lot of parts (of different types) could be taken out of the supermarket in a short time. Subsequently, a lot of kanban could enter the production system at the same time. And this would defeat the purpose of a pull system, as it now overloads the production system again with work (even though the overall inventory is still constrained by the number of kanban). Such an overstuffing of the production system would be bad.

However, there is one way to prevent this: FIFO (First in First Out). A FIFO limits the inventory between two processes. You can never have more inventory between two processes than you have space in your FIFO lane. Hence, limiting the FIFO lanes will limit your overall inventory in your production system. I have written a lot about FIFO before, including Theory and Practice on FiFo Lanes – How Does FiFo Work in Lean Manufacturing?, Why FIFO: The Benefits of First In, First Out, Maintaining Strong FIFO in Parallel FIFO Lanes, Maintaining Weak FIFO in Parallel FIFO Lanes, Delivery Sequences: FIFO, LIFO, and Others, Supermarket vs. FiFo – What Requires Less Inventory?, Top Five Cases When NOT to Use a FiFo, Ten Rules When to Use a FIFO, When a Supermarket, and more.

So, if we have FIFO, we should be all good, right? For most users yes, but Toyota is not most users, and this is not good enough for Toyota. The FIFO is a buffer to decouple fluctuations between processes (or more generally between before the FIFO and after the FIFO). It fluctuates. Hence, the size of the FIFO is determined by the size of the fluctuations (read Determining the Size of Your FiFo Lane – The FiFo Formula if you want the math behind it. Warning! Integrals, theory, and not really a practical application!).

Hence, the risk is if you overstuff your production system, all FIFO lanes may be full. While this is okay from the decoupling point of view, it is too much inventory, especially for a extra-lean company like Toyota.

What Is AB Control?

So, after much ado I finally get to explain AB control. It is simply a tool to limit the number of parts or jobs in your system at any given time. The hard limit would be the total capacity of all the FIFOs in the system, but if you want to have even less inventory (meaning some FIFOs always have some empty spaces), then you need to count what goes in and what comes out. For example, if your system could physically fit 300 parts, but you want it to have only 100 parts, you need to track the parts in the system. Only if you have less than 100 parts or jobs in the system, another part or job can enter the system. This is the basics of AB control.

An AB control can be established between two processes (if you want less parts than you have space in the FIFO), but then it would be just easier to make the FIFO smaller. Ideally, the AB control spans multiple processes, maybe even the entire line. But overall, it limits your inventory even more than what would physically be possible. As such, it is very similar to a CONWIP Systems (Constant Work in Progress). But, before you think about installing an AB control system, first try to get a proper pull system running. Once it is running (running well, mind you), you can think about AB control, but it is one of the more higher end techniques.

Now, go out, get your inventory under control, and organize your industry!

PS: After 13 years and 660 posts with one post per week like clockwork I need a bit of a break. I will still write more posts, but from now on somewhat irregular and no longer every week.Hence, it may be a few weeks (or more?) til the next post. Many thanks for your understanding 🙂