PullScheduling.com

I. The Case for Pull Systems

This Chapter will review Pull System fundamentals and the benefits of applying Pull Systems in your business.

Let’s Get Started

My first exposure to Pull Systems was in the mid 1980’s when I was working for a subway car re-manufacturer. In re-manufacturing, you take used products and rework and upgrade them so they are like new (or better). In the subway car industry, the cars are still in revenue service when the Transit Authority contracts to have them re-manufactured. In order to get them worked on, they have to remove them from revenue service, which they dislike doing (for obvious reasons). When they do remove them from service, they want them back in the shortest lead-time possible.

Most re-manufacture contracts in the subway car industry have requirements for these three items:

Work in Progress (WIP) – The number of subway cars the Transit Authority will allow the re-manufacturing company to have out of revenue service at any given time.

Lead-Time (LT) – The amount of time from when the subway car is released from service until it is returned to service.

Rate – The number of subway cars delivered to the Transit Authority per unit of time.

During this time, I was responsible for both the scheduling of the re-manufacturing process, and the monitoring and analysis of costs related to this process. After a while, I realized that there was a relationship between these three variables:

Lead-Time = WIP / Rate

Later on in my career, I learned that someone named “Little” had discovered this formula before I had and named it “Little’s Law”. This simple little formula is very powerful and is one of the keys to pull systems.

Let’s look at a couple of examples. If my WIP was capped at 10 by contract, and I was able to produce 2 units per week, what would my lead-time be?

LT = WIP / Rate = (10 units) / (2 units per week) = 5 weeks

If my WIP was capped at 20 units, and I was required to maintain a lead-time of 4 weeks, what would my rate have to be?

First transform the formula LT = WIP/Rate to Rate = WIP / LT

Rate = WIP / LT = (20 units) / (4 weeks) = 5 units per week

Later on, when I was exposed more formally to pull systems, I realized that our re-manufacturing customers had forced a Pull System on us. The general definition of a Pull System is:

Pull System – A scheduling system where inventory is limited in some way

Note – Inventory can be raw materials, WIP, finished goods, etc.

Most businesses do not put hard limits on their inventories. Without a limit, the amount and location of inventory at any given time is fairly random and will tend to grow as time goes on. This is considered a Push System, which we will define as:

Push System – A scheduling system where inventory is not formally limited

How did we get to Push Systems?

Since we were young, we have been taught some basic work concepts such as:

• Work hard
• Go fast
• Get ahead

The measures and incentives we often use in business support these concepts. Typically, workers and departments will have measures and incentives for units per day, units per person, schedule attainment, etc.

Figure 1  

Let’s look at a typical business scenario. A worker works at a desk with a computer on it. The worker is given a schedule of what work should be done when. Jobs arrive at the desk and go into an “In-Box”. The worker reviews the schedule, takes the next scheduled Job from the In-Box, processes it using the computer, and puts it in the “Out-Box” when completed. Periodically, the worker will get up and move any Jobs currently in the Out-Box to the In-Box of the next step in the process.

Each week, the worker’s supervisor reviews the number of Jobs completed and whether the Jobs were completed per the schedule. If targets were not met, the supervisor follows up with the worker to take corrective action.

Based on the measurements in place, and the worker’s focus on “Work Hard, Go Fast, Get Ahead”, the worker tries to stay busy. As soon as a job goes to the Out-Box, the next job on the schedule is gathered from the In-Box and started. If the next job is not available, but a later job is, the worker starts that job instead.

What’s wrong with Push?

 There are two assumptions we often make when managing businesses:

1. Our business processes are independent. A change in one process will rarely affect our other processes in any adverse way.
2. Our business processes are additive. To see how well the business is working, see how well each of the processes is working and essentially add them up. To improve performance and profitability, get each process to be as efficient as possible. An improvement to one process will improve the bottom line of the company.

Based on these assumptions, we manage our business in a certain way. We try to optimize each department, area, and individual process. Doing this will maximize performance and profitability. We often schedule each process separately – if they all follow the schedule, the overall schedule will be met. We give each process performance measurements so we can tell if they are being “efficient”.

Here is the problem – These assumptions are wrong! The following two principles replace the flawed assumptions:

1. Our business processes are dependent. A change in one process will tend to affect some or all of the other processes in some way.
2. Our business processes are not additive. An improvement to one process will usually not affect the overall performance and profitability of a business in a positive way. In fact, based on the 1^st principle, what is perceived to be an improvement in one area will actually reduce the overall performance of the system.

Let’s look at an example of this.

Figure 2

The diagram in Figure 2 shows a “Business Circle”. Every business has a series of processes that ultimately link together to form a circle. The names and capabilities of each box will be different for each business, but the general concept will always be the same.

Let’s say that the “Engineering” box represents the worker in the Push System we described previously. If the worker is meeting the 15 per day goal and is following the schedule exactly, how many Jobs will the business ship per day? The answer is 12 per day. In this circle, we can only go as fast as our slowest process, which is currently Purchasing at 12 per day.

Where are the other 3 per day going? They are stacking up in front of Purchasing. What is happening to out lead-time?

LT = WIP / Rate, WIP is growing and Rate is staying the same. Therefore, LT is growing!

Is a growing lead-time a bad thing? Absolutely!

Every Lean Manufacturing presentation from the MEP Network has the following quote from Henry Ford:

“One of the most noteworthy accomplishments in keeping the price of Ford products low is the gradual shortening of the production cycle. The longer an article is in the process of manufacture and the more it is moved about, the greater is its ultimate cost.”

Essentially what Henry Ford is saying is that the longer the lead-time, the greater the cost, which provides the following formula:

Lead-Time = Cost

This formula is the essence of Lean Manufacturing, which is focused on the elimination of non-value added activities. Typically, 95%+ of a products lead-time is non-value-added time. Therefore, if we focus on lead-time reduction, we will inherently reduce non-value-added-time and become “Leaner”.

Lets go back to our Engineer. At the end of the year, his supervisor was real happy because he met his output and scheduling targets. He received a big raise, and then was challenged to improve by getting his daily output up from 15 Jobs per day to 17 Jobs per day. His supervisor said this was necessary because all the supervisors had been tasked by the Big Boss to improve efficiency this year.

After much analysis, hard work, and overtime, the Engineer implemented several changes to his computer programs that enabled him to get his daily output up to 17 per day. At the same time, the other departments didn’t have quite as much success. In fact, Purchasing was still stuck at 12 per day.

What affect did the improvement in Engineering have on the overall business?

LT = WIP / Rate, WIP is growing and Rate is staying the same. Therefore, LT is growing!

LT = Cost, LT is growing. Therefore, Cost is growing!

Wait a minute, what is going on here. Engineering improved, but costs went up! Why is this?

Lets go back to the two principles. Our processes are not independent, as a change in Engineering did affect the other processes in our system. Also, our processes are not additive, so an improvement in Engineering did not improve our bottom line.

So what are the implications of these principles?

1. We must manage the entire system to optimize the business.
2. Typically, one process will limit the effectiveness of the system. This is considered the system constraint. We can make improvements to the overall system by improving the Constraint. All others processes must be subordinated to the constraint.
3. We can improve lead-time and Costs by reducing WIP, as long as we don’t ever starve the Constraint for work.

Instead of making more stuff that wasn’t going anywhere, wouldn’t we have been better off if the Engineer only made what Purchasing needed and then used the spare time to help Purchasing process orders?

Why does reducing lead-time reduce costs?

This is not the way we typically look at costs. In fact, the Cost Accounting systems that 95% of businesses use today do not reflect this at all. These systems, which were developed in the early 1900’s when operations were much different than they are today, allocate overhead based on labor and / or materials. In Traditional Cost Accounting, the only time a Job accumulates cost is during the actual step where labor or materials are applied to the product. This is typically a value added step and accounts for less than 5% of the total lead-time. The time between steps, where the product is not having value added to it, are considered “free” by this Cost Accounting system.

A graph of cumulative product cost throughout the lead-time to complete the Business Circle looks like this:

Figure 3

However, applying Henry Ford’s philosophy of LT=Cost yields a graph that looks like this:

Figure 4

Goldratt, who created the Theory of Constraints, once said:

“Tell me how you measure me and I’ll tell you how I’ll behave.”

If our measurements drive our behaviors, what type of behavior is traditional cost accounting driving? Instead of focusing on lead-time, traditional cost accounting causes us to focus on the small portion of time when labor and material are applied to the product.

The benefits of reducing processing lead-time are too numerous to list here. Insyte Consulting, which is the MEP Center in Buffalo, NY, has a spreadsheet on their website that documents 31 benefits of reducing lead-time. Click here for a link to this spreadsheet. These benefits generally roll up into 6 categories:

1. Increased sales
2. Improved quality
3. Reduced operating costs
4. Increased capacity and throughput
5. Reduced invested assets
6. Increased employee satisfaction (morale)

Let’s use a production work center at a manufacturing company as an example. The work center is located half way through the overall product routing. Say the work center has 1 month of WIP valued at $100,000 stacked up in front of it.

Figure 5

According to the Traditional Cost Accounting system, this WIP is not costing the company anything. In fact, it is included as an asset on the balance sheet. Also, if aggregate WIP grows during any particular year, it will appear to make the company more profitable on an accrual accounting income statement.

According to Lean Cost Accounting, which is a more correct way of looking at this, this WIP is costing the company money. Why is that? The following are a few of the reasons why:

• Cost of capital – A typical company will have a line of credit and / or wish to lease or purchase equipment. If the $100,000 was not sitting here, the company could reduce it’s credit line by that amount or invest in new equipment or other assets.
• Cost of space – That month’s worth of WIP will take up space. Manufacturing space costs money. We also have costs associated with heating, cooling, and lighting that space.
• Cost of handling – During the month that these products sit waiting to be worked on, they will usually get moved, stocked, re-sorted, split, combined, etc, numerous times. The cost to do this includes not only the labor, but the resulting damage to the products and the potential for injuries.
• Cost of quality – If we discover that something is wrong with a product in this WIP inventory, how many might be bad? Potentially all of them! These will have to be re-worked or scrapped and re-made. Since the previous manufacturing steps occurred a month or more ago, our chances of identifying the root cause of the problem and taking effective corrective action are low.
• Lost throughput – The presence of random WIP all over the place effectively masks where our slowest operation is. Without knowing where this is and focusing our efforts on it, we loose overall throughput in the system
• Lost sales opportunities – Many customer orders are awarded based on supplier lead-time. In general, customers are poor planners. Since LT=WIP/Rate, this WIP is adding to lead-time, which is taking away this market opportunity.
• Reduced on-time delivery – The longer your processing lead-time, the harder it is to deliver on-time. Also, the more likely it is that customers will change designs, quantities, due dates, etc.

A full introduction to Lean or TOC cost accounting is beyond the scope of this document. However, a good general rule when making business decisions is to focus on the decisions effect on Cash Position and Cash Flow versus accrual accounting profit. Decisions that improve overall Cash Position and / or Cash Flow will ultimately improve your business. Also use this method when evaluating relative success after implementation of any improvement idea.

How does all of this relate to Pull Systems?

There are many things that contribute to a long processing lead-time. However, the 80/20 rule would suggest that a few policies / systems contribute the most to our overall lead-time. In my experience, the low hanging fruit that has the biggest impact on our lead-time is:

• Scheduling systems
• Batch sizing

Pull systems address both scheduling and batching. The application of a few relatively simple pull techniques can quickly reduce processing lead-times by 50% or more.

Many Lean practitioners consider Pull System to be advanced techniques to be applied late (if ever) in an overall Lean implementation. TOC practitioners consider Pull Systems to be a foundational step that is usually applied first in any improvement initiative. I strongly agree with the TOC approach. If lead-time is the main focus, and if Pull Systems are a method of addressing the major contributors to long lead-time, then lead-time improvement initiatives should start with the application of Pull Systems.

You will find that once you have implemented Pull Systems and your lead-times are reduced, the other critical issues in your system will become painfully obvious and easier to address. You will also have a tool to focus your limited improvement resources on initiatives that will really make an impact and not waste them on initiatives with no (or negative) return on investment.

To get the entire eBook, simply add to your cart and complete the free checkout process. The eBook will then be emailed directly to you.

 

 $0.00 Pull Scheduling Systems Overview

 

This eBook is now available in Russian, Simplified Chinese, and Complex Chinese.